Email alert
Current Issue
2024 Vol. 46, No. 5
Display Method:
2024, 46(5): 1529-1545.
doi: 10.11999/JEIT231285
Abstract:
With the widespread application of artificial intelligence technology, the demand for computing power for intelligent computing has grown exponentially. At present, the rapid development of chips has approached the bottleneck of its manufacturing process, and power consumption is also increasing. Therefore, research on high-speed, energy-efficient intelligent computing hardware is an important direction. Computing architectures represented by photonic circuit neural networks and all-optical diffraction neural networks have received widespread attention due to their advantages such as fast calculation and low power consumption. This article reviews the representative work of optical neural networks, and introduces it through the two main lines of development of three-dimensional diffractive neural networks and optical neural network chips. At the same time, it focuses on the bottlenecks and challenges faced by optical diffractive neural networks and photonic neural network chips, such as network scale and Integration degree, etc., analyzes and compares their characteristics, performance and respective advantages and disadvantages. Secondly, taking into account the development needs of generalization, this article further discusses the programmable design of neuromorphic computing hardware, and introduces some representative work on programmable neural networks to each part. In addition to intelligent neural networks in the optical band, this article also discusses the development and application of microwave diffraction neural networks and demonstrates their programmability. Finally, the future direction and development trends of intelligent neuromorphic computing are introduced, as well as its potential applications in wireless communications, information processing and sensing.
With the widespread application of artificial intelligence technology, the demand for computing power for intelligent computing has grown exponentially. At present, the rapid development of chips has approached the bottleneck of its manufacturing process, and power consumption is also increasing. Therefore, research on high-speed, energy-efficient intelligent computing hardware is an important direction. Computing architectures represented by photonic circuit neural networks and all-optical diffraction neural networks have received widespread attention due to their advantages such as fast calculation and low power consumption. This article reviews the representative work of optical neural networks, and introduces it through the two main lines of development of three-dimensional diffractive neural networks and optical neural network chips. At the same time, it focuses on the bottlenecks and challenges faced by optical diffractive neural networks and photonic neural network chips, such as network scale and Integration degree, etc., analyzes and compares their characteristics, performance and respective advantages and disadvantages. Secondly, taking into account the development needs of generalization, this article further discusses the programmable design of neuromorphic computing hardware, and introduces some representative work on programmable neural networks to each part. In addition to intelligent neural networks in the optical band, this article also discusses the development and application of microwave diffraction neural networks and demonstrates their programmability. Finally, the future direction and development trends of intelligent neuromorphic computing are introduced, as well as its potential applications in wireless communications, information processing and sensing.
2024, 46(5): 1546-1569.
doi: 10.11999/JEIT231300
Abstract:
The ciphertext retrieval techniques are designed to provide query services over encrypted data and to improve the availability of encrypted data. However, most of existing methods leak additional information besides the query result, which may be utilized by the attacker to recover plaintext data or queries. How to enhance the privacy-preserving features in ciphertext retrieval and achieve the minimisation of information leakage has received a lot of attention from researchers. In recent years, with the rapid development of hardware chip technology and new cryptographic technology, a number of novel methods were proposed for privacy-preserving ciphertext retrieval. This paper mainly focuses on the research hotspots of diversified ciphertext retrieval, ciphertext retrieval based on trusted execution environment and private information retrieval, and summarizes the future development trends.
The ciphertext retrieval techniques are designed to provide query services over encrypted data and to improve the availability of encrypted data. However, most of existing methods leak additional information besides the query result, which may be utilized by the attacker to recover plaintext data or queries. How to enhance the privacy-preserving features in ciphertext retrieval and achieve the minimisation of information leakage has received a lot of attention from researchers. In recent years, with the rapid development of hardware chip technology and new cryptographic technology, a number of novel methods were proposed for privacy-preserving ciphertext retrieval. This paper mainly focuses on the research hotspots of diversified ciphertext retrieval, ciphertext retrieval based on trusted execution environment and private information retrieval, and summarizes the future development trends.
2024, 46(5): 1570-1581.
doi: 10.11999/JEIT240143
Abstract:
For the first time, a simplified architecture for 5G millimeter-wave retrodirective arrays and its implementation in CMOS chips is reported in this paper. Phase conjugation and retrodirective functions are provided by sub-harmonic mixers in the simplified architecture, eliminating the need for phase-shifting circuits and beam-controlling systems, thereby enabling automatic beam tracking for mobile communications. For validation, a domestic 0.18 μm CMOS process is employed to realize a 5G millimeter-wave retrodirective array chip, comprising core modules such as the transceiver front-ends and tracking phase-locked loop. Measured conversion gains of 19.5 dB for transmitting and 18.7 dB for receiving are achieved by the transceiver front-end chip utilizing sub-harmonic mixing and gm-boosting techniques. Dual-mode operation capabilities, supporting both amplitude modulation and phase modulation based on different reference signals, are provided by the implemented tracking phase-locked loop chip, with measured output signal phase noise lower than –125 dBc/Hz@100 kHz. The feasibility of the proposed 5G millimeter-wave retrodirective array communication architecture and its CMOS chip implementation is validated by the test results presented in this paper, thus offering a new solution for 5G/6G millimeter-wave communication characterized by its extremely simplified architecture, low cost, and high scalability.
For the first time, a simplified architecture for 5G millimeter-wave retrodirective arrays and its implementation in CMOS chips is reported in this paper. Phase conjugation and retrodirective functions are provided by sub-harmonic mixers in the simplified architecture, eliminating the need for phase-shifting circuits and beam-controlling systems, thereby enabling automatic beam tracking for mobile communications. For validation, a domestic 0.18 μm CMOS process is employed to realize a 5G millimeter-wave retrodirective array chip, comprising core modules such as the transceiver front-ends and tracking phase-locked loop. Measured conversion gains of 19.5 dB for transmitting and 18.7 dB for receiving are achieved by the transceiver front-end chip utilizing sub-harmonic mixing and gm-boosting techniques. Dual-mode operation capabilities, supporting both amplitude modulation and phase modulation based on different reference signals, are provided by the implemented tracking phase-locked loop chip, with measured output signal phase noise lower than –125 dBc/Hz@100 kHz. The feasibility of the proposed 5G millimeter-wave retrodirective array communication architecture and its CMOS chip implementation is validated by the test results presented in this paper, thus offering a new solution for 5G/6G millimeter-wave communication characterized by its extremely simplified architecture, low cost, and high scalability.
2024, 46(5): 1582-1590.
doi: 10.11999/JEIT240202
Abstract:
High-resolution satellite remote sensing images have been recognized as an indispensable means for understanding geographical spaces, and their role in areas such as urban mapping, ecological monitoring, and navigation, has become increasingly important. The use of satellite remote sensing images for large-scale 3D reconstruction of the Earth’s surface is currently a subject of active research in the fields of computer vision and photogrammetry. Neural Radiance Fields (NeRF), which utilizes differentiable rendering to learn implicit representations of scenes, has achieved the most realistic visual effects in novel view synthesis tasks of complex scenes and has attracted significant attention in the field of 3D scene reconstruction and rendering. Recent research has been primarily focused on using neural radiance field technology to extract scene representation and reconstruction from satellite remote sensing images. Ray space optimization, scene representation optimization, and efficient model training are mainly focused on by the neural radiance field methods for satellite remote sensing images. The latest progress in the application of neural radiance field technology in satellite remote sensing is comprehensively summarized in this paper. First, the basic concepts of neural radiance field technology and related datasets are introduced. Then a classification framework of neural radiance field methods for satellite remote sensing images is proposed to systematically review and organize the research progress of this technology in the field of satellite remote sensing. The relevant results of the application of neural radiance field technology in actual satellite remote sensing scenarios are detailed. Finally, analysis and discussion are conducted based on the problems and challenges faced by current research, and future development trends and research directions are prospected.
High-resolution satellite remote sensing images have been recognized as an indispensable means for understanding geographical spaces, and their role in areas such as urban mapping, ecological monitoring, and navigation, has become increasingly important. The use of satellite remote sensing images for large-scale 3D reconstruction of the Earth’s surface is currently a subject of active research in the fields of computer vision and photogrammetry. Neural Radiance Fields (NeRF), which utilizes differentiable rendering to learn implicit representations of scenes, has achieved the most realistic visual effects in novel view synthesis tasks of complex scenes and has attracted significant attention in the field of 3D scene reconstruction and rendering. Recent research has been primarily focused on using neural radiance field technology to extract scene representation and reconstruction from satellite remote sensing images. Ray space optimization, scene representation optimization, and efficient model training are mainly focused on by the neural radiance field methods for satellite remote sensing images. The latest progress in the application of neural radiance field technology in satellite remote sensing is comprehensively summarized in this paper. First, the basic concepts of neural radiance field technology and related datasets are introduced. Then a classification framework of neural radiance field methods for satellite remote sensing images is proposed to systematically review and organize the research progress of this technology in the field of satellite remote sensing. The relevant results of the application of neural radiance field technology in actual satellite remote sensing scenarios are detailed. Finally, analysis and discussion are conducted based on the problems and challenges faced by current research, and future development trends and research directions are prospected.
2024, 46(5): 1591-1603.
doi: 10.11999/JEIT240032
Abstract:
Satellite communication plays a crucial role in addressing blind spots and enhancing ubiquitous coverage for future global ubiquitous communication needs. Direct-to-Smartphone(DS) technology, as a technological means to achieve global integrated space-ground and intelligent connection of all things in the future 6G network, has become a hot development topic worldwide in the past two years and has received widespread attention. The development status and mainstream technical routes of DS technology in this article both domestically and internationally are introduced. The development challenges have been analyzed on DS technology in terms of limited frequency resource usage, broadband service requirements of DS, massive user business with time-varying and non-uniform distribution, high-dynamic effect of low earth orbit satellite, ultra-dense multi-beam influence, and high-density integration of smartphone. Key solutions such as satellite-ground co-frequency sharing, ultra-large array spaceborne multi-beam antenna, multi-dimensional resource management and control of satellite and ground, adapting to high dynamic satellite air interface system, on-demand scheduling of extremely narrow beam, and highly integrated smartphone design are proposed in this paper. Finally, the future development of DS technology is discussed.
Satellite communication plays a crucial role in addressing blind spots and enhancing ubiquitous coverage for future global ubiquitous communication needs. Direct-to-Smartphone(DS) technology, as a technological means to achieve global integrated space-ground and intelligent connection of all things in the future 6G network, has become a hot development topic worldwide in the past two years and has received widespread attention. The development status and mainstream technical routes of DS technology in this article both domestically and internationally are introduced. The development challenges have been analyzed on DS technology in terms of limited frequency resource usage, broadband service requirements of DS, massive user business with time-varying and non-uniform distribution, high-dynamic effect of low earth orbit satellite, ultra-dense multi-beam influence, and high-density integration of smartphone. Key solutions such as satellite-ground co-frequency sharing, ultra-large array spaceborne multi-beam antenna, multi-dimensional resource management and control of satellite and ground, adapting to high dynamic satellite air interface system, on-demand scheduling of extremely narrow beam, and highly integrated smartphone design are proposed in this paper. Finally, the future development of DS technology is discussed.
2024, 46(5): 1604-1612.
doi: 10.11999/JEIT231288
Abstract:
Triple Modular Redundancy (TMR), as the most prevalent and effective technique for soft error mitigation technique, inevitably incurs substantial hardware overhead while meeting high fault-tolerance requirements. To achieve the trade-off between area, power and fault coverage and meet the requirement of low-cost and high-reliability circuit design, Approximate Triple Modular Redundancy (ATMR) is investigated and a Dynamic Adjustment Multi-Objective Optimization Framework based on Approximate Gate Library (ApxLib+DAMOO) is investigated. The basic optimization framework employs Non-dominated Sorting Genetic Algorithm II (NSGA-II), achieving rapidly approximation through parity analysis and the pre-established ApxLib. Subsequently, the framework introduces two novel mechanisms: dynamic probability adjustment and parity expansion. The first mechanism dynamically updates the mutation probability of gates in the genetic algorithm based on testability analysis, while the second mechanism performs recognition and reconstruction for binate gates to achieve dual optimization of efficiency and effectiveness in optimization. Experimental results indicate that the proposed optimization framework achieves an additional Soft Error Rate (SER) reduction of up to 10%～20% compared to traditional NSGA-II with the same hardware overhead, while reducing 18.7% of execution time reduction averagely.
Triple Modular Redundancy (TMR), as the most prevalent and effective technique for soft error mitigation technique, inevitably incurs substantial hardware overhead while meeting high fault-tolerance requirements. To achieve the trade-off between area, power and fault coverage and meet the requirement of low-cost and high-reliability circuit design, Approximate Triple Modular Redundancy (ATMR) is investigated and a Dynamic Adjustment Multi-Objective Optimization Framework based on Approximate Gate Library (ApxLib+DAMOO) is investigated. The basic optimization framework employs Non-dominated Sorting Genetic Algorithm II (NSGA-II), achieving rapidly approximation through parity analysis and the pre-established ApxLib. Subsequently, the framework introduces two novel mechanisms: dynamic probability adjustment and parity expansion. The first mechanism dynamically updates the mutation probability of gates in the genetic algorithm based on testability analysis, while the second mechanism performs recognition and reconstruction for binate gates to achieve dual optimization of efficiency and effectiveness in optimization. Experimental results indicate that the proposed optimization framework achieves an additional Soft Error Rate (SER) reduction of up to 10%～20% compared to traditional NSGA-II with the same hardware overhead, while reducing 18.7% of execution time reduction averagely.
2024, 46(5): 1613-1631.
doi: 10.11999/JEIT231224
Abstract:
Considering the integrated air-to-ground access network, based on summarizing the relevant research, the key technologies of future air-to-ground integrated access architecture are elaborated, and the research progress in several key directions, such as air-port technology, multiple-access technology, interference analysis, computation technology, and Artificial Intelligence (AI) technology are analyzed, and a flexible network architecture with the coexistence of multiple access forms is proposed. Considering the key research problems of the access architecture in the current air-to-ground integrated network, an integrated AI-enabled architecture is constructed by combining the user’s quality of service demand, and the large-scale hybrid multiple access and flexible resource adaptation strategy is proposed. Based on the future key research directions of network architecture stereoscopic, network cooperative transmission, integrated network resource management, future air-to-ground access technology, and network cooperative computation are discussed and outlooked.
Considering the integrated air-to-ground access network, based on summarizing the relevant research, the key technologies of future air-to-ground integrated access architecture are elaborated, and the research progress in several key directions, such as air-port technology, multiple-access technology, interference analysis, computation technology, and Artificial Intelligence (AI) technology are analyzed, and a flexible network architecture with the coexistence of multiple access forms is proposed. Considering the key research problems of the access architecture in the current air-to-ground integrated network, an integrated AI-enabled architecture is constructed by combining the user’s quality of service demand, and the large-scale hybrid multiple access and flexible resource adaptation strategy is proposed. Based on the future key research directions of network architecture stereoscopic, network cooperative transmission, integrated network resource management, future air-to-ground access technology, and network cooperative computation are discussed and outlooked.
2024, 46(5): 1632-1644.
doi: 10.11999/JEIT240074
Abstract:
The generation of video from text aims to produce semantically consistent, photo-realistic, temporal consistent, and logically coherent videos based on provided textual descriptions. Firstly, the current state of research in the field of text-to-video generation is elucidated in this paper, providing a detailed overview of three mainstream approaches: methods based on recurrent networks and Generative Adversarial Networks (GAN), methods based on Transformers, and methods based on diffusion models. Each of these models has its strengths and weaknesses in video generation. The recurrent networks and GAN-based methods can generate videos with higher resolution and duration but struggle with generating complex open-domain videos. Transformer-based methods show proficiency in generating open-domain videos but face challenges related to unidirectional biases and accumulated errors, making it difficult to produce high-fidelity videos. Diffusion models exhibit good generalization but are constrained by inference speed and high memory consumption, making it challenging to generate high-definition and lengthy videos. Subsequently, evaluation benchmarks and metrics in the text-to-video generation domain are explored, and the performance of existing methods is compared. Finally, potential future research directions in the field is outlined.
The generation of video from text aims to produce semantically consistent, photo-realistic, temporal consistent, and logically coherent videos based on provided textual descriptions. Firstly, the current state of research in the field of text-to-video generation is elucidated in this paper, providing a detailed overview of three mainstream approaches: methods based on recurrent networks and Generative Adversarial Networks (GAN), methods based on Transformers, and methods based on diffusion models. Each of these models has its strengths and weaknesses in video generation. The recurrent networks and GAN-based methods can generate videos with higher resolution and duration but struggle with generating complex open-domain videos. Transformer-based methods show proficiency in generating open-domain videos but face challenges related to unidirectional biases and accumulated errors, making it difficult to produce high-fidelity videos. Diffusion models exhibit good generalization but are constrained by inference speed and high memory consumption, making it challenging to generate high-definition and lengthy videos. Subsequently, evaluation benchmarks and metrics in the text-to-video generation domain are explored, and the performance of existing methods is compared. Finally, potential future research directions in the field is outlined.
2024, 46(5): 1645-1657.
doi: 10.11999/JEIT231188
Abstract:
Currently, artificial intelligence systems have achieved significant success in various domains, with deep learning technology playing a pivotal role. However, although the deep neural network has strong inference recognition ability, it is still vulnerable to the attack of adversarial examples, showing its vulnerability. Adversarial samples are specially crafted input data designed to attack and mislead the outputs of deep learning models. With the rapid development of 3D sensors such as LiDAR, the use of deep learning technology to address various intelligent tasks in the 3D domain is gaining increasing attention. Ensuring the security and robustness of artificial intelligence systems that process 3D point cloud data, such as deep learning-based autonomous 3D object detection and recognition for self-driving vehicles, is crucial. In order to analyze the methods by which 3D adversarial samples attack deep neural networks, and reveal the interference mechanisms of 3D adversarial samples on deep neural networks, this paper summarizes the research progress on adversarial attack methods for deep neural network models based on 3D point cloud data. The paper first introduces the fundamental principles and implementation methods of adversarial attacks, and then it summarizes and analyzes digital domain adversarial attacks and physical domain adversarial attacks on 3D point clouds. Finally, it discusses the challenges and future research directions in the realm of 3D point cloud adversarial attacks.
Currently, artificial intelligence systems have achieved significant success in various domains, with deep learning technology playing a pivotal role. However, although the deep neural network has strong inference recognition ability, it is still vulnerable to the attack of adversarial examples, showing its vulnerability. Adversarial samples are specially crafted input data designed to attack and mislead the outputs of deep learning models. With the rapid development of 3D sensors such as LiDAR, the use of deep learning technology to address various intelligent tasks in the 3D domain is gaining increasing attention. Ensuring the security and robustness of artificial intelligence systems that process 3D point cloud data, such as deep learning-based autonomous 3D object detection and recognition for self-driving vehicles, is crucial. In order to analyze the methods by which 3D adversarial samples attack deep neural networks, and reveal the interference mechanisms of 3D adversarial samples on deep neural networks, this paper summarizes the research progress on adversarial attack methods for deep neural network models based on 3D point cloud data. The paper first introduces the fundamental principles and implementation methods of adversarial attacks, and then it summarizes and analyzes digital domain adversarial attacks and physical domain adversarial attacks on 3D point clouds. Finally, it discusses the challenges and future research directions in the realm of 3D point cloud adversarial attacks.
2024, 46(5): 1658-1671.
doi: 10.11999/JEIT231201
Abstract:
An overview of multimodal communication as an important information transfer mode that can simultaneously interact with multiple modal forms in different application scenarios is proposed in this paper. The future development prospects of multimodal communication in 6G wireless communication technology is also discussed. Firstly, multimodal communication is classified into three categories, and its key roles in these fields are explored. Furthermore, a deep analysis is conducted on the communication, sensation, computation, and storage resource limitations, as well as cross-domain resource management issues that 6G wireless communication systems may face. It points out that future 6G wireless multimodal communication will achieve deep integration of communication perception, computation, and storage, as well as enhance communication capabilities. In the process of implementing multimodal communication, various aspects must be considered, including multi-transmitter processing, transmission technology, and receiver processing, in order to address challenges in multimodal corpus construction, multimodal information compression, transmission, interference handling, noise reduction, alignment, fusion, and expansion, as well as resource management issues. Finally, the importance of cross-domain multimodal information transfer, complementarity, and collaboration in the 6G network is emphasized. This will better integrate and apply a massive amount of heterogeneous information to meet the future communication demands of high-speed, low-latency, and intelligent interconnection.
An overview of multimodal communication as an important information transfer mode that can simultaneously interact with multiple modal forms in different application scenarios is proposed in this paper. The future development prospects of multimodal communication in 6G wireless communication technology is also discussed. Firstly, multimodal communication is classified into three categories, and its key roles in these fields are explored. Furthermore, a deep analysis is conducted on the communication, sensation, computation, and storage resource limitations, as well as cross-domain resource management issues that 6G wireless communication systems may face. It points out that future 6G wireless multimodal communication will achieve deep integration of communication perception, computation, and storage, as well as enhance communication capabilities. In the process of implementing multimodal communication, various aspects must be considered, including multi-transmitter processing, transmission technology, and receiver processing, in order to address challenges in multimodal corpus construction, multimodal information compression, transmission, interference handling, noise reduction, alignment, fusion, and expansion, as well as resource management issues. Finally, the importance of cross-domain multimodal information transfer, complementarity, and collaboration in the 6G network is emphasized. This will better integrate and apply a massive amount of heterogeneous information to meet the future communication demands of high-speed, low-latency, and intelligent interconnection.
2024, 46(5): 1672-1683.
doi: 10.11999/JEIT231045
Abstract:
Facing the demand for interconnectivity sensing of three-dimensional coverage for the sixth-Generation mobile communication (6G) networks and the spectrum scarcity issue caused by the widespread access of wireless devices, the multi-dimensional expanded Integrated Sensing and Communication (ISAC), based on Unmanned Aerial Vehicles (UAV) and Intelligent Reflecting Surfaces (IRS), is capable of achieving synergistic communication and sensing functions in the three-dimensional network space. This can effectively enhance spectrum efficiency, hardware resource utilization, and align with the wireless network vision of 6G Internet of Everything. This paper provides an overview of the architecture for the 6G multi-dimensional expanded ISAC. Firstly, it summarizes the theoretical foundations of the 6G network vision and ISAC networks, and the application scenarios, development trends, and performance indicators of multi-dimensional expanded ISAC based on UAV and IRS are discussed. Then, it investigates the potential applications of 6G key technologies, such as ultra-massive multiple-input and multiple-output antenna, terahertz, simultaneous wireless information and power transfer, artificial intelligence, covert communication, and active IRS, in multi-dimensional expanded ISAC networks based on UAV and IRS. Finally, the future development direction and key technical challenges of 6G multi-dimensional expanded ISAC sre prospected.
Facing the demand for interconnectivity sensing of three-dimensional coverage for the sixth-Generation mobile communication (6G) networks and the spectrum scarcity issue caused by the widespread access of wireless devices, the multi-dimensional expanded Integrated Sensing and Communication (ISAC), based on Unmanned Aerial Vehicles (UAV) and Intelligent Reflecting Surfaces (IRS), is capable of achieving synergistic communication and sensing functions in the three-dimensional network space. This can effectively enhance spectrum efficiency, hardware resource utilization, and align with the wireless network vision of 6G Internet of Everything. This paper provides an overview of the architecture for the 6G multi-dimensional expanded ISAC. Firstly, it summarizes the theoretical foundations of the 6G network vision and ISAC networks, and the application scenarios, development trends, and performance indicators of multi-dimensional expanded ISAC based on UAV and IRS are discussed. Then, it investigates the potential applications of 6G key technologies, such as ultra-massive multiple-input and multiple-output antenna, terahertz, simultaneous wireless information and power transfer, artificial intelligence, covert communication, and active IRS, in multi-dimensional expanded ISAC networks based on UAV and IRS. Finally, the future development direction and key technical challenges of 6G multi-dimensional expanded ISAC sre prospected.
2024, 46(5): 1684-1702.
doi: 10.11999/JEIT230941
Abstract:
Low Earth Orbit (LEO) satellite has the characteristics of low communication delay, low deployment cost and wide coverage, and has become an important part of the construction of the future space earth integrated network. However, satellite communication has large end-to-end propagation distance, complex fading and fast terminal movement speed, thus the channel characteristics are very different from the terrestrial cellular network. Based on this, in order to have a more comprehensive understanding of the characteristics and channel model of LEO satellite-ground channel, the current standardization progress of the satellite-ground channel by the international standards organization are summarized, the fading characteristics of the satellite ground channel at different propagation positions are discussed, the existing important channel models are classified and shown according to the modeling method, and finally the prospects for future work are proposed.
Low Earth Orbit (LEO) satellite has the characteristics of low communication delay, low deployment cost and wide coverage, and has become an important part of the construction of the future space earth integrated network. However, satellite communication has large end-to-end propagation distance, complex fading and fast terminal movement speed, thus the channel characteristics are very different from the terrestrial cellular network. Based on this, in order to have a more comprehensive understanding of the characteristics and channel model of LEO satellite-ground channel, the current standardization progress of the satellite-ground channel by the international standards organization are summarized, the fading characteristics of the satellite ground channel at different propagation positions are discussed, the existing important channel models are classified and shown according to the modeling method, and finally the prospects for future work are proposed.
2024, 46(5): 1703-1715.
doi: 10.11999/JEIT231140
Abstract:
The future Sixth-Generation (6G) wireless communication systems are required to support ultra-large-scale user demands, with increasing demands for spectrum efficiency and energy efficiency. In this context, Holographic Multiple-Input Multiple-Output (HMIMO) technology has gained increasing attention due to its potential for intelligent reconfigurability, electromagnetic tunability, high directional gain, cost-effectiveness, and flexible deployment. In holographic MIMO system, large amount small and cheap antenna units are integrated tightly, thus realize high directional gain at a low hardware cost and flexible adjustment of electromagnetic wave at the same time, thereby effectively enhancing the performance of wireless communication. A brief introduction to holographic MIMO technology is provided at the start of this paper, covering its current status, development process, classification, and key characteristics. Subsequently, the channel model for holographic MIMO in line-of-sight scenarios and non-line-of-sight scenarios with spatially smooth scattering is presented. Finally, the challenges and future trends faced by holographic MIMO technology are described, and the article is concluded.
The future Sixth-Generation (6G) wireless communication systems are required to support ultra-large-scale user demands, with increasing demands for spectrum efficiency and energy efficiency. In this context, Holographic Multiple-Input Multiple-Output (HMIMO) technology has gained increasing attention due to its potential for intelligent reconfigurability, electromagnetic tunability, high directional gain, cost-effectiveness, and flexible deployment. In holographic MIMO system, large amount small and cheap antenna units are integrated tightly, thus realize high directional gain at a low hardware cost and flexible adjustment of electromagnetic wave at the same time, thereby effectively enhancing the performance of wireless communication. A brief introduction to holographic MIMO technology is provided at the start of this paper, covering its current status, development process, classification, and key characteristics. Subsequently, the channel model for holographic MIMO in line-of-sight scenarios and non-line-of-sight scenarios with spatially smooth scattering is presented. Finally, the challenges and future trends faced by holographic MIMO technology are described, and the article is concluded.
2024, 46(5): 1716-1741.
doi: 10.11999/JEIT231195
Abstract:
The rapid development of swarm intelligence, represented by evolutionary computation, has triggered a new wave of technological transformation in the field of artificial intelligence. To meet the diverse application needs of complex systems, artificial intelligence is increasingly moving towards cross-level intelligent and collaborative research. In this paper, the concept of swarm intelligence cooperation oriented towards evolutionary computation is proposed. Based on the hierarchical levels of swarm intelligence cooperation, artificial intelligence research across different levels is categorized into micro-level cooperation, meso-level cooperation, and macro-level cooperation. From the perspective of swarm intelligence cooperation, a summary is provided on recent research in the aforementioned branches. Firstly, the micro-level cooperation is discussed by analyzing decision variable level cooperation and global/local level cooperation. Secondly, the meso-level cooperation is summarized from the dimensions of objective-level cooperation and task-level cooperation. Furthermore, an analysis of macro-level cooperation is conducted through the examination of space-air-ground-sea cooperation, vehicle-road-cloud cooperation, and edge-cloud cooperation in intelligent collaborative systems. Finally, the research challenges in the field of swarm intelligence cooperation oriented towards evolutionary computation are identified, and future directions for related fields are proposed.
The rapid development of swarm intelligence, represented by evolutionary computation, has triggered a new wave of technological transformation in the field of artificial intelligence. To meet the diverse application needs of complex systems, artificial intelligence is increasingly moving towards cross-level intelligent and collaborative research. In this paper, the concept of swarm intelligence cooperation oriented towards evolutionary computation is proposed. Based on the hierarchical levels of swarm intelligence cooperation, artificial intelligence research across different levels is categorized into micro-level cooperation, meso-level cooperation, and macro-level cooperation. From the perspective of swarm intelligence cooperation, a summary is provided on recent research in the aforementioned branches. Firstly, the micro-level cooperation is discussed by analyzing decision variable level cooperation and global/local level cooperation. Secondly, the meso-level cooperation is summarized from the dimensions of objective-level cooperation and task-level cooperation. Furthermore, an analysis of macro-level cooperation is conducted through the examination of space-air-ground-sea cooperation, vehicle-road-cloud cooperation, and edge-cloud cooperation in intelligent collaborative systems. Finally, the research challenges in the field of swarm intelligence cooperation oriented towards evolutionary computation are identified, and future directions for related fields are proposed.
2024, 46(5): 1742-1760.
doi: 10.11999/JEIT231207
Abstract:
With the increasing development of marine resources and underwater operations, sonar image-based underwater target recognition has become a hot research area. This article provides a comprehensive review of the current status and future trends in this field. Initially, the background and significance of sonar image-based underwater target recognition are emphasized, noting that the complexity of the underwater environment and the scarcity of samples increase the task difficulty. Subsequently, typical imaging sonar technologies are delved, including forward-looking sonar, side-scan sonar, synthetic aperture sonar, multibeam echo sounder, interferometric synthetic aperture sonar, and forward-looking 3D sonar. Following that, 2D and 3D sonar image-based underwater target recognition methods are systematically examined, the strengths and weaknesses of different algorithms are compared, and methods for the correlated recognition of sonar image sequences are discussed. Finally, the major challenges in the current field and future research directions are summarized, aiming to foster the development of the underwater sonar target recognition field.
With the increasing development of marine resources and underwater operations, sonar image-based underwater target recognition has become a hot research area. This article provides a comprehensive review of the current status and future trends in this field. Initially, the background and significance of sonar image-based underwater target recognition are emphasized, noting that the complexity of the underwater environment and the scarcity of samples increase the task difficulty. Subsequently, typical imaging sonar technologies are delved, including forward-looking sonar, side-scan sonar, synthetic aperture sonar, multibeam echo sounder, interferometric synthetic aperture sonar, and forward-looking 3D sonar. Following that, 2D and 3D sonar image-based underwater target recognition methods are systematically examined, the strengths and weaknesses of different algorithms are compared, and methods for the correlated recognition of sonar image sequences are discussed. Finally, the major challenges in the current field and future research directions are summarized, aiming to foster the development of the underwater sonar target recognition field.
2024, 46(5): 1761-1773.
doi: 10.11999/JEIT231142
Abstract:
With the drastic increase in the number of users, the existing wireless resources have become unsustainable. Therefore, the reactivation of Power Line Communication (PLC) has attracted the attention of major research institutes and industries. The development of PLC has been slow due to the complexity of the channel environment and the complexity and high cost of existing processing solutions. The most extensive work has been done on impulse noise, and it is particularly important to achieve robustness of data transmission against impulse noise at low cost. Firstly, the mainstream noise in PLC environment and its classification are introduced in this paper, and then the Differential Chaos Shift Keying (DCSK) and M-ary DCSK (MDCSK) modulation techniques with low cost and low complexity are described. The characteristics of this system in PLCs are presented and analysed, as well as the advantages and improvements that exist for various types of impulse noise. Secondly, some relevant new coding and modulation schemes are introduced in order to improve the transmission quality in band-limited environments. The results show that these optimisations significantly improve the system performance. Subsequently, modulation and coded modulation transmission optimisation schemes for PLC overall channel characteristics system parameters will be a hot topic for future work.
With the drastic increase in the number of users, the existing wireless resources have become unsustainable. Therefore, the reactivation of Power Line Communication (PLC) has attracted the attention of major research institutes and industries. The development of PLC has been slow due to the complexity of the channel environment and the complexity and high cost of existing processing solutions. The most extensive work has been done on impulse noise, and it is particularly important to achieve robustness of data transmission against impulse noise at low cost. Firstly, the mainstream noise in PLC environment and its classification are introduced in this paper, and then the Differential Chaos Shift Keying (DCSK) and M-ary DCSK (MDCSK) modulation techniques with low cost and low complexity are described. The characteristics of this system in PLCs are presented and analysed, as well as the advantages and improvements that exist for various types of impulse noise. Secondly, some relevant new coding and modulation schemes are introduced in order to improve the transmission quality in band-limited environments. The results show that these optimisations significantly improve the system performance. Subsequently, modulation and coded modulation transmission optimisation schemes for PLC overall channel characteristics system parameters will be a hot topic for future work.
2024, 46(5): 1774-1789.
doi: 10.11999/JEIT230703
Abstract:
With the application and popularization of IoT, cloud computing, and artificial intelligence, data security and privacy protection have become the focus of attention. Fully homomorphic encryption, as an effective solution to the privacy security problem, allows performing arbitrary homomorphic computation on encrypted data, and is a powerful encryption tool with a wide range of potential applications. The paper summarizes the proposed fully homomorphic encryption schemes since 2009, and divides them into four technical routes based on the core technologies of the schemes, analyzes and discusses the key constructs, algorithm optimization processes, and future development directions of each type of scheme. The paper firstly introduces fully homomorphic encryption-related mathematical principles, covering the basic assumptions and security features of fully homomorphic encryption schemes. Subsequently, according to the technical routes of the four fully homomorphic encryption schemes, it summarizes the structural general formulas of the encryption schemes, summarizes the core steps of the bootstrap algorithms, discusses the latest research progress, and on the basis of this, comprehensively analyzes and compares the storage efficiencies and computing speeds of various schemes. The paper finally shows the application implementation of homomorphic algorithm library for encryption schemes under each technical route, analyzes the opportunities and challenges of fully homomorphic encryption schemes in the current era, and makes an outlook on the future research prospects.
With the application and popularization of IoT, cloud computing, and artificial intelligence, data security and privacy protection have become the focus of attention. Fully homomorphic encryption, as an effective solution to the privacy security problem, allows performing arbitrary homomorphic computation on encrypted data, and is a powerful encryption tool with a wide range of potential applications. The paper summarizes the proposed fully homomorphic encryption schemes since 2009, and divides them into four technical routes based on the core technologies of the schemes, analyzes and discusses the key constructs, algorithm optimization processes, and future development directions of each type of scheme. The paper firstly introduces fully homomorphic encryption-related mathematical principles, covering the basic assumptions and security features of fully homomorphic encryption schemes. Subsequently, according to the technical routes of the four fully homomorphic encryption schemes, it summarizes the structural general formulas of the encryption schemes, summarizes the core steps of the bootstrap algorithms, discusses the latest research progress, and on the basis of this, comprehensively analyzes and compares the storage efficiencies and computing speeds of various schemes. The paper finally shows the application implementation of homomorphic algorithm library for encryption schemes under each technical route, analyzes the opportunities and challenges of fully homomorphic encryption schemes in the current era, and makes an outlook on the future research prospects.
2024, 46(5): 1790-1805.
doi: 10.11999/JEIT230448
Abstract:
Fully Homomorphic Encryption (FHE) is a multi-party secure computation protocol characterized by its high computational complexity and low interaction requirements. Although there is no need for multiple rounds of interactions and extensive communications between computing participants in protocols based on FHE, the processing time of encrypted data is typically\begin{document}$ {10}^{3} $\end{document} \begin{document}$ {10}^{6} $\end{document}
Fully Homomorphic Encryption (FHE) is a multi-party secure computation protocol characterized by its high computational complexity and low interaction requirements. Although there is no need for multiple rounds of interactions and extensive communications between computing participants in protocols based on FHE, the processing time of encrypted data is typically
2024, 46(5): 1806-1826.
doi: 10.11999/JEIT231249
Abstract:
Radar possesses the capability for all-day, all-weather observation and can generate radar target images through image processing. It serves as an indispensable piece of remote sensing equipment in various civil and military applications, including earth observation, and surveillance. High-resolution radar images can provide a detailed outline and fine structure of the target, which is conducive to subsequent applications such as target classification and recognition. For the acquired radar images, how to use theoretical methods such as signal and information processing to further improve the resolution and break through the Rayleigh limit has important scientific research and practical application value. On the other hand, polarization, a crucial attribute of electromagnetic waves, plays a significant role in the acquisition and analysis of target characteristics, and can provide rich information for super-resolution reconstruction. Accordingly, this work initially elucidates the concept of polarimetric radar image super-resolution reconstruction, summarizes the performance evaluation metrics, and primarily focuses on the methods of polarimetric radar image super-resolution reconstruction and their applications. Lastly, the limitations of existing methods are summarized and potential future trends in technology are forecasted.
Radar possesses the capability for all-day, all-weather observation and can generate radar target images through image processing. It serves as an indispensable piece of remote sensing equipment in various civil and military applications, including earth observation, and surveillance. High-resolution radar images can provide a detailed outline and fine structure of the target, which is conducive to subsequent applications such as target classification and recognition. For the acquired radar images, how to use theoretical methods such as signal and information processing to further improve the resolution and break through the Rayleigh limit has important scientific research and practical application value. On the other hand, polarization, a crucial attribute of electromagnetic waves, plays a significant role in the acquisition and analysis of target characteristics, and can provide rich information for super-resolution reconstruction. Accordingly, this work initially elucidates the concept of polarimetric radar image super-resolution reconstruction, summarizes the performance evaluation metrics, and primarily focuses on the methods of polarimetric radar image super-resolution reconstruction and their applications. Lastly, the limitations of existing methods are summarized and potential future trends in technology are forecasted.
2024, 46(5): 1827-1842.
doi: 10.11999/JEIT231133
Abstract:
In the future communication network, the sixth generation mobile communication system technology(6G), which is widely expected, will face many challenges, including the issue of ultra-reliable communication in high-speed mobile scenarios. Orthogonal Time Frequency Space (OTFS) modulation technology overcomes the multi-path and Doppler effects of traditional communication systems in high-speed mobile environments, and provides a new possibility for realizing 6G ultra-reliable communication. This paper first introduces the basic principle, mathematical model, interference and advantage analysis of OTFS. Then, the research status of OTFS technology in synchronization, channel estimation and signal detection is summarized and analyzed. Subsequently, the application trend of OTFS is analyzed from four typical application scenarios of vehicle networking, unmanned aerial vehicle, satellite communication and marine communication. Finally, the difficulties and challenges to be overcome in future OTFS research are discussed from four aspects: reducing multi-dimensional matching filter, phase demodulation and channel estimation, hardware implementation complexity and improving the high utilization of time-frequency resources.
In the future communication network, the sixth generation mobile communication system technology(6G), which is widely expected, will face many challenges, including the issue of ultra-reliable communication in high-speed mobile scenarios. Orthogonal Time Frequency Space (OTFS) modulation technology overcomes the multi-path and Doppler effects of traditional communication systems in high-speed mobile environments, and provides a new possibility for realizing 6G ultra-reliable communication. This paper first introduces the basic principle, mathematical model, interference and advantage analysis of OTFS. Then, the research status of OTFS technology in synchronization, channel estimation and signal detection is summarized and analyzed. Subsequently, the application trend of OTFS is analyzed from four typical application scenarios of vehicle networking, unmanned aerial vehicle, satellite communication and marine communication. Finally, the difficulties and challenges to be overcome in future OTFS research are discussed from four aspects: reducing multi-dimensional matching filter, phase demodulation and channel estimation, hardware implementation complexity and improving the high utilization of time-frequency resources.
2024, 46(5): 1843-1852.
doi: 10.11999/JEIT231245
Abstract:
In the field of signal processing, approximate computing techniques have garnered significant attention. Complex algorithms and massive data impose limitations on processing speed and increase system hardware consumption. Since signals often contain redundancy, precise results are not always necessary, and achieving results acceptable to users is sufficient. Therefore, employing approximate computing techniques can effectively reduce computational complexity, enhance computational efficiency, and improve system performance. This paper takes a hierarchical approach to the design of approximate computing techniques. It first introduces the characteristics of signal processing applications, reviews recent research progress in approximate computing techniques at the algorithm and circuit levels, and investigates approximate computing solutions in signal processing directions such as communication, video imaging, and radar. Finally, it discusses and prospects the development direction of this field, providing insights to promote the application of approximate computing techniques in signal processing.
In the field of signal processing, approximate computing techniques have garnered significant attention. Complex algorithms and massive data impose limitations on processing speed and increase system hardware consumption. Since signals often contain redundancy, precise results are not always necessary, and achieving results acceptable to users is sufficient. Therefore, employing approximate computing techniques can effectively reduce computational complexity, enhance computational efficiency, and improve system performance. This paper takes a hierarchical approach to the design of approximate computing techniques. It first introduces the characteristics of signal processing applications, reviews recent research progress in approximate computing techniques at the algorithm and circuit levels, and investigates approximate computing solutions in signal processing directions such as communication, video imaging, and radar. Finally, it discusses and prospects the development direction of this field, providing insights to promote the application of approximate computing techniques in signal processing.
2024, 46(5): 1853-1864.
doi: 10.11999/JEIT231365
Abstract:
Understanding the species composition, abundance and temporal and spatial distribution of fish on a global scale will help their biodiversity conservation. Underwater image acquisition is one of the main means to survey fish species diversity, but image data analysis is time-consuming and labor-intensive. Since 2015, a series of progress has been made in updating the datasets of marine fish images and optimizing the algorithm of deep learning models, but the performance of fine-grained classification is still insufficient, and the production practice application of research results is relatively weak. Therefore, the need for automated fish image classification in marine investigations is firstly studied. Then a comprehensive introduction to fish image datasets and deep learning algorithm applications is provided, and the main challenges and the corresponding solutions are analyzed. Finally, the importance of automated classification of marine fish images for related image information processing research is discussed, and its prospects in the field of marine monitoring are summarized.
Understanding the species composition, abundance and temporal and spatial distribution of fish on a global scale will help their biodiversity conservation. Underwater image acquisition is one of the main means to survey fish species diversity, but image data analysis is time-consuming and labor-intensive. Since 2015, a series of progress has been made in updating the datasets of marine fish images and optimizing the algorithm of deep learning models, but the performance of fine-grained classification is still insufficient, and the production practice application of research results is relatively weak. Therefore, the need for automated fish image classification in marine investigations is firstly studied. Then a comprehensive introduction to fish image datasets and deep learning algorithm applications is provided, and the main challenges and the corresponding solutions are analyzed. Finally, the importance of automated classification of marine fish images for related image information processing research is discussed, and its prospects in the field of marine monitoring are summarized.
2024, 46(5): 1865-1873.
doi: 10.11999/JEIT231211
Abstract:
The recent researches on vortex beam multimode integration and dynamic regulation are summarized in this article. Starting from the passive metasurface lens, utilizing the comprehensive control effect of propagation phase and geometric phase, fractional modes of vortex beam with high-purity is realized on a single metasurface platform. Furthermore, based on the theory and method about multimode vortex beam integration by using passive metasurfaces, active tunable electromagnetic devices such as varactors are loaded into the metasurface unit cell to adopt dynamic switching and manual regulation of vortex beams. On this basis, a theoretical analysis and evaluation of the performance of vortex communication systems based on metasurface is conducted through channel modeling, laying a theoretical foundation for improving the channel capacity and information transmission rate of modern communication systems.
The recent researches on vortex beam multimode integration and dynamic regulation are summarized in this article. Starting from the passive metasurface lens, utilizing the comprehensive control effect of propagation phase and geometric phase, fractional modes of vortex beam with high-purity is realized on a single metasurface platform. Furthermore, based on the theory and method about multimode vortex beam integration by using passive metasurfaces, active tunable electromagnetic devices such as varactors are loaded into the metasurface unit cell to adopt dynamic switching and manual regulation of vortex beams. On this basis, a theoretical analysis and evaluation of the performance of vortex communication systems based on metasurface is conducted through channel modeling, laying a theoretical foundation for improving the channel capacity and information transmission rate of modern communication systems.
2024, 46(5): 1874-1887.
doi: 10.11999/JEIT231208
Abstract:
Sigma-Delta (Σ-Δ) Analog-to-Digital Converter (ADC) is based on oversampling and noise shaping techniques to achieve high-resolution, and is characterized by low passive component matching requirements and simple structure. In high-resolution audio applications, Σ-Δ ADC has gained widespread attention and applications since it can achieve high dynamic range with good power efficiency. Recently, there has been a growing research trend in designing low-power, high-resolution audio ADCs using advanced processes and technologies. However, with process technology going to lower nodes and the reduction of supply voltages, the circuit design becomes more challenging. This paper reviews the state-of-the-art of the discrete-time and continuous-time design of high-resolution audio Sigma-Delta modulators, provides theoretical background for the design of high-resolution audio Sigma-Delta modulators, and gives research prospects.
Sigma-Delta (Σ-Δ) Analog-to-Digital Converter (ADC) is based on oversampling and noise shaping techniques to achieve high-resolution, and is characterized by low passive component matching requirements and simple structure. In high-resolution audio applications, Σ-Δ ADC has gained widespread attention and applications since it can achieve high dynamic range with good power efficiency. Recently, there has been a growing research trend in designing low-power, high-resolution audio ADCs using advanced processes and technologies. However, with process technology going to lower nodes and the reduction of supply voltages, the circuit design becomes more challenging. This paper reviews the state-of-the-art of the discrete-time and continuous-time design of high-resolution audio Sigma-Delta modulators, provides theoretical background for the design of high-resolution audio Sigma-Delta modulators, and gives research prospects.
2024, 46(5): 1888-1895.
doi: 10.11999/JEIT231216
Abstract:
SC DC-DC converters have wide applications. Previous works proposed multiple topologies for a high voltage conversion ratio scenario, such as Dickson, Cockcroft-Walton, Series-Parallel, Ladder, Fibonacci and Divider. They have their own features, fitting different applications. However, it is unclear how these topologies are generated, what the main differences are between them, and what the advantages and disadvantages are. Therefore, this paper starts from the transformer model of the SC converters, analyzing the difference among them. Then, it is demonstrated how the SC converters evolve from the model to the real circuitries, where multiple intuitive understanding can be obtained.
SC DC-DC converters have wide applications. Previous works proposed multiple topologies for a high voltage conversion ratio scenario, such as Dickson, Cockcroft-Walton, Series-Parallel, Ladder, Fibonacci and Divider. They have their own features, fitting different applications. However, it is unclear how these topologies are generated, what the main differences are between them, and what the advantages and disadvantages are. Therefore, this paper starts from the transformer model of the SC converters, analyzing the difference among them. Then, it is demonstrated how the SC converters evolve from the model to the real circuitries, where multiple intuitive understanding can be obtained.
2024, 46(5): 1896-1907.
doi: 10.11999/JEIT231337
Abstract:
In Unmanned Aerial Vehicle (UAV) relay networks, communication resource allocation and motion planning of UAV are the key problems that should be solved. In order to improve the communication efficiency of UAV relay communication system, a joint planning method of UAV relay power allocation and trajectory design is proposed based on proximal policy optimization algorithm. The joint planning problem of UAV relay power allocation and trajectory design in the user movement scenario is modelled as a Markov decision-making process. Considering the inaccurate acquisition of user location information, the reward function is set with the maximum throughput of the relay communication system as the optimization goal under the premise of satisfying the user interruption probability constraint. Then, a deep reinforcement learning algorithm with high convergence speed—the Proximal Policy Optimization (PPO) algorithm, is used to solve the problem and realized the flight trajectory optimization of relay UAV and the reasonable and effective allocation of relay transmission power. The simulation experimental results show that for the scenario of UAV relay communication with random users movement, the proposed method improves system throughput by 22% and 15%, respectively, compared to the methods based on random strategy and traditional Deep Deterministic Policy Gradient (DDPG). The results show that the proposed method can effectively improve the communication efficiency of the system.
In Unmanned Aerial Vehicle (UAV) relay networks, communication resource allocation and motion planning of UAV are the key problems that should be solved. In order to improve the communication efficiency of UAV relay communication system, a joint planning method of UAV relay power allocation and trajectory design is proposed based on proximal policy optimization algorithm. The joint planning problem of UAV relay power allocation and trajectory design in the user movement scenario is modelled as a Markov decision-making process. Considering the inaccurate acquisition of user location information, the reward function is set with the maximum throughput of the relay communication system as the optimization goal under the premise of satisfying the user interruption probability constraint. Then, a deep reinforcement learning algorithm with high convergence speed—the Proximal Policy Optimization (PPO) algorithm, is used to solve the problem and realized the flight trajectory optimization of relay UAV and the reasonable and effective allocation of relay transmission power. The simulation experimental results show that for the scenario of UAV relay communication with random users movement, the proposed method improves system throughput by 22% and 15%, respectively, compared to the methods based on random strategy and traditional Deep Deterministic Policy Gradient (DDPG). The results show that the proposed method can effectively improve the communication efficiency of the system.
2024, 46(5): 1908-1919.
doi: 10.11999/JEIT231318
Abstract:
The electromagnetic security cooperative suppression technology in the security area of major events under the complex urban environment is presented in this paper. Firstly, the complex urban electromagnetic environment is modeled by using the radio wave propagation model suitable for dense urban environment. Secondly, aiming at the problem of efficient electromagnetic suppression and effective avoidance of harmful interference, the potential game method is used to design the cooperative deployment algorithm of electromagnetic suppression equipment. Building upon this, the power optimization method of suppression equipment based on genetic algorithm is proposed to achieve the efficient delivery of interference power under the cooperative work of electromagnetic suppression equipment. The simulation results indicate that the proposed electromagnetic suppression equipment deployment algorithm can obtain outstanding performance similar to the theoretical optimal method (i.e., traversal algorithm), with lower computational complexity. Moreover, under identical interference effectiveness, the proposed power optimization algorithm reduces transmission power by over 50% compared to the traditional interference power allocation methods, thereby achieving precise collaborative control.
The electromagnetic security cooperative suppression technology in the security area of major events under the complex urban environment is presented in this paper. Firstly, the complex urban electromagnetic environment is modeled by using the radio wave propagation model suitable for dense urban environment. Secondly, aiming at the problem of efficient electromagnetic suppression and effective avoidance of harmful interference, the potential game method is used to design the cooperative deployment algorithm of electromagnetic suppression equipment. Building upon this, the power optimization method of suppression equipment based on genetic algorithm is proposed to achieve the efficient delivery of interference power under the cooperative work of electromagnetic suppression equipment. The simulation results indicate that the proposed electromagnetic suppression equipment deployment algorithm can obtain outstanding performance similar to the theoretical optimal method (i.e., traversal algorithm), with lower computational complexity. Moreover, under identical interference effectiveness, the proposed power optimization algorithm reduces transmission power by over 50% compared to the traditional interference power allocation methods, thereby achieving precise collaborative control.
2024, 46(5): 1920-1930.
doi: 10.11999/JEIT231192
Abstract:
The Aerial-Ground Integrated Networks (AGIN) take full advantage of the flexible deployment of Aerial Base Stations (ABSs) to provide on-demand coverage and high-quality services in hotspot areas. However, the high dynamics of ABSs pose a great challenge to service continuity assurance in AGIN. Furthermore, given the energy constraints of ABSs, ensuring service continuity with low power consumption becomes an increasingly formidable challenge. This is attributed to the inherent contradiction between enhancing service continuity and reducing power consumption, which typically necessitates distinct flight actions. Focusing on the problem mentioned above, a communication and control joint optimization approach based on Federated Deep Reinforcement Learning (FDRL) is proposed to obtain low-power service continuity assurance in AGIN. The proposed approach ensures service continuity by jointly optimizing the flight actions of ABSs, user associations, and power allocation. To cope with the high dynamics of ABSs, an environmental state experience pool is designed to capture the spatiotemporal correlation of channels, and the rate variance is introduced into the reward function to ensure service continuity. Taking into account the power consumption differences associated with various flight actions, the proposed approach optimizes the flight actions of ABSs to reduce their power consumption. Simulation results demonstrate that, under the premise of satisfying requirements for user rate and rate variance, the proposed approach can effectively reduce network power consumption. Additionally, the performance of FDRL is close to that of centralized reinforcement learning.
The Aerial-Ground Integrated Networks (AGIN) take full advantage of the flexible deployment of Aerial Base Stations (ABSs) to provide on-demand coverage and high-quality services in hotspot areas. However, the high dynamics of ABSs pose a great challenge to service continuity assurance in AGIN. Furthermore, given the energy constraints of ABSs, ensuring service continuity with low power consumption becomes an increasingly formidable challenge. This is attributed to the inherent contradiction between enhancing service continuity and reducing power consumption, which typically necessitates distinct flight actions. Focusing on the problem mentioned above, a communication and control joint optimization approach based on Federated Deep Reinforcement Learning (FDRL) is proposed to obtain low-power service continuity assurance in AGIN. The proposed approach ensures service continuity by jointly optimizing the flight actions of ABSs, user associations, and power allocation. To cope with the high dynamics of ABSs, an environmental state experience pool is designed to capture the spatiotemporal correlation of channels, and the rate variance is introduced into the reward function to ensure service continuity. Taking into account the power consumption differences associated with various flight actions, the proposed approach optimizes the flight actions of ABSs to reduce their power consumption. Simulation results demonstrate that, under the premise of satisfying requirements for user rate and rate variance, the proposed approach can effectively reduce network power consumption. Additionally, the performance of FDRL is close to that of centralized reinforcement learning.
2024, 46(5): 1931-1939.
doi: 10.11999/JEIT230707
Abstract:
Aiming at the problems that the traditional UAV formation routing algorithm cannot effectively utilize the advance predictability of topology changes, and the high cost is caused by acquiring the link connection by sending detection packets, a UAV formation shortest delay routing protocol based on discrete time aggregation graph is proposed by introducing the time-varying graph model. Firstly, using the prior knowledge of the UAV formation network, such as the movement trajectory of nodes and the network topology changes, the network link resources and network topology are characterized by using the discrete time aggregation graph. Secondly, the routing decision algorithm is designed based on the graph model. The delay in the process of route discovery is used as the link weight to solve the shortest delay route from the source node to the destination node of the network. Finally, the simulation performance shows that the routing protocol improves the packet delivery rate, reduces the end-to-end delay and diminishes the network control overhead compared with the traditional Ad-hoc On-Demand Distance Vector routing protocol.
Aiming at the problems that the traditional UAV formation routing algorithm cannot effectively utilize the advance predictability of topology changes, and the high cost is caused by acquiring the link connection by sending detection packets, a UAV formation shortest delay routing protocol based on discrete time aggregation graph is proposed by introducing the time-varying graph model. Firstly, using the prior knowledge of the UAV formation network, such as the movement trajectory of nodes and the network topology changes, the network link resources and network topology are characterized by using the discrete time aggregation graph. Secondly, the routing decision algorithm is designed based on the graph model. The delay in the process of route discovery is used as the link weight to solve the shortest delay route from the source node to the destination node of the network. Finally, the simulation performance shows that the routing protocol improves the packet delivery rate, reduces the end-to-end delay and diminishes the network control overhead compared with the traditional Ad-hoc On-Demand Distance Vector routing protocol.
2024, 46(5): 1940-1950.
doi: 10.11999/JEIT231219
Abstract:
Holographic Multiple-Input Multiple-Output (HMIMO) is an emerging technology for 6G communications. This type of array is composed of densely distributed antenna elements within a fixed aperture area. It is an extension of Massive MIMO technology under the practical constraints of antenna aperture. HMIMO systems have great potential in significantly improving wireless communication performance. However, due to the presence of closely spaced antennas, and the distane between antennas is less than half of the length, severe coupling effects are inevitable and traditional assumption of independent and identically distributed channel is invalid. Thus, designing an effective and practical channel model becomes one of the most challenging problems in HMIMO researches. To address these challenges, this paper investigates four channel modeling approaches based on electromagnetic field theory. The first approach is based on the plane Green’s function and models the integral of Green’s functions between planes with high complexity. The second and third approaches approximate the communication channel in HMIMO using plane wave expansion and spherical wave expansion, respectively, with lower complexity. The channel modeling based on plane wave expansion is relatively simple and is more suitable for far field, but would underestimate the maximum capacity of the system under strong coupling between antennas. The channel modeling based on spherical wave expansion better captures the characteristics of the electromagnetic wave channel but comes with higher complexity. Finally, a channel modeling method based on random Green’s functions is introduced, primarily describing the random characteristics of electromagnetic waves in rich scattering environments or Rayleigh channels.
Holographic Multiple-Input Multiple-Output (HMIMO) is an emerging technology for 6G communications. This type of array is composed of densely distributed antenna elements within a fixed aperture area. It is an extension of Massive MIMO technology under the practical constraints of antenna aperture. HMIMO systems have great potential in significantly improving wireless communication performance. However, due to the presence of closely spaced antennas, and the distane between antennas is less than half of the length, severe coupling effects are inevitable and traditional assumption of independent and identically distributed channel is invalid. Thus, designing an effective and practical channel model becomes one of the most challenging problems in HMIMO researches. To address these challenges, this paper investigates four channel modeling approaches based on electromagnetic field theory. The first approach is based on the plane Green’s function and models the integral of Green’s functions between planes with high complexity. The second and third approaches approximate the communication channel in HMIMO using plane wave expansion and spherical wave expansion, respectively, with lower complexity. The channel modeling based on plane wave expansion is relatively simple and is more suitable for far field, but would underestimate the maximum capacity of the system under strong coupling between antennas. The channel modeling based on spherical wave expansion better captures the characteristics of the electromagnetic wave channel but comes with higher complexity. Finally, a channel modeling method based on random Green’s functions is introduced, primarily describing the random characteristics of electromagnetic waves in rich scattering environments or Rayleigh channels.
2024, 46(5): 1951-1957.
doi: 10.11999/JEIT231036
Abstract:
The multi-functional integrated platform with simultaneous transmit and receive capability faces the strong Self-Interference (SI) coupled between the adjacent transmit and receive arrays. In this paper, a wideband SI cancellation method in the space domain for fully digital phased array systems is designed. A non-convex optimization problem is formulated to minimize the residual SI and noise power while constraining the loss of beamforming gain in the desired direction, and an alternate optimization method is proposed to jointly determine the transmit and receive beamforming weights, and the SI cancellation performance of the proposed algorithm is analyzed. Theoretical analysis and simulation results show that a 60-element array can achieve an Effective Isotropic Isolation (EII) of 168 dB when the central frequency is 2.4 GHz, the bandwidth is 100 MHz, and the beamforming gain loss is limited to 3 dB, which is 7 dB below the EII upper bound.
The multi-functional integrated platform with simultaneous transmit and receive capability faces the strong Self-Interference (SI) coupled between the adjacent transmit and receive arrays. In this paper, a wideband SI cancellation method in the space domain for fully digital phased array systems is designed. A non-convex optimization problem is formulated to minimize the residual SI and noise power while constraining the loss of beamforming gain in the desired direction, and an alternate optimization method is proposed to jointly determine the transmit and receive beamforming weights, and the SI cancellation performance of the proposed algorithm is analyzed. Theoretical analysis and simulation results show that a 60-element array can achieve an Effective Isotropic Isolation (EII) of 168 dB when the central frequency is 2.4 GHz, the bandwidth is 100 MHz, and the beamforming gain loss is limited to 3 dB, which is 7 dB below the EII upper bound.
2024, 46(5): 1958-1967.
doi: 10.11999/JEIT231199
Abstract:
Unmanned Aerial Vehicles (UAV) technology holds significant importance and offers extensive potential for application in the field of inspection. Taking into account the limited endurance of the UAV, it needs to fly from the nest to the designated inspection area, complete the inspection of the transmission tower, and then return to the nest safely before the battery is exhausted. For large-scale inspection scenarios, a multi-UAV inspection method is proposed to minimize the inspection time. In detail, the k-means++ algorithm is used to optimize task allocation of the UAVs and the modified simulated annealing algorithm is utilized to optimize the inspection trajectory to improve the inspection efficiency. Finally, based on the tower pole distribution data from a simulated real-world environment, the proposed algorithm is employed to assign tasks of the UAVs and design trajectories. The simulation results confirm that the proposed algorithm can significantly reduce the total inspection time through multi-UAV task allocation and trajectory design.
Unmanned Aerial Vehicles (UAV) technology holds significant importance and offers extensive potential for application in the field of inspection. Taking into account the limited endurance of the UAV, it needs to fly from the nest to the designated inspection area, complete the inspection of the transmission tower, and then return to the nest safely before the battery is exhausted. For large-scale inspection scenarios, a multi-UAV inspection method is proposed to minimize the inspection time. In detail, the k-means++ algorithm is used to optimize task allocation of the UAVs and the modified simulated annealing algorithm is utilized to optimize the inspection trajectory to improve the inspection efficiency. Finally, based on the tower pole distribution data from a simulated real-world environment, the proposed algorithm is employed to assign tasks of the UAVs and design trajectories. The simulation results confirm that the proposed algorithm can significantly reduce the total inspection time through multi-UAV task allocation and trajectory design.
2024, 46(5): 1968-1976.
doi: 10.11999/JEIT231086
Abstract:
Space-air-ground integrated network has the advantages of extensive coverage, high throughput, and strong elasticity. A resource allocation algorithm for dense scenarios is proposed to solve the problems of network congestion and deterioration of service quality caused by concurrent access of many users and network load imbalance. Firstly, the user utility function is constructed based on the user demand and the preferences of different types of user tasks. Then, load balancing is realized based on the matching game network selection algorithm and the power control algorithm combined with the dual ascending method, and the resource allocation scheme is optimized. Experimental results show that compared with the traditional strategy, the proposed strategy increases the overall user access rate by at least 35%, and improves the performance of delay and throughput by more than 50%. Load balancing is more effective in dense scenarios and network performance is improved.
Space-air-ground integrated network has the advantages of extensive coverage, high throughput, and strong elasticity. A resource allocation algorithm for dense scenarios is proposed to solve the problems of network congestion and deterioration of service quality caused by concurrent access of many users and network load imbalance. Firstly, the user utility function is constructed based on the user demand and the preferences of different types of user tasks. Then, load balancing is realized based on the matching game network selection algorithm and the power control algorithm combined with the dual ascending method, and the resource allocation scheme is optimized. Experimental results show that compared with the traditional strategy, the proposed strategy increases the overall user access rate by at least 35%, and improves the performance of delay and throughput by more than 50%. Load balancing is more effective in dense scenarios and network performance is improved.
2024, 46(5): 1977-1985.
doi: 10.11999/JEIT231115
Abstract:
The short packet covert communication using a shared-aperture jammer assisted Time-Modulated Array (TMA) is investigated for the first time in this paper. Firstly, a TMA architecture for shared-aperture jammer is proposed and an optimization method is introduced that maximizes the gain of the target direction while forming interference in non-target directions. Based on this model, closed-form expressions for the covertness constraint and covert throughput are derived. Furthermore, the transmission power and blocklength are optimized to maximize the covert throughput. Simulation results show that there exists an optimum blocklength that maximizes the covert throughput, and the proposed scheme outperforms the benchmark scheme in terms of covert communication performance.
The short packet covert communication using a shared-aperture jammer assisted Time-Modulated Array (TMA) is investigated for the first time in this paper. Firstly, a TMA architecture for shared-aperture jammer is proposed and an optimization method is introduced that maximizes the gain of the target direction while forming interference in non-target directions. Based on this model, closed-form expressions for the covertness constraint and covert throughput are derived. Furthermore, the transmission power and blocklength are optimized to maximize the covert throughput. Simulation results show that there exists an optimum blocklength that maximizes the covert throughput, and the proposed scheme outperforms the benchmark scheme in terms of covert communication performance.
2024, 46(5): 1986-1995.
doi: 10.11999/JEIT231169
Abstract:
In order to solve the problems of short-distance communication, lower system throughput and the effects of channel uncertainties in traditional Backscatter Communication (BackCom) systems, a robust resource allocation algorithm for a Reconfigurable Intelligent Surface (RIS)-assisted backscatter communication system with statistical Channel State Information (CSI) is proposed in this paper. A system weighting and sum throughput-maximization robust resource allocation model is formulated by considering the maximum transmit power constraint of the power station, the energy outage constraint and throughput outage constraint of backscatter nodes, the reflection coefficient constraint, the phase shift constraint of the RIS and the information transmission time constraint; Then, the original non-convex problem is transformed into a convex optimization problem by using the methods of Bernstein-type inequality, the alternating optimization, and the semi-definite relaxation technique. An iteration-based robust throughput maximization algorithm is designed. Simulation results show that the proposed algorithm had stronger robustness and higher throughput compared it with the traditional non-robust resource allocation algorithm and the resource allocation algorithm without RIS.
In order to solve the problems of short-distance communication, lower system throughput and the effects of channel uncertainties in traditional Backscatter Communication (BackCom) systems, a robust resource allocation algorithm for a Reconfigurable Intelligent Surface (RIS)-assisted backscatter communication system with statistical Channel State Information (CSI) is proposed in this paper. A system weighting and sum throughput-maximization robust resource allocation model is formulated by considering the maximum transmit power constraint of the power station, the energy outage constraint and throughput outage constraint of backscatter nodes, the reflection coefficient constraint, the phase shift constraint of the RIS and the information transmission time constraint; Then, the original non-convex problem is transformed into a convex optimization problem by using the methods of Bernstein-type inequality, the alternating optimization, and the semi-definite relaxation technique. An iteration-based robust throughput maximization algorithm is designed. Simulation results show that the proposed algorithm had stronger robustness and higher throughput compared it with the traditional non-robust resource allocation algorithm and the resource allocation algorithm without RIS.
2024, 46(5): 1996-2003.
doi: 10.11999/JEIT231175
Abstract:
In many monitoring and control tasks, it is difficult for the control center to get the real-time status information directly because of the distance between the monitored target and the control center. The Unmanned Aerial Vehicles (UAV) can make full use of its advantages of high mobility, reduce the sensing and communication distance, and then improve the sensing and communication capabilities, which provides a new idea for real-time acquisition of remote target status information. In this paper, the optimization problem of Age of Information (AoI) analysis in UAV-assisted integrated sensing and communication system is studied. Firstly, the status update process of control center is analyzed, and then the closed-form expression of average peak AoI is derived. Further, in the multi-UAV multi-target scenario, the average peak AoI of the system is further reduced by optimizing the perception position and communication position of the UAV in the air, as well as the matching relationship between the UAV and the target, and the real-time status update is improved. The simulation results verify the correctness of the theoretical analysis, and show that the proposed optimization method can effectively improve the AoI performance of the system compared with the benchmark methods.
In many monitoring and control tasks, it is difficult for the control center to get the real-time status information directly because of the distance between the monitored target and the control center. The Unmanned Aerial Vehicles (UAV) can make full use of its advantages of high mobility, reduce the sensing and communication distance, and then improve the sensing and communication capabilities, which provides a new idea for real-time acquisition of remote target status information. In this paper, the optimization problem of Age of Information (AoI) analysis in UAV-assisted integrated sensing and communication system is studied. Firstly, the status update process of control center is analyzed, and then the closed-form expression of average peak AoI is derived. Further, in the multi-UAV multi-target scenario, the average peak AoI of the system is further reduced by optimizing the perception position and communication position of the UAV in the air, as well as the matching relationship between the UAV and the target, and the real-time status update is improved. The simulation results verify the correctness of the theoretical analysis, and show that the proposed optimization method can effectively improve the AoI performance of the system compared with the benchmark methods.
2024, 46(5): 2004-2010.
doi: 10.11999/JEIT231196
Abstract:
To better solve the fading and severe inter-symbol interference problems in underwater acoustic channels, a Joint Multi-branch Merging and Equalization algorithm based on Deep Learning (JMME-DL) is proposed in this paper. The algorithm jointly implements multi-branch merging and equalization with the help of the nonlinear fitting ability of the deep learning network. The merging and equalization are not independent of each other, in the implementation of the algorithm, the total error is first calculated based on the total output of the deep learning network, and then the network parameters of each part are jointly adjusted with the total error, and the dataset is constructed based on the statistical underwater acoustic channel model. Simulation results show that the proposed algorithm achieves faster convergence speed and better BER performance compared to the existing algorithms, making it better adapted to underwater acoustic channels.
To better solve the fading and severe inter-symbol interference problems in underwater acoustic channels, a Joint Multi-branch Merging and Equalization algorithm based on Deep Learning (JMME-DL) is proposed in this paper. The algorithm jointly implements multi-branch merging and equalization with the help of the nonlinear fitting ability of the deep learning network. The merging and equalization are not independent of each other, in the implementation of the algorithm, the total error is first calculated based on the total output of the deep learning network, and then the network parameters of each part are jointly adjusted with the total error, and the dataset is constructed based on the statistical underwater acoustic channel model. Simulation results show that the proposed algorithm achieves faster convergence speed and better BER performance compared to the existing algorithms, making it better adapted to underwater acoustic channels.
2024, 46(5): 2011-2017.
doi: 10.11999/JEIT231226
Abstract:
Sparse Code Multiple Access (SCMA) technology is a highly valued code domain-based Non-Orthogonal Multiple Access (NOMA) technology. In order to solve the problem that the existing SCMA codebook design fails to combine the properties of data and decoder and the high complexity of MPA, a compressed sensing-assisted low-complexity SCMA system optimization design scheme is proposed. First, a codebook self-updating method is designed based on the system bit error rate optimization goal, which uses the gradient descent method to achieve self-updating of the codebook during the sparse vector reconstruction training process. Second, a compressed sensing-assisted multi-user detection algorithm is designed: Sign Decision Orthogonal Matching Pursuit (SD-OMP) algorithm. By sparse processing of the transmitted signal at the transmitting end, the compressed sensing technology is used at the receiving end to efficiently detect and reconstruct multi-user sparse signals, this results in a reduction of conflicts between users and a reduction in system complexity. The simulation results show that under Gaussian channel conditions, the compressed sensing-assisted low-complexity SCMA system optimization and design scheme can effectively reduce the complexity of multi-user detection, and can show better bit error rate performance when the system user part is active.
Sparse Code Multiple Access (SCMA) technology is a highly valued code domain-based Non-Orthogonal Multiple Access (NOMA) technology. In order to solve the problem that the existing SCMA codebook design fails to combine the properties of data and decoder and the high complexity of MPA, a compressed sensing-assisted low-complexity SCMA system optimization design scheme is proposed. First, a codebook self-updating method is designed based on the system bit error rate optimization goal, which uses the gradient descent method to achieve self-updating of the codebook during the sparse vector reconstruction training process. Second, a compressed sensing-assisted multi-user detection algorithm is designed: Sign Decision Orthogonal Matching Pursuit (SD-OMP) algorithm. By sparse processing of the transmitted signal at the transmitting end, the compressed sensing technology is used at the receiving end to efficiently detect and reconstruct multi-user sparse signals, this results in a reduction of conflicts between users and a reduction in system complexity. The simulation results show that under Gaussian channel conditions, the compressed sensing-assisted low-complexity SCMA system optimization and design scheme can effectively reduce the complexity of multi-user detection, and can show better bit error rate performance when the system user part is active.
2024, 46(5): 2018-2027.
doi: 10.11999/JEIT230958
Abstract:
Aiming at the problems of high false alarm and sensitivity to channel error of Robust Principal Component Analysis (RPCA) algorithm in multi-channel Ground Moving Target Indication (GMTI), this paper proposes a data reconstruction and Velocity Synthetic Aperture Radar (VSAR)-RPCA joint processing method. Firstly, the sample selection and joint pixel method are used to complete the accurate reconstruction of inter-channel data; then a new RPCA optimization model is proposed by combining the VSAR detection mode, and the sparse matrix in the spatial frequency domain is obtained by solving the new RPCA optimization model with the alternating projection multiplier method, and then the differences in the distribution characteristics of the moving target and the strong clutter residuals in the spatial frequency domain channel are used to realize the strong clutter residuals rejection and the detection of the moving target; finally, the radial velocity of the target is estimated by the Along-Track Interferometry algorithm to complete the moving target relocation. Compared with the traditional RPCA algorithm, the proposed algorithm significantly reduces the false alarm rate under the background of non-ideal strong clutter. Theoretical analyses and experiments verify the effectiveness of the proposed algorithm.
Aiming at the problems of high false alarm and sensitivity to channel error of Robust Principal Component Analysis (RPCA) algorithm in multi-channel Ground Moving Target Indication (GMTI), this paper proposes a data reconstruction and Velocity Synthetic Aperture Radar (VSAR)-RPCA joint processing method. Firstly, the sample selection and joint pixel method are used to complete the accurate reconstruction of inter-channel data; then a new RPCA optimization model is proposed by combining the VSAR detection mode, and the sparse matrix in the spatial frequency domain is obtained by solving the new RPCA optimization model with the alternating projection multiplier method, and then the differences in the distribution characteristics of the moving target and the strong clutter residuals in the spatial frequency domain channel are used to realize the strong clutter residuals rejection and the detection of the moving target; finally, the radial velocity of the target is estimated by the Along-Track Interferometry algorithm to complete the moving target relocation. Compared with the traditional RPCA algorithm, the proposed algorithm significantly reduces the false alarm rate under the background of non-ideal strong clutter. Theoretical analyses and experiments verify the effectiveness of the proposed algorithm.
2024, 46(5): 2028-2035.
doi: 10.11999/JEIT231137
Abstract:
With the rapid developments of positioning systems, high-precision and low-cost direction-finding technologies are urgently needed. The hardware complexity and economic cost of traditional direction-finding methods have hindered their wide applications. Recently, direction finding based on Time-Modulated Arrays (TMAs) has overcome the shortcomings of traditional direction-finding methods. Nevertheless, to ensure measurement accuracy, one has to keep an adequate number of array elements in common TMAs. Consequently, a question arises, i.e., is it possible to reduce the number of array elements in TMAs, thus making the hardware complexity as low as possible? A novel direction-finding method based on the TMA with a single radio frequency channel and composite baselines is proposed in this paper. In the method, four antennas are meticulously arranged at specific intervals to form double-long baselines, and accurate and low-cost direction finding is realized with the ingenious usage of Field Programmable Gate Array (FPGA) and single receiving channel. To verify the effectiveness of the method, a prototype system in the S band is designed, fabricated, and measured. Detailed comparisons with the existing methods are provided. The work will benefit the development and application of high-precision and low-cost direction-finding systems.
With the rapid developments of positioning systems, high-precision and low-cost direction-finding technologies are urgently needed. The hardware complexity and economic cost of traditional direction-finding methods have hindered their wide applications. Recently, direction finding based on Time-Modulated Arrays (TMAs) has overcome the shortcomings of traditional direction-finding methods. Nevertheless, to ensure measurement accuracy, one has to keep an adequate number of array elements in common TMAs. Consequently, a question arises, i.e., is it possible to reduce the number of array elements in TMAs, thus making the hardware complexity as low as possible? A novel direction-finding method based on the TMA with a single radio frequency channel and composite baselines is proposed in this paper. In the method, four antennas are meticulously arranged at specific intervals to form double-long baselines, and accurate and low-cost direction finding is realized with the ingenious usage of Field Programmable Gate Array (FPGA) and single receiving channel. To verify the effectiveness of the method, a prototype system in the S band is designed, fabricated, and measured. Detailed comparisons with the existing methods are provided. The work will benefit the development and application of high-precision and low-cost direction-finding systems.
2024, 46(5): 2036-2047.
doi: 10.11999/JEIT231215
Abstract:
To estimate parameters of parameterized scattering center models is one of the basic methods for Synthetic Aperture Radar Advanced Information Retrieval (SAR AIR). Traditional Attributed Scattering Center (ASC) parameter estimation algorithms usually suffer from issues such as slow computation speed, high algorithm complexity, and high sensitivity to initial values of parameters. In this paper, a novel end-to-end framework for inverting ASC parameters from radar images based on unsupervised deep learning is proposed. Firstly, an autoencoder network structure is employed to effectively extract image features of targets, alleviating the difficulties solving directly caused by the complex non-convex optimization space and resolving the sensitivity to initial values. Secondly, the ASC model is embedded as a physical decoder to constrain the encoder output to correct ASC parameters. Finally, the end-to-end architecture are utlized to train and infer the model, achieving the purpose of reducing algorithm complexity and improving estimation speed. Through testing on simulated and measured data, experimental results indicate that the estimation error obtained on the SAR image test set with a resolution of 0.15 m is less than 0.1 m while the average processing time is 0.06 s for the inversion of one single scattering center, which demonstrate the effectiveness, efficiency, and robustness of the proposed approach.
To estimate parameters of parameterized scattering center models is one of the basic methods for Synthetic Aperture Radar Advanced Information Retrieval (SAR AIR). Traditional Attributed Scattering Center (ASC) parameter estimation algorithms usually suffer from issues such as slow computation speed, high algorithm complexity, and high sensitivity to initial values of parameters. In this paper, a novel end-to-end framework for inverting ASC parameters from radar images based on unsupervised deep learning is proposed. Firstly, an autoencoder network structure is employed to effectively extract image features of targets, alleviating the difficulties solving directly caused by the complex non-convex optimization space and resolving the sensitivity to initial values. Secondly, the ASC model is embedded as a physical decoder to constrain the encoder output to correct ASC parameters. Finally, the end-to-end architecture are utlized to train and infer the model, achieving the purpose of reducing algorithm complexity and improving estimation speed. Through testing on simulated and measured data, experimental results indicate that the estimation error obtained on the SAR image test set with a resolution of 0.15 m is less than 0.1 m while the average processing time is 0.06 s for the inversion of one single scattering center, which demonstrate the effectiveness, efficiency, and robustness of the proposed approach.
2024, 46(5): 2048-2056.
doi: 10.11999/JEIT231333
Abstract:
In modern radar technology, a key research area is the design of special waveforms to prevent non-cooperative electronic reconnaissance systems from intercepting and detecting radar signals. This paper focuses on reducing the power interception probability of electronic reconnaissance systems while maintaining the radiation energy. Specifically, waveform design techniques are explored for passive countermeasures, considering the time-frequency distribution of energy and the characteristics of Short-Time Fourier Transform (STFT) wideband digital reconnaissance receivers. To address this, a low-intercept model for STFT wideband digital reconnaissance receiver is established, and then the low-intercept problem is converted into a constant envelope sequence iterative optimization problem using a quadratic optimization model. To improve autocorrelation performance, an auxiliary scalar is employed to transform the optimization model into a quadratic form and a sequence of low-interception waveforms are generated through an iterative algorithm. Furthermore, the computational complexity of our proposed method is discussed. The simulation results, demonstrate that our sequence exhibits superior low-interception capability compared to commonly used phase-coded signals at the same receive SNR. Additionally, Pareto weights are introduced to control the autocorrelation characteristics of the proposed sequence, thereby enhancing the design flexibility.
In modern radar technology, a key research area is the design of special waveforms to prevent non-cooperative electronic reconnaissance systems from intercepting and detecting radar signals. This paper focuses on reducing the power interception probability of electronic reconnaissance systems while maintaining the radiation energy. Specifically, waveform design techniques are explored for passive countermeasures, considering the time-frequency distribution of energy and the characteristics of Short-Time Fourier Transform (STFT) wideband digital reconnaissance receivers. To address this, a low-intercept model for STFT wideband digital reconnaissance receiver is established, and then the low-intercept problem is converted into a constant envelope sequence iterative optimization problem using a quadratic optimization model. To improve autocorrelation performance, an auxiliary scalar is employed to transform the optimization model into a quadratic form and a sequence of low-interception waveforms are generated through an iterative algorithm. Furthermore, the computational complexity of our proposed method is discussed. The simulation results, demonstrate that our sequence exhibits superior low-interception capability compared to commonly used phase-coded signals at the same receive SNR. Additionally, Pareto weights are introduced to control the autocorrelation characteristics of the proposed sequence, thereby enhancing the design flexibility.
2024, 46(5): 2057-2064.
doi: 10.11999/JEIT231293
Abstract:
In the broad and unknown environments, mobile deployment of passive radar often faces challenges in promptly obtaining the precise location information of third-party transmitter stations. To address this issue, a transmitter localization method based on cooperative Unmanned Aerial Vehicle (UAV) path planning is proposed. Firstly, a single UAV is used as a cooperative target to establish the localization model and measurement equation in a two-dimensional scenario, and the Levenberg-Marquardt (LM) algorithm is employed for solution. Then, an optimization function is constructed by integrating Fisher information and control parameter constraints to dynamically plan the UAV trajectory, thereby improving the accuracy of transmitter localization and the practicality of this method. Finally, simulation experiments show that under the maximum control distance constraint, the positioning result of the proposed method is better than that of straight-line track and typical optimized track, and the final positioning accuracy is less than the standard deviation of the bistatic distance difference measurements, which can meet the application requirements of the passive radar system.
In the broad and unknown environments, mobile deployment of passive radar often faces challenges in promptly obtaining the precise location information of third-party transmitter stations. To address this issue, a transmitter localization method based on cooperative Unmanned Aerial Vehicle (UAV) path planning is proposed. Firstly, a single UAV is used as a cooperative target to establish the localization model and measurement equation in a two-dimensional scenario, and the Levenberg-Marquardt (LM) algorithm is employed for solution. Then, an optimization function is constructed by integrating Fisher information and control parameter constraints to dynamically plan the UAV trajectory, thereby improving the accuracy of transmitter localization and the practicality of this method. Finally, simulation experiments show that under the maximum control distance constraint, the positioning result of the proposed method is better than that of straight-line track and typical optimized track, and the final positioning accuracy is less than the standard deviation of the bistatic distance difference measurements, which can meet the application requirements of the passive radar system.
2024, 46(5): 2065-2075.
doi: 10.11999/JEIT231242
Abstract:
In order to enhance the radio frequency stealth performance of radar networks, a Joint Detection Threshold and Power Allocation Optimization (JDTPAO) Strategy for Multi-Target Tracking (MTT) in radar networks is studied in this paper. Firstly, based on the integrated structure of detection and tracking, the average detection probability in associated gate and the predicted Bayesian Cramér-Rao Lower Bound are respectively derived as the performance metrics for target detection and MTT. Subsequently, a joint detection threshold and power allocation optimization model for minimizing the total power consumption is formulated, in which the node selection, detection threshold and power allocation are to be optimized, and the predetermined detection and MTT performance requirements, as well as the limited transmit resource are considered as the constraints. To address this problem, an approach combining improved probabilistic data association algorithm and sequential quadratic programming is proposed. Simulation results indicate that the proposed strategy effectively reduces the power consumption of radar network while meeting the requirements for target detection and tracking, thereby improving the radio frequency stealth performance.
In order to enhance the radio frequency stealth performance of radar networks, a Joint Detection Threshold and Power Allocation Optimization (JDTPAO) Strategy for Multi-Target Tracking (MTT) in radar networks is studied in this paper. Firstly, based on the integrated structure of detection and tracking, the average detection probability in associated gate and the predicted Bayesian Cramér-Rao Lower Bound are respectively derived as the performance metrics for target detection and MTT. Subsequently, a joint detection threshold and power allocation optimization model for minimizing the total power consumption is formulated, in which the node selection, detection threshold and power allocation are to be optimized, and the predetermined detection and MTT performance requirements, as well as the limited transmit resource are considered as the constraints. To address this problem, an approach combining improved probabilistic data association algorithm and sequential quadratic programming is proposed. Simulation results indicate that the proposed strategy effectively reduces the power consumption of radar network while meeting the requirements for target detection and tracking, thereby improving the radio frequency stealth performance.
2024, 46(5): 2076-2086.
doi: 10.11999/JEIT231309
Abstract:
Multiband Fusion is an effective way to broaden bandwidth of radar, which plays a key role in the detection and recognition of small-scale target. However, the existing multiband fusion algorithms still face the problems of slow operation and low precision. Therefore, a super-resolution technique of multiband fusion based on optimal dictionary selection and orthogonal matching pursuit is proposed in this paper. Firstly, the parametric model of multiband radar signal is conducted. Next, Snake Optimizer (SO) is applied to the coherent processing. Then, an Orthogonal Matching Pursuit (OMP) algorithm based on the optimal Geometrical Theory of Diffraction (GTD) dictionary selection is used to extrapolate the vacant spectrum. Experiment results of simulated and measured data are given. Experimental results show that the proposed method can effectively achieve super-resolution. This method combines simplified model rough estimation with complete model fine estimation, effectively reducing the amount of computation and realizing fast and accurate multiband fusion extrapolation processing.
Multiband Fusion is an effective way to broaden bandwidth of radar, which plays a key role in the detection and recognition of small-scale target. However, the existing multiband fusion algorithms still face the problems of slow operation and low precision. Therefore, a super-resolution technique of multiband fusion based on optimal dictionary selection and orthogonal matching pursuit is proposed in this paper. Firstly, the parametric model of multiband radar signal is conducted. Next, Snake Optimizer (SO) is applied to the coherent processing. Then, an Orthogonal Matching Pursuit (OMP) algorithm based on the optimal Geometrical Theory of Diffraction (GTD) dictionary selection is used to extrapolate the vacant spectrum. Experiment results of simulated and measured data are given. Experimental results show that the proposed method can effectively achieve super-resolution. This method combines simplified model rough estimation with complete model fine estimation, effectively reducing the amount of computation and realizing fast and accurate multiband fusion extrapolation processing.
2024, 46(5): 2087-2094.
doi: 10.11999/JEIT231223
Abstract:
Compared with traditional Two-Dimensional (2D) Synthetic Aperture Radar (SAR) imaging, Three-Dimensional (3D) SAR imaging technology can overcome problems such as overlay and geometric distortion, thus having broad development space. As a typical 3D imaging system, the elevation resolution of array SAR is generally limited by the array aperture in theory, which is much lower than the range and azimuth resolution. To address this issue, an assumption of consistency in elevation between neighboring pixels is introduced and a re-weighted locally joint sparsity based Compressed Sensing (CS) approach is proposed for the array super-resolution imaging in the height dimension. Then, typical clustering methods such as K-means and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) are used to achieve clustering analysis of specific targets (such as buildings and vehicles) in the observation scene. Finally, the experimental analysis using measured data is performed to confirm the effectiveness of the proposed algorithm.
Compared with traditional Two-Dimensional (2D) Synthetic Aperture Radar (SAR) imaging, Three-Dimensional (3D) SAR imaging technology can overcome problems such as overlay and geometric distortion, thus having broad development space. As a typical 3D imaging system, the elevation resolution of array SAR is generally limited by the array aperture in theory, which is much lower than the range and azimuth resolution. To address this issue, an assumption of consistency in elevation between neighboring pixels is introduced and a re-weighted locally joint sparsity based Compressed Sensing (CS) approach is proposed for the array super-resolution imaging in the height dimension. Then, typical clustering methods such as K-means and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) are used to achieve clustering analysis of specific targets (such as buildings and vehicles) in the observation scene. Finally, the experimental analysis using measured data is performed to confirm the effectiveness of the proposed algorithm.
2024, 46(5): 2095-2103.
doi: 10.11999/JEIT230878
Abstract:
As a result of the characteristics of UnderWater Acoustic (UWA) channel, such as rapid fading of Channel Frequency Response (CFR) due to large delay spreading, the development of UnderWater Acoustic Communication (UWAC) technology is challenged. The acquisition of effective and reliable Channel State Information (CSI) at the transmitter is a prerequisite for adaptive communication. To meet the needs of UWA adaptive OFDM communication, a CSI-Grouping-Sequencing-Fitting-Feedback (CSI-GSFF) based on sequenced codebook algorithm is proposed, which consists of three steps, including grouping, sequencing, and data fitting. Firstly, adjacent pilot subcarriers are divided into several groups and each group is seen as a feedback cell. Then, the pilot subcarriers within each group are sorted according to the channel gains to mitigate adverse effects such as high feedback overhead caused by the rapid fading of CFR. Finally, polynomial fitting is performed, and the sorting operation effectively reduces the fitting order. Through the simulation of time-varying channel data in sea trials, the results show that the CSI-GSFF algorithm can achieve the Bit Error Rate (BER) performance of the UWA adaptive OFDM communication system under the perfect CSI, while effectively reduce the feedback overhead.
As a result of the characteristics of UnderWater Acoustic (UWA) channel, such as rapid fading of Channel Frequency Response (CFR) due to large delay spreading, the development of UnderWater Acoustic Communication (UWAC) technology is challenged. The acquisition of effective and reliable Channel State Information (CSI) at the transmitter is a prerequisite for adaptive communication. To meet the needs of UWA adaptive OFDM communication, a CSI-Grouping-Sequencing-Fitting-Feedback (CSI-GSFF) based on sequenced codebook algorithm is proposed, which consists of three steps, including grouping, sequencing, and data fitting. Firstly, adjacent pilot subcarriers are divided into several groups and each group is seen as a feedback cell. Then, the pilot subcarriers within each group are sorted according to the channel gains to mitigate adverse effects such as high feedback overhead caused by the rapid fading of CFR. Finally, polynomial fitting is performed, and the sorting operation effectively reduces the fitting order. Through the simulation of time-varying channel data in sea trials, the results show that the CSI-GSFF algorithm can achieve the Bit Error Rate (BER) performance of the UWA adaptive OFDM communication system under the perfect CSI, while effectively reduce the feedback overhead.
2024, 46(5): 2104-2110.
doi: 10.11999/JEIT231160
Abstract:
Traditional multireceiver Synthetic Aperture Sonar (SAS) imaging algorithms based on Phase Center Approximation (PCA) neglect the spatial variance of approximation error in the azimuth dimension. The distortion would be introduced in the focused results of distributed. To solve this problem, a two-way slant range considering the azimuth variance of approximation error is deduced based on the geometry models of transmitter/receiver bistatic sampling and PCA sampling. The system function in the 2D frequency domain is further decomposed into transmitter/receiver bistatic deformation term and quasi monostatic term. Based on that, the complex multiplication and interpolation are adopted to compensate the bistatic deformation term. Then, the range-Doppler imaging algorithm is used to focus the targets. Compared to traditional methods, much smaller appropriation error across the whole mapping swath is obtained by using the proposed method. Besides, the position deviation in the azimuth dimension is not introduced by the proposed method. The imaging result which is identical to practical target position can be obtained.
Traditional multireceiver Synthetic Aperture Sonar (SAS) imaging algorithms based on Phase Center Approximation (PCA) neglect the spatial variance of approximation error in the azimuth dimension. The distortion would be introduced in the focused results of distributed. To solve this problem, a two-way slant range considering the azimuth variance of approximation error is deduced based on the geometry models of transmitter/receiver bistatic sampling and PCA sampling. The system function in the 2D frequency domain is further decomposed into transmitter/receiver bistatic deformation term and quasi monostatic term. Based on that, the complex multiplication and interpolation are adopted to compensate the bistatic deformation term. Then, the range-Doppler imaging algorithm is used to focus the targets. Compared to traditional methods, much smaller appropriation error across the whole mapping swath is obtained by using the proposed method. Besides, the position deviation in the azimuth dimension is not introduced by the proposed method. The imaging result which is identical to practical target position can be obtained.
2024, 46(5): 2111-2121.
doi: 10.11999/JEIT230874
Abstract:
Directed Acyclic Graph(DAG)-based blockchain can significantly improve system performance and have become a research topic in both academia and industry. Compared with the traditional chain-based blockchains with serialization, DAG-based blockchains can process multiple blocks concurrently to package significant transactions into the chain. With the surge in transaction throughput, DAG-based blockchain faces the issue of low transaction execution efficiency, i.e., the demand for state data access for massive transaction execution increases dramatically, resulting in high Input/Output(I/O) overhead. Enabling low I/O state access mainly encounters two new challenges. On the one hand, if DAG-based blockchain directly adopts the traditional state prefetch mechanism, it will introduce a large number of stale reads due to inconsistent execution logic. On the other hand, state access for different accounts causes duplicate I/O overhead in the upper nodes of the Merkle tree. To this end, an efficient transaction execution mechanism based on pipelining—ErlangShen is designed, including the epoch granularity state prefetch mechanism and Merkle high-level path buffer mechanism to reduce the number of stale reads and duplicate I/O overhead, respectively. Specifically, ErlangShen leverages the complicated logic and severe conflicts of transactions accessing hotspot states to parallelize the execution of hotspot transactions and the prefetch of states accessed by cold transactions, to avoid the implication of the state prefetching on the transaction execution. Furthermore, the customized concurrency control methods is designed according to the data access pattern of hotspot and cold states to further improve the system throughput. Experimental results show that ErlangShen can reduce the number of stale reads by about 90% and improve transaction processing performance by 3～4 times compared to Nezha, the state-of-the-art DAG-based blockchain transaction processing solution.
Directed Acyclic Graph(DAG)-based blockchain can significantly improve system performance and have become a research topic in both academia and industry. Compared with the traditional chain-based blockchains with serialization, DAG-based blockchains can process multiple blocks concurrently to package significant transactions into the chain. With the surge in transaction throughput, DAG-based blockchain faces the issue of low transaction execution efficiency, i.e., the demand for state data access for massive transaction execution increases dramatically, resulting in high Input/Output(I/O) overhead. Enabling low I/O state access mainly encounters two new challenges. On the one hand, if DAG-based blockchain directly adopts the traditional state prefetch mechanism, it will introduce a large number of stale reads due to inconsistent execution logic. On the other hand, state access for different accounts causes duplicate I/O overhead in the upper nodes of the Merkle tree. To this end, an efficient transaction execution mechanism based on pipelining—ErlangShen is designed, including the epoch granularity state prefetch mechanism and Merkle high-level path buffer mechanism to reduce the number of stale reads and duplicate I/O overhead, respectively. Specifically, ErlangShen leverages the complicated logic and severe conflicts of transactions accessing hotspot states to parallelize the execution of hotspot transactions and the prefetch of states accessed by cold transactions, to avoid the implication of the state prefetching on the transaction execution. Furthermore, the customized concurrency control methods is designed according to the data access pattern of hotspot and cold states to further improve the system throughput. Experimental results show that ErlangShen can reduce the number of stale reads by about 90% and improve transaction processing performance by 3～4 times compared to Nezha, the state-of-the-art DAG-based blockchain transaction processing solution.
2024, 46(5): 2122-2136.
doi: 10.11999/JEIT231048
Abstract:
Dissimilar redundancy architecture is widely used in information systems to improve their security and reliability. When the system operates normally, the executers behave consistently, but when faced with malicious attacks, the executers exhibit inconsistency. The architecture improves the security and reliability of the system by comparing the performance of the executers to monitor the system and perceive threats. The synchronization of executers is a challenge that all dissimilar redundancy architectures need to address. There is currently no systematic description and summary of synchronization technology. This article is a review of executer synchronization techniques in dissimilar redundancy architectures. First, the importance of synchronization in dissimilar redundancy systems is explained and a standardized description of synchronization is provided. Then, a synchronization technology classification method based on synchronization points is proposed and the basic process, popularity, advantages and disadvantages of each class are summarized separately. This article also proposes three important indicators that affect synchronization performance, namely synchronization point, false alarm rate, and performance, and provides a mathematical model for synchronization technology, which can be used for design evaluation. Finally, this article combines the development of cyber resilience and software defined system on wafer, and points out the potential and possible directions for the future development of synchronous technology.
Dissimilar redundancy architecture is widely used in information systems to improve their security and reliability. When the system operates normally, the executers behave consistently, but when faced with malicious attacks, the executers exhibit inconsistency. The architecture improves the security and reliability of the system by comparing the performance of the executers to monitor the system and perceive threats. The synchronization of executers is a challenge that all dissimilar redundancy architectures need to address. There is currently no systematic description and summary of synchronization technology. This article is a review of executer synchronization techniques in dissimilar redundancy architectures. First, the importance of synchronization in dissimilar redundancy systems is explained and a standardized description of synchronization is provided. Then, a synchronization technology classification method based on synchronization points is proposed and the basic process, popularity, advantages and disadvantages of each class are summarized separately. This article also proposes three important indicators that affect synchronization performance, namely synchronization point, false alarm rate, and performance, and provides a mathematical model for synchronization technology, which can be used for design evaluation. Finally, this article combines the development of cyber resilience and software defined system on wafer, and points out the potential and possible directions for the future development of synchronous technology.
2024, 46(5): 2137-2148.
doi: 10.11999/JEIT231197
Abstract:
Wireless sensor networks use public wireless channels and their storage and computing resources are limited, making them vulnerable to active attacks and passive attacks. Identity authentication acts as the first line to ensure the security of information systems. Then, how to design multi-factor authentication schemes for wireless sensor devices is currently a hot topic. Nowadays, most existing schemes are based on foreign cryptographic standards that do not comply with the autonomous and controllable cyberspace security development strategy. SM9 is an identity-based cryptographic algorithm that has become a Chinese cryptographic standard recently. Therefore, this paper focuses on how to combine passwords, biometrics, and smart cards to design a multi-factor authentication scheme that can be used for wireless sensor networks under the framework of SM9. The proposed scheme applies the fuzzy verifier technique and the honeyword method to resist password guessing attacks and further enables session key negotiation and password update. The security is proved under the Random Oracle Model (ROM) and a heuristic security analysis is provided additionally. The comparison results show that the proposed scheme can be deployed to wireless sensor networks.
Wireless sensor networks use public wireless channels and their storage and computing resources are limited, making them vulnerable to active attacks and passive attacks. Identity authentication acts as the first line to ensure the security of information systems. Then, how to design multi-factor authentication schemes for wireless sensor devices is currently a hot topic. Nowadays, most existing schemes are based on foreign cryptographic standards that do not comply with the autonomous and controllable cyberspace security development strategy. SM9 is an identity-based cryptographic algorithm that has become a Chinese cryptographic standard recently. Therefore, this paper focuses on how to combine passwords, biometrics, and smart cards to design a multi-factor authentication scheme that can be used for wireless sensor networks under the framework of SM9. The proposed scheme applies the fuzzy verifier technique and the honeyword method to resist password guessing attacks and further enables session key negotiation and password update. The security is proved under the Random Oracle Model (ROM) and a heuristic security analysis is provided additionally. The comparison results show that the proposed scheme can be deployed to wireless sensor networks.
2024, 46(5): 2149-2158.
doi: 10.11999/JEIT231231
Abstract:
Integral attack is one of the most powerful cryptanalytic methods after differential and linear cryptanalysis, which was presented by Daemen et al. in 1997 (doi: 10.1007/BFb0052343). As the winning block cipher of China’s National Cipher Designing Competition in 2018, the security strength of uBlock against integral attack has received much attention. To better understand the integral property, this paper constructs the Mixed Integer Linear Programming (MILP) models for monomial prediction to search for the integral distinguishers and uses the partial sum techniques to perform key-recovery attacks. For uBlock-128/128 and uBlock-128/256, this paper gives the first 11 and 12-round attacks based on a 9-round integral distinguisher, respectively. The data complexity is\begin{document}$ {2}^{127} $\end{document} \begin{document}$ {2}^{127.06} $\end{document} \begin{document}$ {2}^{224} $\end{document} \begin{document}$ {2}^{44.58} $\end{document} \begin{document}$ {2}^{138} $\end{document} \begin{document}$ {2}^{253} $\end{document} \begin{document}$ {2}^{253.06} $\end{document} \begin{document}$ {2}^{44.46} $\end{document}
Integral attack is one of the most powerful cryptanalytic methods after differential and linear cryptanalysis, which was presented by Daemen et al. in 1997 (doi: 10.1007/BFb0052343). As the winning block cipher of China’s National Cipher Designing Competition in 2018, the security strength of uBlock against integral attack has received much attention. To better understand the integral property, this paper constructs the Mixed Integer Linear Programming (MILP) models for monomial prediction to search for the integral distinguishers and uses the partial sum techniques to perform key-recovery attacks. For uBlock-128/128 and uBlock-128/256, this paper gives the first 11 and 12-round attacks based on a 9-round integral distinguisher, respectively. The data complexity is
A Multi-party Vertically Partitioned Data Synthesis Mechanism with Personalized Differential Privacy
2024, 46(5): 2159-2176.
doi: 10.11999/JEIT231158
Abstract:
In today’s era, with the rapid development of big data technology and the continuous increase in data volume, large amounts of data are constantly collected by different companies or institutions, aggregating and publishing data owned by different companies or institutions helps to better provide services and support decision-making. However, their respective data may contain privacy information with different degrees of sensitivity, thus personalized privacy protection requirements need to be met while aggregating and publishing data from all parties. To solve the problem of multi-party data publication while ensuring that different privacy protection needs of all parties are met, a Multi-party Vertically partitioned Data Synthesis mechanism with Personalized Differential Privacy (PDP-MVDS) is proposed. Low-dimensional marginal distributions are firstly generated to reduce the dimension of high-dimensional data, then a randomly initialized dataset with these marginal distributions are updated, and finally a synthesized dataset whose distribution is similar to that of the real aggregated dataset from all parties is published. Personalized differential privacy protection is achieved by dividing the privacy budget; Secure scalar product protocol and threshold Paillier encryption algorithm are used to ensure the privacy of each party’s data in the aggregation process; Distributed Laplace perturbation mechanism is used to effectively protect the privacy of marginal distributions that aggregated from those parties. Through rigorous theoretical analysis, it is proved that PDP-MVDS can ensure the security of each participant’s data and the finally published dataset. Furthermore, the experimental results on public datasets show that PDP-MVDS mechanism can obtain a multi-party synthesized dataset with high utility under low overhead.
In today’s era, with the rapid development of big data technology and the continuous increase in data volume, large amounts of data are constantly collected by different companies or institutions, aggregating and publishing data owned by different companies or institutions helps to better provide services and support decision-making. However, their respective data may contain privacy information with different degrees of sensitivity, thus personalized privacy protection requirements need to be met while aggregating and publishing data from all parties. To solve the problem of multi-party data publication while ensuring that different privacy protection needs of all parties are met, a Multi-party Vertically partitioned Data Synthesis mechanism with Personalized Differential Privacy (PDP-MVDS) is proposed. Low-dimensional marginal distributions are firstly generated to reduce the dimension of high-dimensional data, then a randomly initialized dataset with these marginal distributions are updated, and finally a synthesized dataset whose distribution is similar to that of the real aggregated dataset from all parties is published. Personalized differential privacy protection is achieved by dividing the privacy budget; Secure scalar product protocol and threshold Paillier encryption algorithm are used to ensure the privacy of each party’s data in the aggregation process; Distributed Laplace perturbation mechanism is used to effectively protect the privacy of marginal distributions that aggregated from those parties. Through rigorous theoretical analysis, it is proved that PDP-MVDS can ensure the security of each participant’s data and the finally published dataset. Furthermore, the experimental results on public datasets show that PDP-MVDS mechanism can obtain a multi-party synthesized dataset with high utility under low overhead.
2024, 46(5): 2177-2186.
doi: 10.11999/JEIT231214
Abstract:
Metaverses is a new type of internet social ecosystem that promotes user interaction, provides virtual services, and enables digital asset transactions. Blockchain, as the underlying technology of metaverses, supports the circulation of digital assets such as Non-Fungible Token (NFT) within the metaverse. However, the increase in consensus nodes can decrease the consensus efficiency of digital asset transactions. Therefore, a multi-metaverse digital assets transaction management framework based on edge computing and cross-chain technology is proposed. Firstly, cross-chain technology is utilized to connect multiple sub-metaverses into a multi sub-metaverse system. Secondly, edge devices are allocated as miners in various sub-metaverses, contributing idle computational resources to enhance the efficiency of digital asset transactions. Additionally, the paper models the edge device allocation problem as a multi-knapsack problem and designs a miner selection approach. To address the dynamic allocation problem caused by environmental changes, the Deep Reinforcement Learning Proximal Policy Optimization (DRL-PPO) algorithm from deep reinforcement learning is employed. Simulation results demonstrate the effectiveness of the proposed scheme in achieving secure, efficient, and flexible cross-chain NFT transactions and sub-metaverse management.
Metaverses is a new type of internet social ecosystem that promotes user interaction, provides virtual services, and enables digital asset transactions. Blockchain, as the underlying technology of metaverses, supports the circulation of digital assets such as Non-Fungible Token (NFT) within the metaverse. However, the increase in consensus nodes can decrease the consensus efficiency of digital asset transactions. Therefore, a multi-metaverse digital assets transaction management framework based on edge computing and cross-chain technology is proposed. Firstly, cross-chain technology is utilized to connect multiple sub-metaverses into a multi sub-metaverse system. Secondly, edge devices are allocated as miners in various sub-metaverses, contributing idle computational resources to enhance the efficiency of digital asset transactions. Additionally, the paper models the edge device allocation problem as a multi-knapsack problem and designs a miner selection approach. To address the dynamic allocation problem caused by environmental changes, the Deep Reinforcement Learning Proximal Policy Optimization (DRL-PPO) algorithm from deep reinforcement learning is employed. Simulation results demonstrate the effectiveness of the proposed scheme in achieving secure, efficient, and flexible cross-chain NFT transactions and sub-metaverse management.
2024, 46(5): 2187-2197.
doi: 10.11999/JEIT240220
Abstract:
In recent years, the semi-supervised element extraction task in remote sensing, which utilizes unlabeled data to assist training with a small amount of labeled data, has been widely explored. Most existing approaches adopt self-training or consistency regularization methods to enhance element extraction performance. However, there still exists a significant discrepancy in accuracy among different categories due to the imbalanced distribution of data classes. Therefore, a feature extraction Framework Integrated with Distribution-Aligned Sampling (FIDAS) framework is proposed in this paper. By leveraging historical data class distributions, the framework adjusts the training difficulty for different categories while guiding the model to learn the true data distribution. Specifically, it utilizes historical data distribution information to sample from each category, increasing the probability of difficult-category instances passing through thresholds and enabling the model to capture more features of difficult categories. Furthermore, a distribution alignment loss is designed to improve the alignment between the learned category distribution and the true data category distribution, enhancing model robustness. Additionally, to reduce the computational overhead introduced by the Transformer model, an image feature block adaptive aggregation network is proposed, which aggregates redundant input image features to accelerate model training. Experiments are conducted on the remote sensing element extraction dataset Potsdam. Under the setting of a 1/32 semi-supervised data ratio, a 4.64% improvement in mean Intersection over Union (mIoU) is achieved by the proposed approach compared to state-of-the-art methods. Moreover, while the essential element extraction accuracy is maintained, the training time is reduced by approximately 30%. The effectiveness and performance advantages of the proposed method in semi-supervised remote sensing element extraction tasks are demonstrated by these results.
In recent years, the semi-supervised element extraction task in remote sensing, which utilizes unlabeled data to assist training with a small amount of labeled data, has been widely explored. Most existing approaches adopt self-training or consistency regularization methods to enhance element extraction performance. However, there still exists a significant discrepancy in accuracy among different categories due to the imbalanced distribution of data classes. Therefore, a feature extraction Framework Integrated with Distribution-Aligned Sampling (FIDAS) framework is proposed in this paper. By leveraging historical data class distributions, the framework adjusts the training difficulty for different categories while guiding the model to learn the true data distribution. Specifically, it utilizes historical data distribution information to sample from each category, increasing the probability of difficult-category instances passing through thresholds and enabling the model to capture more features of difficult categories. Furthermore, a distribution alignment loss is designed to improve the alignment between the learned category distribution and the true data category distribution, enhancing model robustness. Additionally, to reduce the computational overhead introduced by the Transformer model, an image feature block adaptive aggregation network is proposed, which aggregates redundant input image features to accelerate model training. Experiments are conducted on the remote sensing element extraction dataset Potsdam. Under the setting of a 1/32 semi-supervised data ratio, a 4.64% improvement in mean Intersection over Union (mIoU) is achieved by the proposed approach compared to state-of-the-art methods. Moreover, while the essential element extraction accuracy is maintained, the training time is reduced by approximately 30%. The effectiveness and performance advantages of the proposed method in semi-supervised remote sensing element extraction tasks are demonstrated by these results.
2024, 46(5): 2198-2216.
doi: 10.11999/JEIT231400
Abstract:
Due to the limitations of imaging principles and processing technology, uncooled infrared imagery detectors suffer from serious non-uniformities, damaging the imaging results. To improve the quality of infrared images, non-uniformity correction techniques are of great significance for practical applications. According to the physical formation and spatial characteristics of the non-uniformity in uncooled infrared imagery detectors, this paper divides these common non-uniformities into three categories: low-frequency non-uniformity, shot non-uniformity, and stripe non-uniformity. The physical mechanisms of these non-uniformities are further explored from the procedure in the optical system, thermal materials, and amplifier circuit of the uncooled infrared imagery detector. Then, the existing shutter-less non-uniformity correction methods are systematically summarized. Based on the principles of the methods, statistical-based, filter-based, optimization-based, and learning-based non-uniformity correction methods are categorized. Besides, the specifics of each method in dealing with different non-uniformities are clarified and distinguished. Finally, the existing problems in the current methods are reviewed and summarized, and the development trend of non-uniformity correction methods in future practical applications is prospected.
Due to the limitations of imaging principles and processing technology, uncooled infrared imagery detectors suffer from serious non-uniformities, damaging the imaging results. To improve the quality of infrared images, non-uniformity correction techniques are of great significance for practical applications. According to the physical formation and spatial characteristics of the non-uniformity in uncooled infrared imagery detectors, this paper divides these common non-uniformities into three categories: low-frequency non-uniformity, shot non-uniformity, and stripe non-uniformity. The physical mechanisms of these non-uniformities are further explored from the procedure in the optical system, thermal materials, and amplifier circuit of the uncooled infrared imagery detector. Then, the existing shutter-less non-uniformity correction methods are systematically summarized. Based on the principles of the methods, statistical-based, filter-based, optimization-based, and learning-based non-uniformity correction methods are categorized. Besides, the specifics of each method in dealing with different non-uniformities are clarified and distinguished. Finally, the existing problems in the current methods are reviewed and summarized, and the development trend of non-uniformity correction methods in future practical applications is prospected.
2024, 46(5): 2217-2227.
doi: 10.11999/JEIT231062
Abstract:
A multi-relation perception network for infrared and visible image fusion is proposed in this paper to fully integrate consistent features and complementary features between infrared and visible images. First, a dual-branch encoder module is used to extract features from the source images. The extracted features are then fed into the fusion strategy module based on multi-relation perception. Finally, a decoder module is used to reconstruct the fused features and generate the final fused image. In this fusion strategy module, the feature relationship perception and the weight relationship perception are constructed by exploring the interactions between the shared relationship, the differential relationship, and the cumulative relationship across different modalities, so as to integrate consistent features and complementary features between different modalities and obtain fused features. To constrain network training and preserve the intrinsic features of the source images, a wavelet transform-based loss function is developed to assist in preserving low-frequency components and high-frequency components of the source images during the fusion process. Experiments indicate that, compared to the state-of-the-art deep learning-based image fusion methods, the proposed method can fully integrate consistent features and complementary features between source images, thereby successfully preserving the background information of visible images and the thermal targets of infrared images. Overall, the fusion performance of the proposed method surpasses that of the compared methods.
A multi-relation perception network for infrared and visible image fusion is proposed in this paper to fully integrate consistent features and complementary features between infrared and visible images. First, a dual-branch encoder module is used to extract features from the source images. The extracted features are then fed into the fusion strategy module based on multi-relation perception. Finally, a decoder module is used to reconstruct the fused features and generate the final fused image. In this fusion strategy module, the feature relationship perception and the weight relationship perception are constructed by exploring the interactions between the shared relationship, the differential relationship, and the cumulative relationship across different modalities, so as to integrate consistent features and complementary features between different modalities and obtain fused features. To constrain network training and preserve the intrinsic features of the source images, a wavelet transform-based loss function is developed to assist in preserving low-frequency components and high-frequency components of the source images during the fusion process. Experiments indicate that, compared to the state-of-the-art deep learning-based image fusion methods, the proposed method can fully integrate consistent features and complementary features between source images, thereby successfully preserving the background information of visible images and the thermal targets of infrared images. Overall, the fusion performance of the proposed method surpasses that of the compared methods.
2024, 46(5): 2228-2236.
doi: 10.11999/JEIT231304
Abstract:
Group activity recognition aims to identify behaviors involving multiple individuals. In real-world applications, group behavior is often treated as a hierarchical structure, which consists group, subgroups and individuals. Previous researches have been focused on modeling relationships between individuals, without in-depth relationship analysis between subgroups. Therefore, a novel hierarchical group activity recognition framework based on Multi-scale Sub-group Interaction Relationships (MSIR) is proposed, and an innovative multi-scale interaction features extraction method between subgroups is presented as specified below. A sub-group division module is implemented. It aggregates individuals with potential correlations based on their appearance features and spatial positions, then dynamically generates subgroups of different scales using semantic information. A sub-group interactive feature extraction module is developed to extract more discriminative subgroup features. It constructs interaction matrices between different subgroups and leverages the relational reasoning capabilities of graph neural networks. Compared with existing twelve methods on benchmark datasets for group behavior recognition, including volleyball and collective activity datasets, the methodology of this paper demonstrates superior performance. This research presents an easily extendable and adaptable group activity recognition framework, exhibiting strong generalization capabilities across different datasets.
Group activity recognition aims to identify behaviors involving multiple individuals. In real-world applications, group behavior is often treated as a hierarchical structure, which consists group, subgroups and individuals. Previous researches have been focused on modeling relationships between individuals, without in-depth relationship analysis between subgroups. Therefore, a novel hierarchical group activity recognition framework based on Multi-scale Sub-group Interaction Relationships (MSIR) is proposed, and an innovative multi-scale interaction features extraction method between subgroups is presented as specified below. A sub-group division module is implemented. It aggregates individuals with potential correlations based on their appearance features and spatial positions, then dynamically generates subgroups of different scales using semantic information. A sub-group interactive feature extraction module is developed to extract more discriminative subgroup features. It constructs interaction matrices between different subgroups and leverages the relational reasoning capabilities of graph neural networks. Compared with existing twelve methods on benchmark datasets for group behavior recognition, including volleyball and collective activity datasets, the methodology of this paper demonstrates superior performance. This research presents an easily extendable and adaptable group activity recognition framework, exhibiting strong generalization capabilities across different datasets.
2024, 46(5): 2237-2248.
doi: 10.11999/JEIT231209
Abstract:
HyperSpectral Image (HSI) classification is one of the most prominent research topics in geoscience and remote sensing image processing tasks. In recent years, the combination of Convolutional Neural Network (CNN) and vision transformer has achieved success in HSI classification tasks by comprehensively considering local-global information. Nevertheless, the ground objects of HSIs vary in scale, containing rich texture information and complex structures. The current methods based on the combination of CNN and vision transformer usually have limited capability to extract texture and structural information of multi-scale ground objects. To overcome the above limitations, a CNN and vision transformer-driven cross-layer multi-scale fusion network is proposed for HSI classification. Firstly, from the perspective of combining CNN and visual transformer, a cross-layer multi-scale local-global feature extraction module branch is constructed, which is composed of a convolution embedded vision transformer architecture and a cross-layer feature fusion module. Specifically, to enhance attention to multi-scale ground objects in HSIs, the convolution embedded vision transformer captures multi-scale local-global features effectively by organically combining multi-scale CNN and vision transformer. Furthermore, the cross-layer feature fusion module aggregates hierarchical multi-scale local-global features, thereby combining shallow texture information and deep structural information of ground objects. Secondly, a group multi-scale convolution module branch is designed to explore the potential multi-scale features from abundant spectral bands in HSIs. Finally, to mine local spectral details and global spectral information in HSIs, a residual group convolution module is designed to extract local-global spectral features. Experimental results on Indian Pines, Houston 2013, and Salinas Valley datasets confirm the effectiveness of the proposed method.
HyperSpectral Image (HSI) classification is one of the most prominent research topics in geoscience and remote sensing image processing tasks. In recent years, the combination of Convolutional Neural Network (CNN) and vision transformer has achieved success in HSI classification tasks by comprehensively considering local-global information. Nevertheless, the ground objects of HSIs vary in scale, containing rich texture information and complex structures. The current methods based on the combination of CNN and vision transformer usually have limited capability to extract texture and structural information of multi-scale ground objects. To overcome the above limitations, a CNN and vision transformer-driven cross-layer multi-scale fusion network is proposed for HSI classification. Firstly, from the perspective of combining CNN and visual transformer, a cross-layer multi-scale local-global feature extraction module branch is constructed, which is composed of a convolution embedded vision transformer architecture and a cross-layer feature fusion module. Specifically, to enhance attention to multi-scale ground objects in HSIs, the convolution embedded vision transformer captures multi-scale local-global features effectively by organically combining multi-scale CNN and vision transformer. Furthermore, the cross-layer feature fusion module aggregates hierarchical multi-scale local-global features, thereby combining shallow texture information and deep structural information of ground objects. Secondly, a group multi-scale convolution module branch is designed to explore the potential multi-scale features from abundant spectral bands in HSIs. Finally, to mine local spectral details and global spectral information in HSIs, a residual group convolution module is designed to extract local-global spectral features. Experimental results on Indian Pines, Houston 2013, and Salinas Valley datasets confirm the effectiveness of the proposed method.
2024, 46(5): 2249-2263.
doi: 10.11999/JEIT231330
Abstract:
For automatic recognition of the depression intensity in patients, the existing deep learning based methods typically face two main challenges: (1) It is difficult for deep models to effectively capture the global context information relevant to the level of depression intensity from facial expressions, and (2) the semantic consistency between the global semantic information and the local one associated with depression intensity is often ignored. One new deep neural network for recognizing the severity of depressive symptoms, by combining the Perceptually Locally-Enhanced Global Depression Features and the Fused Global-Local Semantic Correlation Features (PLEGDF-FGLSCF), is proposed in this paper. Firstly, the PLEGDF module for the extraction of global depression features with local perceptual enhancement, is designed to extract the semantic correlations among local facial regions, to promote the interactions between depression-relevant information in different local regions, and thus to enhance the expressiveness of the global depression features driven by the local ones. Secondly, in order to achieve full integration of global and local semantic features related to depression severity, the FGLSCF module is proposed, aiming to capture the correlation of global and local semantic information and thus to ensure the semantic consistency in describing the depression intensity by means of global and local semantic features. Finally, on the AVEC2013 and AVEC2014 datasets, the PLEGDF-FGLSCF model achieved recognition results in terms of the Root Mean Square Error (RMSE) and the Mean Absolute Error (MAE) with the values of 7.75/5.96 and 7.49/5.99, respectively, demonstrating its superiority to most existing benchmark methods, verifying the rationality and effectiveness of our approach.
For automatic recognition of the depression intensity in patients, the existing deep learning based methods typically face two main challenges: (1) It is difficult for deep models to effectively capture the global context information relevant to the level of depression intensity from facial expressions, and (2) the semantic consistency between the global semantic information and the local one associated with depression intensity is often ignored. One new deep neural network for recognizing the severity of depressive symptoms, by combining the Perceptually Locally-Enhanced Global Depression Features and the Fused Global-Local Semantic Correlation Features (PLEGDF-FGLSCF), is proposed in this paper. Firstly, the PLEGDF module for the extraction of global depression features with local perceptual enhancement, is designed to extract the semantic correlations among local facial regions, to promote the interactions between depression-relevant information in different local regions, and thus to enhance the expressiveness of the global depression features driven by the local ones. Secondly, in order to achieve full integration of global and local semantic features related to depression severity, the FGLSCF module is proposed, aiming to capture the correlation of global and local semantic information and thus to ensure the semantic consistency in describing the depression intensity by means of global and local semantic features. Finally, on the AVEC2013 and AVEC2014 datasets, the PLEGDF-FGLSCF model achieved recognition results in terms of the Root Mean Square Error (RMSE) and the Mean Absolute Error (MAE) with the values of 7.75/5.96 and 7.49/5.99, respectively, demonstrating its superiority to most existing benchmark methods, verifying the rationality and effectiveness of our approach.
2024, 46(5): 2264-2273.
doi: 10.11999/JEIT231298
Abstract:
Tunneling Field Effect Transistor (TFET)-based low-power microcontroller design combines devices, circuits, and systems to achieve extremely low leakage power consumption under non-operating conditions through devices with ultra-low subthreshold swing characteristics, avoiding the power bottleneck brought about by the theoretical limit of the subthreshold swing of MOSFETs, and solving the problem of low power consumption of battery-powered microcontrollers. TFETs differ greatly from traditional MOSFETs in their operating mechanism, mainly in that they have lower leakage current after shutdown and can operate at lower voltages, which makes them suitable for IoT application scenarios with low-power requirements for long-dormant battery power supply. The paper investigates the research of TFET devices in low-power circuit design in recent years, introduces the structure of traditional microcontrollers and the source of power consumption, and at the same time explains the working principle, characteristics and design challenges of TFET devices, examines the research and development process of TFETs in the fields of digital circuits, analog circuits, and system design, and analyzes the strengths and weaknesses of each design scheme, and analyzes the advantages and disadvantages of TFETs in low-power circuits, in combination with the research of the literature. The future outlook of TFETs in the field of low-power microcontrollers is analyzed.
Tunneling Field Effect Transistor (TFET)-based low-power microcontroller design combines devices, circuits, and systems to achieve extremely low leakage power consumption under non-operating conditions through devices with ultra-low subthreshold swing characteristics, avoiding the power bottleneck brought about by the theoretical limit of the subthreshold swing of MOSFETs, and solving the problem of low power consumption of battery-powered microcontrollers. TFETs differ greatly from traditional MOSFETs in their operating mechanism, mainly in that they have lower leakage current after shutdown and can operate at lower voltages, which makes them suitable for IoT application scenarios with low-power requirements for long-dormant battery power supply. The paper investigates the research of TFET devices in low-power circuit design in recent years, introduces the structure of traditional microcontrollers and the source of power consumption, and at the same time explains the working principle, characteristics and design challenges of TFET devices, examines the research and development process of TFETs in the fields of digital circuits, analog circuits, and system design, and analyzes the strengths and weaknesses of each design scheme, and analyzes the advantages and disadvantages of TFETs in low-power circuits, in combination with the research of the literature. The future outlook of TFETs in the field of low-power microcontrollers is analyzed.
2024, 46(5): 2274-2280.
doi: 10.11999/JEIT231313
Abstract:
As a novel hardware security primitive, Physical Unclonable Function (PUF) extracts process deviations to generate a unique response sequence, providing a root of trust for computing systems. However, existing PUFs based on Field Programmable Gate Arrays (FPGAs) cannot maintain high reliability over a wide range of temperatures and voltages. In this work, we propose a Self-Timed Ring (STR) based Self-adaption Deviation Locking PUF (SDL PUF). Firstly, the PUF response is generated utilizing the oscillation frequency difference caused by the STR delay. Secondly, the adaptive configuration in the initialization stage can effectively expand the deviation of the event arrival time in the STR, substantially enhancing the reliability of PUF. Finally, a comparator obfuscation strategy is proposed, automatically configuring the comparator by extracting the process deviation to resist the side-channel attack. The proposed structure is implemented on a Xilinx Virtex-6 FPGA. Experimental results show that the proposed SDL PUF achieves 0 bit error rate in the temperature range of 0°C～80°C and the voltage range of 0.85～1.15V, and ensures 49.29% uniqueness and 49.84% uniformity while maintaining high reliability.
As a novel hardware security primitive, Physical Unclonable Function (PUF) extracts process deviations to generate a unique response sequence, providing a root of trust for computing systems. However, existing PUFs based on Field Programmable Gate Arrays (FPGAs) cannot maintain high reliability over a wide range of temperatures and voltages. In this work, we propose a Self-Timed Ring (STR) based Self-adaption Deviation Locking PUF (SDL PUF). Firstly, the PUF response is generated utilizing the oscillation frequency difference caused by the STR delay. Secondly, the adaptive configuration in the initialization stage can effectively expand the deviation of the event arrival time in the STR, substantially enhancing the reliability of PUF. Finally, a comparator obfuscation strategy is proposed, automatically configuring the comparator by extracting the process deviation to resist the side-channel attack. The proposed structure is implemented on a Xilinx Virtex-6 FPGA. Experimental results show that the proposed SDL PUF achieves 0 bit error rate in the temperature range of 0°C～80°C and the voltage range of 0.85～1.15V, and ensures 49.29% uniqueness and 49.84% uniformity while maintaining high reliability.
2024, 46(5): 2281-2288.
doi: 10.11999/JEIT231129
Abstract:
Physical Unclonable Function (PUF) has broad application prospects in the field of hardware security, but it is susceptible to modeling attacks based on machine learning. By studying the strong PUF circuit structure and chaotic mapping mechanism, a PUF circuit that can effectively resist machine learning modeling attacks is proposed. This circuit takes the original excitation as the initial value of the chaotic mapping, utilizes the internal relationship between the PUF excitation response mapping time and the iteration depth of the chaotic algorithm to generate unpredictable chaotic values, and uses PUF intermediate response feedback to encrypt the excitation. It can further improve the complexity of excitation and response mapping, thereby enhancing the resistance of PUF to machine learning attacks. The PUF is implemented using Artix-7 FPGA. The test results show that even with up to 1 million sets of excitation response pairs selected, the attack prediction rate based on logistic regression, support vector machine, and artificial neural network is still close to the ideal value of 50%. And the PUF has good randomness, uniqueness, and stability.
Physical Unclonable Function (PUF) has broad application prospects in the field of hardware security, but it is susceptible to modeling attacks based on machine learning. By studying the strong PUF circuit structure and chaotic mapping mechanism, a PUF circuit that can effectively resist machine learning modeling attacks is proposed. This circuit takes the original excitation as the initial value of the chaotic mapping, utilizes the internal relationship between the PUF excitation response mapping time and the iteration depth of the chaotic algorithm to generate unpredictable chaotic values, and uses PUF intermediate response feedback to encrypt the excitation. It can further improve the complexity of excitation and response mapping, thereby enhancing the resistance of PUF to machine learning attacks. The PUF is implemented using Artix-7 FPGA. The test results show that even with up to 1 million sets of excitation response pairs selected, the attack prediction rate based on logistic regression, support vector machine, and artificial neural network is still close to the ideal value of 50%. And the PUF has good randomness, uniqueness, and stability.
2024, 46(5): 2289-2297.
doi: 10.11999/JEIT231166
Abstract:
True Random Number Generator (TRNG), as an important hardware security primitive, is used in key generation, initialization vector and identity authentication in protocols. In order to design a lightweight TRNG with high throughput, the method of generating metastability is studied by using the switching characteristics of MUltipleXer (MUX) and XOR gate, and a TRNG design based on Metastability Superposition (MS-TRNG) cell (MS-cell) is proposed. It superimposes MUX and XOR gate guided metastases, thereby increasing the entropy of TRNG. The proposed TRNG is implemented in Xilinx Virtex-7 and Xilinx Artix-7 FPGA development boards, respectively, without the need for post-processing circuits. Compared to other advanced TRNGS, the proposed TRNG has the highest throughput and extremely low hardware overhead, and the random sequences it generates pass NIST testing and a series of performance tests.
True Random Number Generator (TRNG), as an important hardware security primitive, is used in key generation, initialization vector and identity authentication in protocols. In order to design a lightweight TRNG with high throughput, the method of generating metastability is studied by using the switching characteristics of MUltipleXer (MUX) and XOR gate, and a TRNG design based on Metastability Superposition (MS-TRNG) cell (MS-cell) is proposed. It superimposes MUX and XOR gate guided metastases, thereby increasing the entropy of TRNG. The proposed TRNG is implemented in Xilinx Virtex-7 and Xilinx Artix-7 FPGA development boards, respectively, without the need for post-processing circuits. Compared to other advanced TRNGS, the proposed TRNG has the highest throughput and extremely low hardware overhead, and the random sequences it generates pass NIST testing and a series of performance tests.
