Email alert
2016 Vol. 38, No. 8
Display Method:
2016, 38(8): 1843-1849.
doi: 10.11999/JEIT151244
Abstract:
In-network caching is one of the key technologies of Content-Centric Networking (CCN), which is widely concerned recently. However, most existing works are targeted for optimizing network resource utilization, and the energy consumption aspect is largely ignored. In this paper, first an energy consumption model for content distribution is built and a judging condition for energy efficiency optimization in caching is designed. On this basis and in combination with content popularity and node centrality, an Energy-Efficiency Aware Probabilistic Caching (E2APC) scheme is proposed. Simulation results show that the proposed scheme can effectively reduce the whole energy consumption, while guaranteeing comparatively high cache hit rate and few average response hops.
In-network caching is one of the key technologies of Content-Centric Networking (CCN), which is widely concerned recently. However, most existing works are targeted for optimizing network resource utilization, and the energy consumption aspect is largely ignored. In this paper, first an energy consumption model for content distribution is built and a judging condition for energy efficiency optimization in caching is designed. On this basis and in combination with content popularity and node centrality, an Energy-Efficiency Aware Probabilistic Caching (E2APC) scheme is proposed. Simulation results show that the proposed scheme can effectively reduce the whole energy consumption, while guaranteeing comparatively high cache hit rate and few average response hops.
2016, 38(8): 1850-1857.
doi: 10.11999/JEIT151230
Abstract:
Recovery of link failure is not only the issue of selecting a connected backup path, but the QoS requirement in the process of failure recovery of the network service should be also taken into account. The probabilistically correlated failure model and rerouting traffic disruption optimization target are built based on multiple backup paths strategy. Furthermore, a mathematical model of the failure recovery problem is modeled, which takes the minimum rerouting traffic disruption as the target and the QoS requirement as the constrain, and a link failure recovery algorithm based on multiple backup paths with QoS constrain is proposed. The proposed algorithm takes reducing rerouting traffic disruption farthest as the target and adopts the improved k shortest path algorithm with QoS constrain to splice the single backup path, and it gives the links more protection resource with high priority. Moreover, the correctness of the proposed algorithm is proved, and the time complex and the space complex are computed. The simulation results under NS2 show that the proposed algorithm significantly improves link failure recovery rate and the QoS satisfaction rate of the rerouting traffic, and it performs better when the QoS constrain is stronger.
Recovery of link failure is not only the issue of selecting a connected backup path, but the QoS requirement in the process of failure recovery of the network service should be also taken into account. The probabilistically correlated failure model and rerouting traffic disruption optimization target are built based on multiple backup paths strategy. Furthermore, a mathematical model of the failure recovery problem is modeled, which takes the minimum rerouting traffic disruption as the target and the QoS requirement as the constrain, and a link failure recovery algorithm based on multiple backup paths with QoS constrain is proposed. The proposed algorithm takes reducing rerouting traffic disruption farthest as the target and adopts the improved k shortest path algorithm with QoS constrain to splice the single backup path, and it gives the links more protection resource with high priority. Moreover, the correctness of the proposed algorithm is proved, and the time complex and the space complex are computed. The simulation results under NS2 show that the proposed algorithm significantly improves link failure recovery rate and the QoS satisfaction rate of the rerouting traffic, and it performs better when the QoS constrain is stronger.
2016, 38(8): 1858-1864.
doi: 10.11999/JEIT151203
Abstract:
Traditional mobile target localization algorithms are not suitable for wireless sensor networks as they need to collect, store, and process a mass of data. To address this issue, a mobile target localization algorithm based on compressive sensing is proposed. Two sparse representation bases are designed by exploiting the movement characteristics of mobile targets, therefore the mobile target localization issue is transferred into a sparse signal recovery issue. To avoid the unpractical limitation of traditional measurement matrices, two sparse measurement matrices are proposed that are practical and lowly coherent with the designed representation bases. The characteristic of this algorithm is that mobile target localization can be achieved by collecting a little data, thus prolonging the lifetime of wireless sensor networks. Simulation results indicate that the proposed localization algorithm based on compressive sensing is highly efficient.
Traditional mobile target localization algorithms are not suitable for wireless sensor networks as they need to collect, store, and process a mass of data. To address this issue, a mobile target localization algorithm based on compressive sensing is proposed. Two sparse representation bases are designed by exploiting the movement characteristics of mobile targets, therefore the mobile target localization issue is transferred into a sparse signal recovery issue. To avoid the unpractical limitation of traditional measurement matrices, two sparse measurement matrices are proposed that are practical and lowly coherent with the designed representation bases. The characteristic of this algorithm is that mobile target localization can be achieved by collecting a little data, thus prolonging the lifetime of wireless sensor networks. Simulation results indicate that the proposed localization algorithm based on compressive sensing is highly efficient.
2016, 38(8): 1865-1871.
doi: 10.11999/JEIT151205
Abstract:
Previous network-trace-based methods only consider the plaintext format of payload data, and are not suitable for security protocols which include a large number of ciphertext data; therefore, a novel approach named CFIA (Ciphertext Field Identification Approach) is proposed based on entropy estimation for unknown security protocols. On the basis of keywords sequences extraction, CFIA utilizes byte sample entropy and entropy estimation to pre-locate ciphertext filed, and further searches ciphertext length field to identify ciphertext field. The experimental results show that without using dynamic binary analysis, the proposed method can effectively identify ciphertext fields purely from network traces, and the inferred formats are highly accurate in identifying the protocols.
Previous network-trace-based methods only consider the plaintext format of payload data, and are not suitable for security protocols which include a large number of ciphertext data; therefore, a novel approach named CFIA (Ciphertext Field Identification Approach) is proposed based on entropy estimation for unknown security protocols. On the basis of keywords sequences extraction, CFIA utilizes byte sample entropy and entropy estimation to pre-locate ciphertext filed, and further searches ciphertext length field to identify ciphertext field. The experimental results show that without using dynamic binary analysis, the proposed method can effectively identify ciphertext fields purely from network traces, and the inferred formats are highly accurate in identifying the protocols.
2016, 38(8): 1872-1879.
doi: 10.11999/JEIT151264
Abstract:
An efficient range matching method based on Bloom Filter algorithm and Ternary Content Addressable Memory (BF-TCAM) technology is proposed to resolve the problem that there generally exit low memory using ratio and high power dissipation in current TCAM range matching methods. An algorithm of Segmented Match on Longest Common Prefix (SMLCP) splits range matching into two stepsprefix matching and feature range comparation, resulting in 100% TCAM space using ratio. BF-TCAM is designed according to SMLCP algorithm, which filters searching key words by Bloom filter to avoid that unrelated items participate in comparation, so as to reduce greatly power dissipation. Critical paths are streamlined so that searching operation can be completed during one clock cycle. Research results demonstrate that BF-TCAM achieves zero range expansion, meanwhile power dissipation falls more than 50%.
An efficient range matching method based on Bloom Filter algorithm and Ternary Content Addressable Memory (BF-TCAM) technology is proposed to resolve the problem that there generally exit low memory using ratio and high power dissipation in current TCAM range matching methods. An algorithm of Segmented Match on Longest Common Prefix (SMLCP) splits range matching into two stepsprefix matching and feature range comparation, resulting in 100% TCAM space using ratio. BF-TCAM is designed according to SMLCP algorithm, which filters searching key words by Bloom filter to avoid that unrelated items participate in comparation, so as to reduce greatly power dissipation. Critical paths are streamlined so that searching operation can be completed during one clock cycle. Research results demonstrate that BF-TCAM achieves zero range expansion, meanwhile power dissipation falls more than 50%.
2016, 38(8): 1880-1887.
doi: 10.11999/JEIT151201
Abstract:
With the access distance of users increasing in hybrid Time and Wavelength Division Multiplexing Passive Optical Networks (TWDM-PONs), the round trip time becomes very long, and idle time is increased in the network. It leads to the problem of bandwidth waste. In order to solve this issue, an efficient and hybrid dynamic resource allocation strategy with Grey Prediction (GP) is proposed. In order to decrease the network delay performance and avoid the waste of resource, the Grey Prediction model is utilized to predict the arrival data during the idle time and dynamically allocate bandwidth to optical network units. Utilizing the finishing time difference of data transmission on wavelengths, the higher transmission efficiency and load balance of wavelengths can be achieved by constantly adjusting the mode of wavelengths transmission periodically. The simulation results show that the proposed hybrid strategy can effectively eliminate the idle time to avoid waste of bandwidth and reduce network delay while making wavelengths efficiently used to improve resource utilization rate.
With the access distance of users increasing in hybrid Time and Wavelength Division Multiplexing Passive Optical Networks (TWDM-PONs), the round trip time becomes very long, and idle time is increased in the network. It leads to the problem of bandwidth waste. In order to solve this issue, an efficient and hybrid dynamic resource allocation strategy with Grey Prediction (GP) is proposed. In order to decrease the network delay performance and avoid the waste of resource, the Grey Prediction model is utilized to predict the arrival data during the idle time and dynamically allocate bandwidth to optical network units. Utilizing the finishing time difference of data transmission on wavelengths, the higher transmission efficiency and load balance of wavelengths can be achieved by constantly adjusting the mode of wavelengths transmission periodically. The simulation results show that the proposed hybrid strategy can effectively eliminate the idle time to avoid waste of bandwidth and reduce network delay while making wavelengths efficiently used to improve resource utilization rate.
2016, 38(8): 1888-1893.
doi: 10.11999/JEIT151259
Abstract:
Sparse Code Multiple Access (SCMA) is a promising air-interface technology for 5 G wireless communication networks, which can enable massive connectivity. The existing multiuser detection schemes are based on a parallel message updating for Message Passing Algorithm (MPA), thus it is not efficient in terms of convergence. In this paper, an efficient multiuser detection scheme for uplink SCMA is proposed based on serial updating of function nodes, messages. Compared to the existing detection schemes, the proposed scheme accelerates the convergence due to that the updated messages can join the belief propagation immediately in current iteration, which avoids being used in the next iteration. Furthermore, the proposed scheme can reduce the storage burden, which fuses message passing process on the basis of the relationship between messages. Numerical results show that the proposed scheme can offer a good trade-off between complexity and Bit Error Rate (BER) performance.
Sparse Code Multiple Access (SCMA) is a promising air-interface technology for 5 G wireless communication networks, which can enable massive connectivity. The existing multiuser detection schemes are based on a parallel message updating for Message Passing Algorithm (MPA), thus it is not efficient in terms of convergence. In this paper, an efficient multiuser detection scheme for uplink SCMA is proposed based on serial updating of function nodes, messages. Compared to the existing detection schemes, the proposed scheme accelerates the convergence due to that the updated messages can join the belief propagation immediately in current iteration, which avoids being used in the next iteration. Furthermore, the proposed scheme can reduce the storage burden, which fuses message passing process on the basis of the relationship between messages. Numerical results show that the proposed scheme can offer a good trade-off between complexity and Bit Error Rate (BER) performance.
2016, 38(8): 1894-1900.
doi: 10.11999/JEIT151207
Abstract:
Concurrent data transmission from massive Machine Type Communications (MTC) devices makes the traffic pattern of MTC more bursty, which invalidates the commonly-used methodologies of traffic engineering for multichannel S-ALOHA under the assumption of homogeneous or compound Poisson process. By usage of the number of contending devices that transmit the j-th preamble at the i-th Random Access (RA) slot as state variable, an innovative iterative process with its simplified form is proposed to acquire the dynamic process of multichannel S-ALOHA. It reveals the direct relation between the number of contending devices that transmit the j-th preamble at the i-th RA slot and the newly arrived devices before i-th RA slot. It also presents an analytical way to compute the probability density function, cumulative density function and mean of access delay. Numerical results by the use of MTC traffic models proposed by 3GPP are conducted to validate the effectiveness of the proposed iterative process and its simplified form. These works provide engineers insights to design enhanced overload control mechanism for MTC applications.
Concurrent data transmission from massive Machine Type Communications (MTC) devices makes the traffic pattern of MTC more bursty, which invalidates the commonly-used methodologies of traffic engineering for multichannel S-ALOHA under the assumption of homogeneous or compound Poisson process. By usage of the number of contending devices that transmit the j-th preamble at the i-th Random Access (RA) slot as state variable, an innovative iterative process with its simplified form is proposed to acquire the dynamic process of multichannel S-ALOHA. It reveals the direct relation between the number of contending devices that transmit the j-th preamble at the i-th RA slot and the newly arrived devices before i-th RA slot. It also presents an analytical way to compute the probability density function, cumulative density function and mean of access delay. Numerical results by the use of MTC traffic models proposed by 3GPP are conducted to validate the effectiveness of the proposed iterative process and its simplified form. These works provide engineers insights to design enhanced overload control mechanism for MTC applications.
2016, 38(8): 1901-1907.
doi: 10.11999/JEIT151091
Abstract:
A dynamic pilot allocation scheme is proposed in case of the pilot contamination existing in massive MIMO system. Based on the signal to interference difference between the aim cell user and the interference cell user, the interference cell is divided intoUin and Uout. Specifically, in order to improve the average downlink achievable sum rates, the users in theUin are operated with the optimal pilot allocation, and the users in theUout are operated with the random pilot allocation. Simultaneously, the proposed pilot allocation scheme is further optimized with an extral set of orthogonal pilots. Simulation results show that the proposed dynamic pilot allocation scheme can enhance the downlink performance of the massive MIMO system effectively.
A dynamic pilot allocation scheme is proposed in case of the pilot contamination existing in massive MIMO system. Based on the signal to interference difference between the aim cell user and the interference cell user, the interference cell is divided intoUin and Uout. Specifically, in order to improve the average downlink achievable sum rates, the users in theUin are operated with the optimal pilot allocation, and the users in theUout are operated with the random pilot allocation. Simultaneously, the proposed pilot allocation scheme is further optimized with an extral set of orthogonal pilots. Simulation results show that the proposed dynamic pilot allocation scheme can enhance the downlink performance of the massive MIMO system effectively.
2016, 38(8): 1908-1912.
doi: 10.11999/JEIT151161
Abstract:
This paper investigates transceiver design for dual-hop up-link multi-user Multiple-Input Multiple- Output (MIMO) system with Amplify-and-Forward (AF) relay to improve the performance of Bit Error Ratio (BER) under the imperfect Channel State Information (CSI). Since the antenna correlation at both ends of the channel and the channel estimation errors are taken into account, a robust transceiver design is proposed. The optimization issue is formed with Minimum Mean-Square Error (MMSE) rule firstly. Then the relaying matrix and the optimal source precoding matrix are optimized at the maximum power constraint of the transmitter and relay station subsequently. Finally, the equalizer is optimized by a gradient-based line search algorithm. Simulation results indicate that the proposed design approach achieves better robustness against antenna correlation and channel estimation errors than existing methods.
This paper investigates transceiver design for dual-hop up-link multi-user Multiple-Input Multiple- Output (MIMO) system with Amplify-and-Forward (AF) relay to improve the performance of Bit Error Ratio (BER) under the imperfect Channel State Information (CSI). Since the antenna correlation at both ends of the channel and the channel estimation errors are taken into account, a robust transceiver design is proposed. The optimization issue is formed with Minimum Mean-Square Error (MMSE) rule firstly. Then the relaying matrix and the optimal source precoding matrix are optimized at the maximum power constraint of the transmitter and relay station subsequently. Finally, the equalizer is optimized by a gradient-based line search algorithm. Simulation results indicate that the proposed design approach achieves better robustness against antenna correlation and channel estimation errors than existing methods.
2016, 38(8): 1913-1919.
doi: 10.11999/JEIT151236
Abstract:
The physical layer security performances of low-complexity opportunistic transmission strategy based on multiple antenna are investigated for cooperative adaptive decode-and-forward relaying system in Nakagami-m fading channels. To fully utilize the antenna diversity gain to improve the system security performance, the transmitting nodes apply the transmit antenna selection strategy, and the receiving nodes apply the maximal ratio combining strategy. The closed-form expressions of secrecy outage probability are derived, the asymptotic analysis of secrecy performance is further provided, and the secrecy diversity order are also obtained. Simulation results verify the correctness of theoretical analysis and identify the effects of several system parameters on the secrecy performance of the opportunistic transmission strategy. It is shown that the system secrecy performance can be greatly improved by increasing the number of antennas at the legitimate nodes and increasing the Nakagami fading channel parameters of legitimate channels.
The physical layer security performances of low-complexity opportunistic transmission strategy based on multiple antenna are investigated for cooperative adaptive decode-and-forward relaying system in Nakagami-m fading channels. To fully utilize the antenna diversity gain to improve the system security performance, the transmitting nodes apply the transmit antenna selection strategy, and the receiving nodes apply the maximal ratio combining strategy. The closed-form expressions of secrecy outage probability are derived, the asymptotic analysis of secrecy performance is further provided, and the secrecy diversity order are also obtained. Simulation results verify the correctness of theoretical analysis and identify the effects of several system parameters on the secrecy performance of the opportunistic transmission strategy. It is shown that the system secrecy performance can be greatly improved by increasing the number of antennas at the legitimate nodes and increasing the Nakagami fading channel parameters of legitimate channels.
2016, 38(8): 1920-1927.
doi: 10.11999/JEIT151158
Abstract:
Multiple-Input-Multiple-Output (MIMO) under water acoustic communication is capable of improving the channel capacity in extremely limited bandwidth. However, the performance of traditional channel estimation algorithms, such as Least Squares (LS) method, Compressed Sensing (CS) method decreases rapidly because of the simultaneous presence of the Co-channel Interference (CoI) and multipath. As the sparse multipath structures between adjacent data blocks exhibit temporal correlation features, in this paper, the temporal correlation of sparse multipath structures is exploited to establish temporal joint sparse MIMO channel estimation model, and the Simultaneous Orthogonal Matching Pursuit (SOMP) algorithm is utilized for compressed sensing estimation of MIMO channels. Simulation and sea trial results validate the effectiveness of the proposed method.
Multiple-Input-Multiple-Output (MIMO) under water acoustic communication is capable of improving the channel capacity in extremely limited bandwidth. However, the performance of traditional channel estimation algorithms, such as Least Squares (LS) method, Compressed Sensing (CS) method decreases rapidly because of the simultaneous presence of the Co-channel Interference (CoI) and multipath. As the sparse multipath structures between adjacent data blocks exhibit temporal correlation features, in this paper, the temporal correlation of sparse multipath structures is exploited to establish temporal joint sparse MIMO channel estimation model, and the Simultaneous Orthogonal Matching Pursuit (SOMP) algorithm is utilized for compressed sensing estimation of MIMO channels. Simulation and sea trial results validate the effectiveness of the proposed method.
2016, 38(8): 1928-1934.
doi: 10.11999/JEIT151255
Abstract:
This paper shows the rebound attack on the Feistel-SPS structure, which has the Feistel network with a Substitution-Permutation-Substitution (SPS) round function. A 6-round known-key truncated differential distinguisher is obtained by studying the diffusion properties of differences. Based on the distinguisher, a near- collision attack on the compression functions of this structure embedding the Matyas-Meyer-Oseas (MMO) and Miyaguchi-Preneel (MP) modes is given. Besides, the 6-round distinguisher is extended and a 7-round truncated differential path is constructed to get a 7-round truncated differential distinguisher of the compression function for the two modes mentioned before.
This paper shows the rebound attack on the Feistel-SPS structure, which has the Feistel network with a Substitution-Permutation-Substitution (SPS) round function. A 6-round known-key truncated differential distinguisher is obtained by studying the diffusion properties of differences. Based on the distinguisher, a near- collision attack on the compression functions of this structure embedding the Matyas-Meyer-Oseas (MMO) and Miyaguchi-Preneel (MP) modes is given. Besides, the 6-round distinguisher is extended and a 7-round truncated differential path is constructed to get a 7-round truncated differential distinguisher of the compression function for the two modes mentioned before.
2016, 38(8): 1935-1941.
doi: 10.11999/JEIT151212
Abstract:
Dataspace is a new type of data management, which can manage the mass, heterogeneous, and dynamic data in a pay-as-you-go fashion. However, it is difficult to construct an effective access control mechanism in dataspace environment, because of the data dynamic evolution, the fine-grained and extremely loose data description. A fine-grained and dynamic access control mechanism supporting secure updates is presented in this paper for very loosely structured data model which is commonly used in dataspace. Firstly, a set of update operations are defined for modifying data in the dataspace, and the mapping functions are provided for mapping the updates data into relational databases. Secondly, the fine-grained access control rule supporting secure updates is given, and the consistency of the conversion between this rule and relational database access control rule is analyzed. Thirdly, an access request rewriting algorithm, which is sound and complete, is also presented for dynamically controlling read/write access to the data. The algorithm retrieves the related access control rules based on user's access request, and then rewrites the request by utilizing the relevant authority. Finally, the validity of the work in this paper is proved by the theory and the experiment.
Dataspace is a new type of data management, which can manage the mass, heterogeneous, and dynamic data in a pay-as-you-go fashion. However, it is difficult to construct an effective access control mechanism in dataspace environment, because of the data dynamic evolution, the fine-grained and extremely loose data description. A fine-grained and dynamic access control mechanism supporting secure updates is presented in this paper for very loosely structured data model which is commonly used in dataspace. Firstly, a set of update operations are defined for modifying data in the dataspace, and the mapping functions are provided for mapping the updates data into relational databases. Secondly, the fine-grained access control rule supporting secure updates is given, and the consistency of the conversion between this rule and relational database access control rule is analyzed. Thirdly, an access request rewriting algorithm, which is sound and complete, is also presented for dynamically controlling read/write access to the data. The algorithm retrieves the related access control rules based on user's access request, and then rewrites the request by utilizing the relevant authority. Finally, the validity of the work in this paper is proved by the theory and the experiment.
2016, 38(8): 1942-1949.
doi: 10.11999/JEIT151335
Abstract:
In colored-loading Knowledge Aided STAP (KA-STAP) techniques, the colored-loading factor should be determined according to the performance of the a priori information. The existing Pre-Whitening (PW) colored-loading factor optimization method can not evaluate the accuracy degree of the a priori information of the Cell Under Test (CUT), which makes it not robust to the situation where a priori information for each range bin is different. In this paper, a colored-loading factor optimization method, CUT information involved PW (CPW), is proposed to improve the performance of PW method. In CPW, partial training samples are utilized to evaluate the pre-whitening ability of the colored-loading matrix of CUT. At the same time, non-uniqueness problem of the optimization result of PW is also solved. Simulations are conducted to discuss the performance of CPW under different sample support conditions and different a priori information performance situations. Simulation results demonstrate the effectiveness and robustness of the proposed CPW approach.
In colored-loading Knowledge Aided STAP (KA-STAP) techniques, the colored-loading factor should be determined according to the performance of the a priori information. The existing Pre-Whitening (PW) colored-loading factor optimization method can not evaluate the accuracy degree of the a priori information of the Cell Under Test (CUT), which makes it not robust to the situation where a priori information for each range bin is different. In this paper, a colored-loading factor optimization method, CUT information involved PW (CPW), is proposed to improve the performance of PW method. In CPW, partial training samples are utilized to evaluate the pre-whitening ability of the colored-loading matrix of CUT. At the same time, non-uniqueness problem of the optimization result of PW is also solved. Simulations are conducted to discuss the performance of CPW under different sample support conditions and different a priori information performance situations. Simulation results demonstrate the effectiveness and robustness of the proposed CPW approach.
2016, 38(8): 1950-1955.
doi: 10.11999/JEIT151135
Abstract:
In this paper, a novel algorithm named Compact Sensing Matrix Pursuit (CSMP) is proposed to deal with the high coherence problem encountered in the compressed sensing based radar system with high resolution. The CSMP algorithm is applied to the two dimensional Direction Of Arrival (DOA) estimation of cross-array. The simulation results show that the resolution can be increased largely compared with the MUltiple SIgnal Classification (MUSIC) algorithm, Subspace Pursuit (SP), Basis Pursuit (BP), and the Sparse Bayesian Learning (SBL) algorithms.
In this paper, a novel algorithm named Compact Sensing Matrix Pursuit (CSMP) is proposed to deal with the high coherence problem encountered in the compressed sensing based radar system with high resolution. The CSMP algorithm is applied to the two dimensional Direction Of Arrival (DOA) estimation of cross-array. The simulation results show that the resolution can be increased largely compared with the MUltiple SIgnal Classification (MUSIC) algorithm, Subspace Pursuit (SP), Basis Pursuit (BP), and the Sparse Bayesian Learning (SBL) algorithms.
2016, 38(8): 1956-1961.
doi: 10.11999/JEIT151261
Abstract:
Based on the fact that anisotropy characteristic represents some intrinsic properties of scattering centers, an algorithm for extracting anisotropy characteristic of scattering centers is proposed when wide-angle SAR is used. Firstly, the components of the echo from a single scattering center are analyzed based on attributed scattering center model. Secondly, the estimation of anisotropy characteristic is transformed into an inverse issue of solving variation of single scattering centers amplitude by using identity matrix as orthonormal basis. Finally, the inverse issue is solved under the constraints that the amplitude of scattering centers should be real and continuous. As a result, anisotropy characteristic is extracted from the solution of the inverse issue. Estimation results of both Matlab simulated and electromagnetic computation data validate that the proposed algorithm is not only precise but also robust. Moreover, the proposed method is more efficient compared with traditional methods.
Based on the fact that anisotropy characteristic represents some intrinsic properties of scattering centers, an algorithm for extracting anisotropy characteristic of scattering centers is proposed when wide-angle SAR is used. Firstly, the components of the echo from a single scattering center are analyzed based on attributed scattering center model. Secondly, the estimation of anisotropy characteristic is transformed into an inverse issue of solving variation of single scattering centers amplitude by using identity matrix as orthonormal basis. Finally, the inverse issue is solved under the constraints that the amplitude of scattering centers should be real and continuous. As a result, anisotropy characteristic is extracted from the solution of the inverse issue. Estimation results of both Matlab simulated and electromagnetic computation data validate that the proposed algorithm is not only precise but also robust. Moreover, the proposed method is more efficient compared with traditional methods.
2016, 38(8): 1962-1968.
doi: 10.11999/JEIT151231
Abstract:
To compensate the translation motion of precession ballistic targets, the characters of translation and precession are analyzed and it is found out that the precession has the center symmetry character. Using this character, a method based on conjugate multiplication of center symmetry data is proposed to estimate the translation motion of precession ballistic target. The precession motion is canceled out due to the conjugate multiplication and the translation motion parameters can be estimated by the spectral peak position of the conjugate multiplied signal. Therefore, the translation motion can be compensated based on the estimated translation motion parameters. Simulation results verify the validity of the proposed algorithm.
To compensate the translation motion of precession ballistic targets, the characters of translation and precession are analyzed and it is found out that the precession has the center symmetry character. Using this character, a method based on conjugate multiplication of center symmetry data is proposed to estimate the translation motion of precession ballistic target. The precession motion is canceled out due to the conjugate multiplication and the translation motion parameters can be estimated by the spectral peak position of the conjugate multiplied signal. Therefore, the translation motion can be compensated based on the estimated translation motion parameters. Simulation results verify the validity of the proposed algorithm.
2016, 38(8): 1969-1976.
doi: 10.11999/JEIT151194
Abstract:
Traditional adaptive detectors are mostly derived under the assumption of independent and identically distributed texture. However, the texture correlation along the range cell exists in real sea clutter datasets. A new coherent detector based on texture structure is proposed by adding the information of texture correlation into the likelihood ratio test. Based on the prior knowledge that the texture correlation along range is generated by the swell modulation, the number of range cells related to the texture of the Cell Under Test (CUT) is determined, and this number provides the information for the texture of CUT. Experimental results using real datasets show that the proposed detector has better performance in comparison with the optimal detector in compound-Gaussian clutter with inverse gamma texture.
Traditional adaptive detectors are mostly derived under the assumption of independent and identically distributed texture. However, the texture correlation along the range cell exists in real sea clutter datasets. A new coherent detector based on texture structure is proposed by adding the information of texture correlation into the likelihood ratio test. Based on the prior knowledge that the texture correlation along range is generated by the swell modulation, the number of range cells related to the texture of the Cell Under Test (CUT) is determined, and this number provides the information for the texture of CUT. Experimental results using real datasets show that the proposed detector has better performance in comparison with the optimal detector in compound-Gaussian clutter with inverse gamma texture.
2016, 38(8): 1977-1983.
doi: 10.11999/JEIT160016
Abstract:
In order to achieve the matched beams goal of the dual-band dual-polarized spaceborne precipitation radar, the solution that a shared aperture feeding array illuminates the parabolic cylindrical reflector is proposed. The shared aperture feeding array using of microstrip patches for Ku band and waveguide slots for Ka band is proposed. The interleaved layout is selected to configurate the shared aperture feeding array. The measured results reveal that the beam width and the beam point are similar with that of the Second Generation Precipitation Radar, which is supported by National Aeronautics and Space Administration. The second generation precipitation radar antenna is a reflector offset-fed by two separated arrays. The volume of the shared aperture feeding array is smaller than that of two separated arrays, which is more applicable to the satellites.
In order to achieve the matched beams goal of the dual-band dual-polarized spaceborne precipitation radar, the solution that a shared aperture feeding array illuminates the parabolic cylindrical reflector is proposed. The shared aperture feeding array using of microstrip patches for Ku band and waveguide slots for Ka band is proposed. The interleaved layout is selected to configurate the shared aperture feeding array. The measured results reveal that the beam width and the beam point are similar with that of the Second Generation Precipitation Radar, which is supported by National Aeronautics and Space Administration. The second generation precipitation radar antenna is a reflector offset-fed by two separated arrays. The volume of the shared aperture feeding array is smaller than that of two separated arrays, which is more applicable to the satellites.
2016, 38(8): 1984-1990.
doi: 10.11999/JEIT151139
Abstract:
Since spoofing send similar navigation satellites C/A code, this kind of interferences can lead to the receiver misled easily, receiver will provide a wrong location information. Considering both navigation signal and spoofing with same navigation signal structure has high self-coherence characteristic, and the power of spoofing is a little higher than authentic signal, this paper proposes a blind spoofing suppression method based on array antenna. Firstly, a cross-correlation matrix is obtained by cross-correlation processing between receive data from multiple antennas and its own delayed reference data. Secondly, the orthogonal projection matrix of interference can be got by the cross-correlation matrix. Finally, the eigenvector corresponding to the biggest eigenvalue from the projected cross-correlation matrix is taken as the optimal weight. Without having known the direction of authentic signal and interference, this method avoid despreading by searching each satellite C/A code sequence before receive. Experimental results show that the array beam can effectively suppress interference with high SNR and receiver performance has no effect by the spoofing.
Since spoofing send similar navigation satellites C/A code, this kind of interferences can lead to the receiver misled easily, receiver will provide a wrong location information. Considering both navigation signal and spoofing with same navigation signal structure has high self-coherence characteristic, and the power of spoofing is a little higher than authentic signal, this paper proposes a blind spoofing suppression method based on array antenna. Firstly, a cross-correlation matrix is obtained by cross-correlation processing between receive data from multiple antennas and its own delayed reference data. Secondly, the orthogonal projection matrix of interference can be got by the cross-correlation matrix. Finally, the eigenvector corresponding to the biggest eigenvalue from the projected cross-correlation matrix is taken as the optimal weight. Without having known the direction of authentic signal and interference, this method avoid despreading by searching each satellite C/A code sequence before receive. Experimental results show that the array beam can effectively suppress interference with high SNR and receiver performance has no effect by the spoofing.
2016, 38(8): 1991-1998.
doi: 10.11999/JEIT151156
Abstract:
Phase retrieval is an issue that tries to recover an image from its Fourier magnitude measurements. Since the Fourier magnitude measurements contain less information, the traditional phase retrieval algorithms can not reconstruct the image efficiently under the scenario that the oversampling ratio is relatively low. Therefore, how to use the suitable image priors to improve the reconstruction quality of the image is the key issue. In this paper, the cartoon-texture model is utilized for phase retrieval algorithm. Two sparse representation methods including both Total Variation (TV) and Dual-Tree Complex Wavelet Transform (DTCWT) are exploited to decompose the image into two parts, namely the cartoon component and the texture component. Moreover, Alternating Direction Method of Multipliers (ADMM) is exploited to solve the corresponding problem. The experimental results show that the proposed algorithm can effectively improve the quality of image reconstruction.
Phase retrieval is an issue that tries to recover an image from its Fourier magnitude measurements. Since the Fourier magnitude measurements contain less information, the traditional phase retrieval algorithms can not reconstruct the image efficiently under the scenario that the oversampling ratio is relatively low. Therefore, how to use the suitable image priors to improve the reconstruction quality of the image is the key issue. In this paper, the cartoon-texture model is utilized for phase retrieval algorithm. Two sparse representation methods including both Total Variation (TV) and Dual-Tree Complex Wavelet Transform (DTCWT) are exploited to decompose the image into two parts, namely the cartoon component and the texture component. Moreover, Alternating Direction Method of Multipliers (ADMM) is exploited to solve the corresponding problem. The experimental results show that the proposed algorithm can effectively improve the quality of image reconstruction.
2016, 38(8): 1999-2006.
doi: 10.11999/JEIT151227
Abstract:
In order to sufficiently exploit the image information residing in the residual image for boosting the denoising performance of the Non-local Means (NLM) algorithm, a novel multi-stage residual filtering method is proposed. Firstly, the Non-Local Means algorithm is applied to a noisy image to produce an initial denoised image and a weight distributing matrix. Then the fixed-weight NLM algorithm is applied to the residual image followed by a Gaussian filtering process, which can extract the image content out from the residual as a compensation image. The compensation image is then added back to the denoised image to generate an enhanced restored image. An iterative scheme, whose principle and feasibility are derived and proved theoretically, is developed for the above filtering procedure; meanwhile a novel stopping criterion with no reference image required is proposed to determine the optimal number of iterations adaptively. Experimental results demonstrate that the proposed stopping criterion behaves similarly as the PSNR rule, and compared with the original NLM approach, the proposed method can boost the denoising performance significantly with 1.2 dB PSNR gains achieved on average and more detail information preserved, while the computational complexity is not apparently increased.
In order to sufficiently exploit the image information residing in the residual image for boosting the denoising performance of the Non-local Means (NLM) algorithm, a novel multi-stage residual filtering method is proposed. Firstly, the Non-Local Means algorithm is applied to a noisy image to produce an initial denoised image and a weight distributing matrix. Then the fixed-weight NLM algorithm is applied to the residual image followed by a Gaussian filtering process, which can extract the image content out from the residual as a compensation image. The compensation image is then added back to the denoised image to generate an enhanced restored image. An iterative scheme, whose principle and feasibility are derived and proved theoretically, is developed for the above filtering procedure; meanwhile a novel stopping criterion with no reference image required is proposed to determine the optimal number of iterations adaptively. Experimental results demonstrate that the proposed stopping criterion behaves similarly as the PSNR rule, and compared with the original NLM approach, the proposed method can boost the denoising performance significantly with 1.2 dB PSNR gains achieved on average and more detail information preserved, while the computational complexity is not apparently increased.
2016, 38(8): 2007-2013.
doi: 10.11999/JEIT151217
Abstract:
To overcome the limitation that the alpha stable distributed variable possesses infinite second-order moment, a novel generalized correntropy is defined and the bounded property of the generalized correntropy for the symmetric alpha stable variable is proved. Furthermore, a novel minimum generalized correntropy criterion based DOA estimation method for impulsive noise is proposed, and an iterative optimization algorithm is presented, the convergence of which is analyzed by simulation experiments. The simulation results demonstrate that the proposed method can get better estimation results than the fractional lower order moments based FLOM-MUSIC, the correntropy-like based CRCO-MUSIC and the lp norm based ACO-MUSIC methods, especially in the highly impulsive noise environments.
To overcome the limitation that the alpha stable distributed variable possesses infinite second-order moment, a novel generalized correntropy is defined and the bounded property of the generalized correntropy for the symmetric alpha stable variable is proved. Furthermore, a novel minimum generalized correntropy criterion based DOA estimation method for impulsive noise is proposed, and an iterative optimization algorithm is presented, the convergence of which is analyzed by simulation experiments. The simulation results demonstrate that the proposed method can get better estimation results than the fractional lower order moments based FLOM-MUSIC, the correntropy-like based CRCO-MUSIC and the lp norm based ACO-MUSIC methods, especially in the highly impulsive noise environments.
2016, 38(8): 2014-2019.
doi: 10.11999/JEIT151213
Abstract:
Traditional double constraint robust Capon beamforming algorithm uses Newton iterative method for solving the optimal loading, presenting the problems of low accuracy and large amount of computation. An improved Double Constraint Robust Capon Beamforming (DCRCB) algorithm is proposed in this paper. The algorithm reconstructs the signal convariance matrix, and by optimizing the projection of signal steering vector onto the noise subspace, it projects the reconstructed interference-plus-noise convariance matrix onto the noise subspace, obtaining the double constraint algorithm model based on the noise subspace. For the norm constraint is accessorial, the algorithm model can be converted into a single constraint issue and be solved into an analytical expression of optimal diagonal loading finally. Simulation results show that the improved algorithm can optimize the side lobe by adjusting the beam width of the main lobe, improve effectively the anti-vector error robustness, and reduce the amount of computation.
Traditional double constraint robust Capon beamforming algorithm uses Newton iterative method for solving the optimal loading, presenting the problems of low accuracy and large amount of computation. An improved Double Constraint Robust Capon Beamforming (DCRCB) algorithm is proposed in this paper. The algorithm reconstructs the signal convariance matrix, and by optimizing the projection of signal steering vector onto the noise subspace, it projects the reconstructed interference-plus-noise convariance matrix onto the noise subspace, obtaining the double constraint algorithm model based on the noise subspace. For the norm constraint is accessorial, the algorithm model can be converted into a single constraint issue and be solved into an analytical expression of optimal diagonal loading finally. Simulation results show that the improved algorithm can optimize the side lobe by adjusting the beam width of the main lobe, improve effectively the anti-vector error robustness, and reduce the amount of computation.
2016, 38(8): 2020-2026.
doi: 10.11999/JEIT151302
Abstract:
Existing dual microphone Voice Activity Detection (VAD) algorithms use normally a fixed threshold. The fixed threshold can not provide an accurate VAD under various noise environments. In such case, it causes voice quality degradation, particularly in handset applications. This paper proposes a new VAD algorithm based on Neural Network (NN). Both sub-band power level difference and inter-microphone cross correlation are used as features. Then the NN based VAD is combined with the method of inter-microphone signal power ratio to get a new voice and noise activity detection algorithm. Furthermore, the algorithm is used into noise suppression in handset to avoid performance degradation caused by VAD misjudgment. Experimental results show that the proposed method provides better noise suppression performance and lower speech distortion compared to the existing method.
Existing dual microphone Voice Activity Detection (VAD) algorithms use normally a fixed threshold. The fixed threshold can not provide an accurate VAD under various noise environments. In such case, it causes voice quality degradation, particularly in handset applications. This paper proposes a new VAD algorithm based on Neural Network (NN). Both sub-band power level difference and inter-microphone cross correlation are used as features. Then the NN based VAD is combined with the method of inter-microphone signal power ratio to get a new voice and noise activity detection algorithm. Furthermore, the algorithm is used into noise suppression in handset to avoid performance degradation caused by VAD misjudgment. Experimental results show that the proposed method provides better noise suppression performance and lower speech distortion compared to the existing method.
2016, 38(8): 2027-2032.
doi: 10.11999/JEIT151082
Abstract:
In order to improve the stability and accuracy of the object tracking under nonlinear conditions, an object tracking algorithm based on Partial Least Squares (PLS) representation and stochastic gradient object optimization tracking is proposed. In this method, object tracking is defined as an optimization task that minimizes the representation error and classification loss. Firstly, it expresses object appearance and background information by PLS theory, learns multiple low dimensional and discriminative subspaces to describe the nonlinear appearance changes of the object. Then, a joint optimization objective function based on deterministic search mechanism is proposed. Furthermore, an stochastic gradient classifier based on incremental features updating is proposed, and make sure that it can achieve a stable tracking. Experiments show favorable performance of the proposed algorithm on sequences where the targets undergo a variety complex changes on foreground and background.
In order to improve the stability and accuracy of the object tracking under nonlinear conditions, an object tracking algorithm based on Partial Least Squares (PLS) representation and stochastic gradient object optimization tracking is proposed. In this method, object tracking is defined as an optimization task that minimizes the representation error and classification loss. Firstly, it expresses object appearance and background information by PLS theory, learns multiple low dimensional and discriminative subspaces to describe the nonlinear appearance changes of the object. Then, a joint optimization objective function based on deterministic search mechanism is proposed. Furthermore, an stochastic gradient classifier based on incremental features updating is proposed, and make sure that it can achieve a stable tracking. Experiments show favorable performance of the proposed algorithm on sequences where the targets undergo a variety complex changes on foreground and background.
2016, 38(8): 2033-2039.
doi: 10.11999/JEIT151137
Abstract:
Process mining is a technology which can extract non-trivial and useful information from airport event logs. However, the airport event logs are always on a detailed level of abstraction, which may not be in line with the expected abstract level of an analyst. Process models generated by these event logs are always spaghetti-like and too hard to comprehend. An approach to overcome this issue is to group low-level events into clusters, which represent the execution of a higher-level activity in the process model. Therefore, this paper presents a new activity mining method which is based on RankClus algorithm to generate activity clusters integrated with ranking. On this basis, the activity-clustered model which is easier to comprehend can be constructed. The experiment results show that this activity-clustered model, which shares a similar level of conformance with the meta model, is significantly less complex.
Process mining is a technology which can extract non-trivial and useful information from airport event logs. However, the airport event logs are always on a detailed level of abstraction, which may not be in line with the expected abstract level of an analyst. Process models generated by these event logs are always spaghetti-like and too hard to comprehend. An approach to overcome this issue is to group low-level events into clusters, which represent the execution of a higher-level activity in the process model. Therefore, this paper presents a new activity mining method which is based on RankClus algorithm to generate activity clusters integrated with ranking. On this basis, the activity-clustered model which is easier to comprehend can be constructed. The experiment results show that this activity-clustered model, which shares a similar level of conformance with the meta model, is significantly less complex.
2016, 38(8): 2040-2046.
doi: 10.11999/JEIT151331
Abstract:
In order to realize intelligent and autonomous navigation of vehicles, a method of spatial place representation based on biological place cells firing is proposed. A relationship is built that the spatial location of vehicle is corresponding with the distances between the vehicle and different landmarks, and a map of place cells is constructed in two different ways of space coverage. Then using the real-time distances sensed inspires place cells firing so as to estimate the location of the vehicle. An analysis is carried out about different parameters in the model influence on spatial location expression and the performance of positioning. The simulation results show that both two ways can realize location expression and autonomous positioning through using the map of place cells. The way that the space is separated equally is influenced by the distance interval greatly, and the way of building the map of place cells through exploring randomly, starting off with the perspective of sensing spatial location autonomously, can finish spatial location expression and autonomous positioning better through choosing appropriate interval and trained time.
In order to realize intelligent and autonomous navigation of vehicles, a method of spatial place representation based on biological place cells firing is proposed. A relationship is built that the spatial location of vehicle is corresponding with the distances between the vehicle and different landmarks, and a map of place cells is constructed in two different ways of space coverage. Then using the real-time distances sensed inspires place cells firing so as to estimate the location of the vehicle. An analysis is carried out about different parameters in the model influence on spatial location expression and the performance of positioning. The simulation results show that both two ways can realize location expression and autonomous positioning through using the map of place cells. The way that the space is separated equally is influenced by the distance interval greatly, and the way of building the map of place cells through exploring randomly, starting off with the perspective of sensing spatial location autonomously, can finish spatial location expression and autonomous positioning better through choosing appropriate interval and trained time.
2016, 38(8): 2047-2053.
doi: 10.11999/JEIT151237
Abstract:
To improve distribution and convergence of the obtained solution set in constrained multi-objective optimization problems, this paper presents a constrained multi-objective optimization algorithm based on adaptive truncation strategy. Firstly, through the proposed truncation strategy, the Pareto optimal solutions and the infeasible solutions with low constraint violation and good objective function values are retained to improve diversity. Besides, both diversity and convergence are coordinated. Secondly, the exponential variation is added for further enhancing the local exploitation ability after mutation and crossover operation. Finally, the improved crowding density estimation chooses a part of the Pareto optimal individuals and the near individuals to take part in the calculation, thus it not only assesses the distribution of the solution set more accurately, but also reduces the computational quantity. The comparative experiment results with another four excellent constrained multi- objective algorithms on the standard constrained multi-objective optimization problems (CTP series) show that diversity and convergence of the proposed algorithm are improved, and it has certain advantages compared with these algorithms.
To improve distribution and convergence of the obtained solution set in constrained multi-objective optimization problems, this paper presents a constrained multi-objective optimization algorithm based on adaptive truncation strategy. Firstly, through the proposed truncation strategy, the Pareto optimal solutions and the infeasible solutions with low constraint violation and good objective function values are retained to improve diversity. Besides, both diversity and convergence are coordinated. Secondly, the exponential variation is added for further enhancing the local exploitation ability after mutation and crossover operation. Finally, the improved crowding density estimation chooses a part of the Pareto optimal individuals and the near individuals to take part in the calculation, thus it not only assesses the distribution of the solution set more accurately, but also reduces the computational quantity. The comparative experiment results with another four excellent constrained multi- objective algorithms on the standard constrained multi-objective optimization problems (CTP series) show that diversity and convergence of the proposed algorithm are improved, and it has certain advantages compared with these algorithms.
2016, 38(8): 2054-2061.
doi: 10.11999/JEIT151209
Abstract:
Conventional fuzzy system modeling methods essentially belong to the single-view learning modality. In multi-view-oriented data scenarios, they can only cope with each view separately, which is prone to incurring their unsatisfactory generalization performance. In response to such problem, the fuzzy system modeling method with the ability of multi-view learning is pursued. To this end, based on the classic L2 norm Takagi-Sugeno-Kang (TSK) fuzzy system, by means of the collaborative learning items qualified for multi-view learning, the core Multi-View TSK Fuzzy System (MV-TSK-FS) modeling method is presented. MV-TSK-FS can not only effectively utilize the independent components composed of the characteristics affiliated to each view, but also take full advantage of the potential information occurred by the interrelated effects among views, which eventually facilitates its relatively strong generalization ability. The experimental results performed on both synthetic and real-life datasets indicate that, compared with some traditional single-view methods, this propounded multi-view fuzzy modeling system owns preferable applicability as well as generalization.
Conventional fuzzy system modeling methods essentially belong to the single-view learning modality. In multi-view-oriented data scenarios, they can only cope with each view separately, which is prone to incurring their unsatisfactory generalization performance. In response to such problem, the fuzzy system modeling method with the ability of multi-view learning is pursued. To this end, based on the classic L2 norm Takagi-Sugeno-Kang (TSK) fuzzy system, by means of the collaborative learning items qualified for multi-view learning, the core Multi-View TSK Fuzzy System (MV-TSK-FS) modeling method is presented. MV-TSK-FS can not only effectively utilize the independent components composed of the characteristics affiliated to each view, but also take full advantage of the potential information occurred by the interrelated effects among views, which eventually facilitates its relatively strong generalization ability. The experimental results performed on both synthetic and real-life datasets indicate that, compared with some traditional single-view methods, this propounded multi-view fuzzy modeling system owns preferable applicability as well as generalization.
2016, 38(8): 2062-2067.
doi: 10.11999/JEIT160101
Abstract:
Based on adaptive control and Lyapunov stability theory, a novel adaptive Generalized Combination Complex Synchronization (GCCS) scheme is proposed for nonidentical hyperchaotic complex systems with unknown parameters. Firstly, the definition of GCCS is presented, and synchronization of drive-response systems is transformed to the zero solution analysis of the error dynamical system. Secondly, a nonlinear feedback controller and parameter update laws are theoretically designed, wherein error feedback gains are introduced to control synchronization speed. Finally, GCCS among the hyperchaotic complex Lorenz system, complex Chen system, and complex L system is carried out to verify the correctness and effectiveness of the proposed scheme by numerical simulation.
Based on adaptive control and Lyapunov stability theory, a novel adaptive Generalized Combination Complex Synchronization (GCCS) scheme is proposed for nonidentical hyperchaotic complex systems with unknown parameters. Firstly, the definition of GCCS is presented, and synchronization of drive-response systems is transformed to the zero solution analysis of the error dynamical system. Secondly, a nonlinear feedback controller and parameter update laws are theoretically designed, wherein error feedback gains are introduced to control synchronization speed. Finally, GCCS among the hyperchaotic complex Lorenz system, complex Chen system, and complex L system is carried out to verify the correctness and effectiveness of the proposed scheme by numerical simulation.
2016, 38(8): 2068-2075.
doi: 10.11999/JEIT151346
Abstract:
With the increasing amount of image data, the image retrieval methods have several drawbacks, such as the low expression ability of visual feature, high dimension of feature, low precision of image retrieval and so on. To solve these problems, a learning method of binary hashing based on deep convolutional neural networks is proposed, which can be used for large-scale image retrieval. The basic idea is to add a hash layer into the deep learning framework and to learn simultaneously image features and hash functions should satisfy independence and quantization error minimized. First, convolutional neural network is employed to learn the intrinsic implications of training images so as to improve the distinguish ability and expression ability of visual feature. Second, the visual feature is putted into the hash layer, in which hash functions are learned. And the learned hash functions should satisfy the classification error and quantization error minimized and the independence constraint. Finally, an input image is given, hash codes are generated by the output of the hash layer of the proposed framework and large scale image retrieval can be accomplished in low-dimensional hamming space. Experimental results on the three benchmark datasets show that the binary hash codes generated by the proposed method has superior performance gains over other state-of-the-art hashing methods.
With the increasing amount of image data, the image retrieval methods have several drawbacks, such as the low expression ability of visual feature, high dimension of feature, low precision of image retrieval and so on. To solve these problems, a learning method of binary hashing based on deep convolutional neural networks is proposed, which can be used for large-scale image retrieval. The basic idea is to add a hash layer into the deep learning framework and to learn simultaneously image features and hash functions should satisfy independence and quantization error minimized. First, convolutional neural network is employed to learn the intrinsic implications of training images so as to improve the distinguish ability and expression ability of visual feature. Second, the visual feature is putted into the hash layer, in which hash functions are learned. And the learned hash functions should satisfy the classification error and quantization error minimized and the independence constraint. Finally, an input image is given, hash codes are generated by the output of the hash layer of the proposed framework and large scale image retrieval can be accomplished in low-dimensional hamming space. Experimental results on the three benchmark datasets show that the binary hash codes generated by the proposed method has superior performance gains over other state-of-the-art hashing methods.
2016, 38(8): 2076-2084.
doi: 10.11999/JEIT151132
Abstract:
For solving the clustering problem of transfer learning, a new algorithm called Transfer Affinity Propagation clustering algorithm is proposed based on Kullback-Leiber distance (TAP_KL). Based on the probabilistic framework, a new interpretation of the objective function of Affinity Propagation (AP) clustering algorithm is proposed. By leveraging Kullback-Leiber distance which is usually used in information theory, TAP_KL measures the similarity relationship between source data and target data. Moreover, TAP_KL algorithm can embed the similarity relationship to the calculation of similarity matrix of target data. Thus, the optimization framework of AP can be directly used to optimize the new target function of TAP_KL. In this case, TAP_KL builds a simple algorithm framework to solve the transfer clustering problem, in which the algorithm just needs to modify the similarity matrix to solve the transfer clustering problem. The experimental results based on both 4 datasets show the effectiveness of the proposed algorithm TAP_KL.
For solving the clustering problem of transfer learning, a new algorithm called Transfer Affinity Propagation clustering algorithm is proposed based on Kullback-Leiber distance (TAP_KL). Based on the probabilistic framework, a new interpretation of the objective function of Affinity Propagation (AP) clustering algorithm is proposed. By leveraging Kullback-Leiber distance which is usually used in information theory, TAP_KL measures the similarity relationship between source data and target data. Moreover, TAP_KL algorithm can embed the similarity relationship to the calculation of similarity matrix of target data. Thus, the optimization framework of AP can be directly used to optimize the new target function of TAP_KL. In this case, TAP_KL builds a simple algorithm framework to solve the transfer clustering problem, in which the algorithm just needs to modify the similarity matrix to solve the transfer clustering problem. The experimental results based on both 4 datasets show the effectiveness of the proposed algorithm TAP_KL.
2016, 38(8): 2085-2090.
doi: 10.11999/JEIT151218
Abstract:
Searching for wavelet invariants is a key issue in multiresolution analysis. On the other hand,the method of moment invariants is fully developed both in the theory and the practice. A kind of wavelet moment invariants are given based on the image invariant moments and wavelet appr-oximation coefficients from the limited number of scales of the image. A fairy complete result on theory and experiment is obtained. At the same time, some problems of the theory and method are pointed out in the practical application.Finally, the application relationship between multi-scale analysis and invariant moment is briefly described.
Searching for wavelet invariants is a key issue in multiresolution analysis. On the other hand,the method of moment invariants is fully developed both in the theory and the practice. A kind of wavelet moment invariants are given based on the image invariant moments and wavelet appr-oximation coefficients from the limited number of scales of the image. A fairy complete result on theory and experiment is obtained. At the same time, some problems of the theory and method are pointed out in the practical application.Finally, the application relationship between multi-scale analysis and invariant moment is briefly described.
2016, 38(8): 2091-2098.
doi: 10.11999/JEIT151049
Abstract:
In order to implement the wisdom city planning and build perfect wisdom network, it is important to design the security Radio Frequency IDentification (RFID) protocol. A secure RFID ownership transfer protocol should be evaluated in terms of the security and privacy properties. In particular, there are two important privacy properties included forward untraceable and backward untraceable in the practical application of RFID system. In order to solve the various security and privacy problems, this paper enhances the unreasonable assumption that the attacker misses the key-update session in the definition of forward untraceable, then proposes the definition of strong forward untraceable. In addition, this paper designs the lightweight RFID ownership transfer protocol based on quadratic residues, and uses the enhanced model and definitions to formal prove the security and privacy properties. Moreover, the proof results not only show that the scheme resists against inner reader malicious im-personation attack, tracing attack, tag impersonation attack and desynchronization attack, but also formally prove that the proposed protocol meets strong forward untraceable and backward untraceable properties. In addition, the analysis results demonstrate that the protocol based on low-cost and high efficiency is superior to other protocols in the security and performance properties.
In order to implement the wisdom city planning and build perfect wisdom network, it is important to design the security Radio Frequency IDentification (RFID) protocol. A secure RFID ownership transfer protocol should be evaluated in terms of the security and privacy properties. In particular, there are two important privacy properties included forward untraceable and backward untraceable in the practical application of RFID system. In order to solve the various security and privacy problems, this paper enhances the unreasonable assumption that the attacker misses the key-update session in the definition of forward untraceable, then proposes the definition of strong forward untraceable. In addition, this paper designs the lightweight RFID ownership transfer protocol based on quadratic residues, and uses the enhanced model and definitions to formal prove the security and privacy properties. Moreover, the proof results not only show that the scheme resists against inner reader malicious im-personation attack, tracing attack, tag impersonation attack and desynchronization attack, but also formally prove that the proposed protocol meets strong forward untraceable and backward untraceable properties. In addition, the analysis results demonstrate that the protocol based on low-cost and high efficiency is superior to other protocols in the security and performance properties.
2016, 38(8): 2099-2106.
doi: 10.11999/JEIT151233
Abstract:
Stochastic computing is a special algorithm that performs mathematical operations with probabilistic values of bit streams rather than traditional deterministic values. The main advantage of stochastic computing is its great simplicity of hardware arithmetic units for mathematical operations to reduce the circuit cost. This paper discusses the principle of the stochastic computing and its main arithmetic logic. It analyzes a two-dimension state transition topology structure, and discusses the Gaussian function implementation method based on the two-dimension Finite State Machin (FSM). Then, a low cost stochastic radial basis function neural network model is proposed. Results from two pattern recognition tests show that the difference of the mean squared error between the stochastic network output value and the corresponding deterministic network output value can be less than 1.3%. FPGA implementation results show that the hardware resource requirement of the proposed stochastic hidden neuron is only 1.2% of the corresponding deterministic hidden neuron with the interpolated look-up table, and is 2.0% of the CORDIC algorithm. The accuracy, speed and power of the stochastic network can be tradeoff dynamically. This network is suitable for the low cost and low power applications like embedded, portable and wearable devices.
Stochastic computing is a special algorithm that performs mathematical operations with probabilistic values of bit streams rather than traditional deterministic values. The main advantage of stochastic computing is its great simplicity of hardware arithmetic units for mathematical operations to reduce the circuit cost. This paper discusses the principle of the stochastic computing and its main arithmetic logic. It analyzes a two-dimension state transition topology structure, and discusses the Gaussian function implementation method based on the two-dimension Finite State Machin (FSM). Then, a low cost stochastic radial basis function neural network model is proposed. Results from two pattern recognition tests show that the difference of the mean squared error between the stochastic network output value and the corresponding deterministic network output value can be less than 1.3%. FPGA implementation results show that the hardware resource requirement of the proposed stochastic hidden neuron is only 1.2% of the corresponding deterministic hidden neuron with the interpolated look-up table, and is 2.0% of the CORDIC algorithm. The accuracy, speed and power of the stochastic network can be tradeoff dynamically. This network is suitable for the low cost and low power applications like embedded, portable and wearable devices.
2016, 38(8): 2107-2112.
doi: 10.11999/JEIT151189
Abstract:
A new Virtual Least-Square Active Element Pattern Expansion (VLS-AEPE) method is presented in this paper, which considers each active element pattern of an unequally spaced array as the one radiated by some of equally spaced elements of a virtual array. Using the help of this method, the pattern of an unequally spaced array including mutual coupling can be efficiently calculated by FFT. In addition, this method is combined with the Genetic Algorithm (GA) to deal with the shaped pattern synthesis problem for unequally spaced linear arrays. Two synthesis experiments including the synthesis of flat-top pattern for an unequally spaced dipole array and the synthesis of cosec-squared pattern for an unequally spaced microstrip array are conducted to verify the effectiveness and advantages of the proposed algorithm.
A new Virtual Least-Square Active Element Pattern Expansion (VLS-AEPE) method is presented in this paper, which considers each active element pattern of an unequally spaced array as the one radiated by some of equally spaced elements of a virtual array. Using the help of this method, the pattern of an unequally spaced array including mutual coupling can be efficiently calculated by FFT. In addition, this method is combined with the Genetic Algorithm (GA) to deal with the shaped pattern synthesis problem for unequally spaced linear arrays. Two synthesis experiments including the synthesis of flat-top pattern for an unequally spaced dipole array and the synthesis of cosec-squared pattern for an unequally spaced microstrip array are conducted to verify the effectiveness and advantages of the proposed algorithm.
2016, 38(8): 2113-2121.
doi: 10.11999/JEIT151254
Abstract:
With the shrinking of feature size, soft errors due to Single Event Transient (SET) effect become the main reliability threat for aerospace deep sub-micron VLSI circuits, and the generation and propagation of SET pulse is also a hot issue in the study of soft error. Results of SET pulse propagation on logic chains show that the difference of rise and fall time of SET pulse can make the width of output pulse widened or lessened. The width and amplitude of SET pulse can determine whether it is electrically masked out. A four-value pulse parameters model is proposed to accurately characterize the shape of SET pulse, and then the LUT-based technique is combined with experiential equations to model the transmission process of SET. The proposed four-value pulse parameters model can model the effect of broadening or attenuation of SET pulse, and it has calculation precision improvement of 2.4% compared with single parameter model. This paper applies the graph-based error propagation probability analytic algorithm to estimate the logical masking in pulse propagation. The experimental results on ISCAS89 and ISCAS85 circuits show that the average deviation of this method and HSPICE simulation method is 4.12% and the calculation speed is 10000 times. This method can be used to analyze quickly the soft error rate of large scale integrated circuits.
With the shrinking of feature size, soft errors due to Single Event Transient (SET) effect become the main reliability threat for aerospace deep sub-micron VLSI circuits, and the generation and propagation of SET pulse is also a hot issue in the study of soft error. Results of SET pulse propagation on logic chains show that the difference of rise and fall time of SET pulse can make the width of output pulse widened or lessened. The width and amplitude of SET pulse can determine whether it is electrically masked out. A four-value pulse parameters model is proposed to accurately characterize the shape of SET pulse, and then the LUT-based technique is combined with experiential equations to model the transmission process of SET. The proposed four-value pulse parameters model can model the effect of broadening or attenuation of SET pulse, and it has calculation precision improvement of 2.4% compared with single parameter model. This paper applies the graph-based error propagation probability analytic algorithm to estimate the logical masking in pulse propagation. The experimental results on ISCAS89 and ISCAS85 circuits show that the average deviation of this method and HSPICE simulation method is 4.12% and the calculation speed is 10000 times. This method can be used to analyze quickly the soft error rate of large scale integrated circuits.
2016, 38(8): 2122-2128.
doi: 10.11999/JEIT151256
Abstract:
This paper presents a non-bandgap voltage reference, which contains a pre-regulated circuit with a super source follower. The pre-regulated circuit includes a super source follower, which attenuates the impedance from the supply of the core reference circuit to ground. In this way, the pre-regulated circuit provides a relative stable voltage for the core reference circuit, improving the Power Supply Rejection Ratio (PSRR) of the output voltage of the reference. The proposed reference circuit is implemented in standard 0.35 m CMOS process. Measured results show that the supply range is from 1.8 to 5 V and the quiescent current is only about 13 A at room temperature. The PSRR at low frequency achieves -100 dB and the PSRR below 1 kHz is better than -93 dB. The active area of the proposed reference is only 0.013 mm2.
This paper presents a non-bandgap voltage reference, which contains a pre-regulated circuit with a super source follower. The pre-regulated circuit includes a super source follower, which attenuates the impedance from the supply of the core reference circuit to ground. In this way, the pre-regulated circuit provides a relative stable voltage for the core reference circuit, improving the Power Supply Rejection Ratio (PSRR) of the output voltage of the reference. The proposed reference circuit is implemented in standard 0.35 m CMOS process. Measured results show that the supply range is from 1.8 to 5 V and the quiescent current is only about 13 A at room temperature. The PSRR at low frequency achieves -100 dB and the PSRR below 1 kHz is better than -93 dB. The active area of the proposed reference is only 0.013 mm2.
