Email alert
2000 Vol. 22, No. 5
column
Display Method:
2000, 22(5): 705-714.
Abstract:
In order to improve the tracking performance of local sensors in distributed multisensor data fusion systems,this paper discusses the state estimation technique of multisensor tracking with feedback information.Based on the single sensor Kalman filtering equations and distributed estimation solutions with feedback information,this paper presents the track fusion equations with and without feedback information in multicoordinate systems,and proves that two kinds of track fusion solutions are optimal and equivalent in the form of the theorem.The simulation results show that the multisensor data fusion system can not only increase the global estimation accuracy of target but also greatly improve the tracking performance of local nodes by using the feedback information.With respect to multiradar integration tracking,it is suitable for using 2～4 local sensors in data fusion system.
In order to improve the tracking performance of local sensors in distributed multisensor data fusion systems,this paper discusses the state estimation technique of multisensor tracking with feedback information.Based on the single sensor Kalman filtering equations and distributed estimation solutions with feedback information,this paper presents the track fusion equations with and without feedback information in multicoordinate systems,and proves that two kinds of track fusion solutions are optimal and equivalent in the form of the theorem.The simulation results show that the multisensor data fusion system can not only increase the global estimation accuracy of target but also greatly improve the tracking performance of local nodes by using the feedback information.With respect to multiradar integration tracking,it is suitable for using 2～4 local sensors in data fusion system.
2000, 22(5): 715-722.
Abstract:
This paper deals with estimating the rigid bodiesmotion parameters from token correspondence between two perspective views.Total least square(TLS)method is used to solve the linear equations of motion estimation.Then a minor component analysis (MCA)neuron is constructed to get the TLS solution.Finally a outline resisting version of the learning rule is developed by using the statistical approach.Comparative experiments have been made and results show that our robust rules improve the performance significantly when outliers are presented.
This paper deals with estimating the rigid bodiesmotion parameters from token correspondence between two perspective views.Total least square(TLS)method is used to solve the linear equations of motion estimation.Then a minor component analysis (MCA)neuron is constructed to get the TLS solution.Finally a outline resisting version of the learning rule is developed by using the statistical approach.Comparative experiments have been made and results show that our robust rules improve the performance significantly when outliers are presented.
2000, 22(5): 723-728.
Abstract:
The paper deals with the spectrum features,the fractal characteristics of the shipradiated noise,and researches the relationship between them;also,it takes advantage of them to distinguish the types of the ships.At the same time,a fuzzy neural network is used to be a classifier,and its particular superiorities are discussed.The classification processing experiment achieved satisfactory result.
The paper deals with the spectrum features,the fractal characteristics of the shipradiated noise,and researches the relationship between them;also,it takes advantage of them to distinguish the types of the ships.At the same time,a fuzzy neural network is used to be a classifier,and its particular superiorities are discussed.The classification processing experiment achieved satisfactory result.
2000, 22(5): 729-734.
Abstract:
In this paper, the continuous time recurrent neural network is proposed to solve the functional minimization problem, which is often involved in estimation and control. At first, the continuous time BPTT algorithm corresponding to the problem is presented. Then,an on-line algorithm based on the amendments of the BPTT algorithm is discussed. This on-line algorithm paves the way for parallel realization.
In this paper, the continuous time recurrent neural network is proposed to solve the functional minimization problem, which is often involved in estimation and control. At first, the continuous time BPTT algorithm corresponding to the problem is presented. Then,an on-line algorithm based on the amendments of the BPTT algorithm is discussed. This on-line algorithm paves the way for parallel realization.
2000, 22(5): 735-740.
Abstract:
Super resolution Algorithms can be used to obtain high resolution in ISAR imaging.But in the case of maneuvering target,the Doppler frequency of received signal scattered by target will not be constant,and these algorithms always suffer from the signal model error or fail completely when they are used in the cross range.In this paper,the first order approximate model of Doppler frequency is assumed,and an algorithm based on the extended RELAX method is proposed to estimate the parameters of chirp signals.The Cramer-Rao bounds are derived and simulation results are also present to demonstrate the performance of the algorithm.
Super resolution Algorithms can be used to obtain high resolution in ISAR imaging.But in the case of maneuvering target,the Doppler frequency of received signal scattered by target will not be constant,and these algorithms always suffer from the signal model error or fail completely when they are used in the cross range.In this paper,the first order approximate model of Doppler frequency is assumed,and an algorithm based on the extended RELAX method is proposed to estimate the parameters of chirp signals.The Cramer-Rao bounds are derived and simulation results are also present to demonstrate the performance of the algorithm.
2000, 22(5): 741-746.
Abstract:
With adoption of the hopped-frequency pulses as a radar waveform for imaging targets,a novel algorithm for range-Doppler imaging of rotating targets is proposed.Compared with the stepped-frequency radar Fourier transform(FT) imaging algorithm,the algorithm is characterized by large unambiguous Doppler window that eliminates Doppler confusion occasionally arising via the stepped-frequency radar FT imaging algorithm.Computer simulation results are given to confirm the effectiveness of the proposed algorithm.
With adoption of the hopped-frequency pulses as a radar waveform for imaging targets,a novel algorithm for range-Doppler imaging of rotating targets is proposed.Compared with the stepped-frequency radar Fourier transform(FT) imaging algorithm,the algorithm is characterized by large unambiguous Doppler window that eliminates Doppler confusion occasionally arising via the stepped-frequency radar FT imaging algorithm.Computer simulation results are given to confirm the effectiveness of the proposed algorithm.
2000, 22(5): 747-752.
Abstract:
This paper analyzes the distributed OS-CFAR detection with noncoherent integration of multiple pulses under the conditions of nonhomogeneous background and two local detectors.Compared with the monopulse detection and the single sensor detector and the corresponding optimal fixed threshold detection,the integration of multiple pulses improves the distributed detection obviously,and the increase of number of integrated pulses reduces its difference from the optimal fixed threshold detection.Although the increase of number of integrated pulses weaken the superority of distributed detection over the single sensor detection,the superority still exist and it is more obvious for low SNR.
This paper analyzes the distributed OS-CFAR detection with noncoherent integration of multiple pulses under the conditions of nonhomogeneous background and two local detectors.Compared with the monopulse detection and the single sensor detector and the corresponding optimal fixed threshold detection,the integration of multiple pulses improves the distributed detection obviously,and the increase of number of integrated pulses reduces its difference from the optimal fixed threshold detection.Although the increase of number of integrated pulses weaken the superority of distributed detection over the single sensor detection,the superority still exist and it is more obvious for low SNR.
2000, 22(5): 753-758.
Abstract:
This paper presents an iterative instantaneous frequency estimation based on a cross localized Wigner-Ville distribution(X-LWVD). The convergence of this method is examined analytically and the performance in noise is compared with the peak detection of spectrogram and the peak detection of Wigner-Ville distribution (WVD) by computer simulation. It is demonstrated that this algorithm is an statistically effective method.
This paper presents an iterative instantaneous frequency estimation based on a cross localized Wigner-Ville distribution(X-LWVD). The convergence of this method is examined analytically and the performance in noise is compared with the peak detection of spectrogram and the peak detection of Wigner-Ville distribution (WVD) by computer simulation. It is demonstrated that this algorithm is an statistically effective method.
2000, 22(5): 759-767.
Abstract:
A method of multiple parameters estimation for multiple chirp signals is proposed,a brief performance analysis is also described.Compared with references,the computational load of the method is more attractive,it is favorable to realize for real-time.It is no need for multiple-dimensions search.The results of simulation show that the method can accurately estimate multiple parameters of multiple chirp signals under the lower ratio of signals-to-noise.
A method of multiple parameters estimation for multiple chirp signals is proposed,a brief performance analysis is also described.Compared with references,the computational load of the method is more attractive,it is favorable to realize for real-time.It is no need for multiple-dimensions search.The results of simulation show that the method can accurately estimate multiple parameters of multiple chirp signals under the lower ratio of signals-to-noise.
2000, 22(5): 768-774.
Abstract:
A kind of parallel structure of the FIR filter is gotten by expressing the polynomial signal as the parallel form in this paper.By analyzing the parallel structure of FIR filter,several kinds of parallel processing algorithms of the adaptive FIR filter are presented.At the same time,the corresponding systolic implementation structures of these algorithms are given.
A kind of parallel structure of the FIR filter is gotten by expressing the polynomial signal as the parallel form in this paper.By analyzing the parallel structure of FIR filter,several kinds of parallel processing algorithms of the adaptive FIR filter are presented.At the same time,the corresponding systolic implementation structures of these algorithms are given.
2000, 22(5): 775-779.
Abstract:
This paper discusses the combination method of the optimal array based on the maximum signal to noise ratio principle and multi-user detection technique,and proves in theory that the code error rate of the new method is far superior to the sole multi-user detection.
This paper discusses the combination method of the optimal array based on the maximum signal to noise ratio principle and multi-user detection technique,and proves in theory that the code error rate of the new method is far superior to the sole multi-user detection.
2000, 22(5): 780-784.
Abstract:
In this paper,the wavelet-based fractal block image coding is introduced.A new oriented zerotree wavelet-based fractal image coding is presented through the combination of zerotree wavelet and fractal image coding.Experimental results show that the performance of this new approach outperforms that of fractal image coding with quadtrees and is almost comparable to that of embedded zerotree wavelet coding (EZW) at high compression ratios.
In this paper,the wavelet-based fractal block image coding is introduced.A new oriented zerotree wavelet-based fractal image coding is presented through the combination of zerotree wavelet and fractal image coding.Experimental results show that the performance of this new approach outperforms that of fractal image coding with quadtrees and is almost comparable to that of embedded zerotree wavelet coding (EZW) at high compression ratios.
2000, 22(5): 785-790.
Abstract:
Both Otsu method and KSW entropic thresholding segmentation method are implemented using Genetic Algorithm(GA) in this paper. Optimum parameters suitable for the image segmentation algorithm are also devised. Single thresholding and multithresholding methods are all presented using GA.The results show that GA can shorten the computational time compared with the classical methods.
Both Otsu method and KSW entropic thresholding segmentation method are implemented using Genetic Algorithm(GA) in this paper. Optimum parameters suitable for the image segmentation algorithm are also devised. Single thresholding and multithresholding methods are all presented using GA.The results show that GA can shorten the computational time compared with the classical methods.
2000, 22(5): 791-798.
Abstract:
Two-band continuous time wavelets can be obtained from infinite-level binary tree-structured QMF banks.Similarly,two-band discrete time wavelet transform can be represented as a finite-level binary tree-structured QMF banks.In this paper,two-band discrete time wavelet transforms are generalized to M-band discrete time wavelet transforms.The relationship between M-band discrete time wavelet transforms and M-band paraunitary filter banks is also investigated.It is shown that if the filter banks on each level are parannitary,the wavelet basis generated by this tree is orthonormal.
Two-band continuous time wavelets can be obtained from infinite-level binary tree-structured QMF banks.Similarly,two-band discrete time wavelet transform can be represented as a finite-level binary tree-structured QMF banks.In this paper,two-band discrete time wavelet transforms are generalized to M-band discrete time wavelet transforms.The relationship between M-band discrete time wavelet transforms and M-band paraunitary filter banks is also investigated.It is shown that if the filter banks on each level are parannitary,the wavelet basis generated by this tree is orthonormal.
2000, 22(5): 799-805.
Abstract:
A statistical characteristics model of empirical VBR MPEG-I video streams is presented in this paper. Based on fractional Gaussian noise process, an self-similar model approach which models directly both the short-range dependence and long-range dependence empirical autocorrelation structure is studied. According to the simulation result, the improved self-similar process is a better model of VBR video traffic.
A statistical characteristics model of empirical VBR MPEG-I video streams is presented in this paper. Based on fractional Gaussian noise process, an self-similar model approach which models directly both the short-range dependence and long-range dependence empirical autocorrelation structure is studied. According to the simulation result, the improved self-similar process is a better model of VBR video traffic.
2000, 22(5): 806-810.
Abstract:
On the base of a finite state sequential machine, the well known Cariolaro s algorithm is used, namely the transition probability matrix method, to analyze the power spectrum of 8B/10B line code used in gigabit ethernet, and to give the closed-form solution and the curve of the power spectrum density.
On the base of a finite state sequential machine, the well known Cariolaro s algorithm is used, namely the transition probability matrix method, to analyze the power spectrum of 8B/10B line code used in gigabit ethernet, and to give the closed-form solution and the curve of the power spectrum density.
2000, 22(5): 811-816.
Abstract:
As a result of the rising demand for services and the resulting increase in size,bandwidth and complexity,fault management in todays high speed communication networks is becoming even more difficult.When a network problem or failure occurs,it is possible that a very large volume of alarm messages is generated,while alarm correlation is a potentially complex problem.Though some existing alarm correlation systems nowadays have different drawbacks such as lack of scalability,hindered by solving complexity,or no learning process,etc.This paper presents a fault-identification and alarm-correlation method based on improved GA-NN model in communication networks.The experimental results show that this method is simple,which not only overcomes the disadvantages of normal alarm correlation ways,but also improves the dynamic character,training accuracy and efficiency greatly than BP algorithm,BGA algorithm and AGA algorithm do.
As a result of the rising demand for services and the resulting increase in size,bandwidth and complexity,fault management in todays high speed communication networks is becoming even more difficult.When a network problem or failure occurs,it is possible that a very large volume of alarm messages is generated,while alarm correlation is a potentially complex problem.Though some existing alarm correlation systems nowadays have different drawbacks such as lack of scalability,hindered by solving complexity,or no learning process,etc.This paper presents a fault-identification and alarm-correlation method based on improved GA-NN model in communication networks.The experimental results show that this method is simple,which not only overcomes the disadvantages of normal alarm correlation ways,but also improves the dynamic character,training accuracy and efficiency greatly than BP algorithm,BGA algorithm and AGA algorithm do.
2000, 22(5): 817-823.
Abstract:
In this paper, a decorrelator for multiuser detection derived from ML criteria in CDMA system is proposed for frequency selective fading channels. Simulations prove the better performance of this method than the method proposed by Zvonar in which the paths of the same user are combined after decorrelating. This method utilizes equivalent spreading codes for decorrelating, thus the hardware complexity is reduced.
In this paper, a decorrelator for multiuser detection derived from ML criteria in CDMA system is proposed for frequency selective fading channels. Simulations prove the better performance of this method than the method proposed by Zvonar in which the paths of the same user are combined after decorrelating. This method utilizes equivalent spreading codes for decorrelating, thus the hardware complexity is reduced.
2000, 22(5): 824-830.
Abstract:
This paper systematically introduces how to find secure elliptic curves with the help of Weil theorem, and proposes an algorithm to find base points in the curves. Finally, an efficient method of finding the trace of any element in GF(2lk), which is involved in the algorithm of finding base points, is given.
This paper systematically introduces how to find secure elliptic curves with the help of Weil theorem, and proposes an algorithm to find base points in the curves. Finally, an efficient method of finding the trace of any element in GF(2lk), which is involved in the algorithm of finding base points, is given.
2000, 22(5): 831-835.
Abstract:
In this paper, the constructions for the families of Bent complementary functions (BCF) are further studied. Recursive construction methods of BCFs are given by using matrices with orthogonal columns and perfect arrays.
In this paper, the constructions for the families of Bent complementary functions (BCF) are further studied. Recursive construction methods of BCFs are given by using matrices with orthogonal columns and perfect arrays.
2000, 22(5): 836-841.
Abstract:
This paper gives an improved MBELP speech coding algorithm at 2.4kb/s.Compared with conventional MBELP algorithm,this algorithm has some improvements on pitch estimation and unvoice/voice decision.The synthetical speech of vocoder based on this algorithm is better than that based on conventional one.After a brief introduction,the improved MBELP algorithm is discussed in detail.And the results of pitch estimation and U/V decision based on the improved MBELP are compared with those based on conventional MBELP.
This paper gives an improved MBELP speech coding algorithm at 2.4kb/s.Compared with conventional MBELP algorithm,this algorithm has some improvements on pitch estimation and unvoice/voice decision.The synthetical speech of vocoder based on this algorithm is better than that based on conventional one.After a brief introduction,the improved MBELP algorithm is discussed in detail.And the results of pitch estimation and U/V decision based on the improved MBELP are compared with those based on conventional MBELP.
2000, 22(5): 842-846.
Abstract:
By analyzing the switching characteristic of the multiple- transistor and combining its features of emit-input, emit-output and high working speed, a ternary double-edge-triggered flip-flops is designed in the paper. Computer simulation shows that this design has correct logic function and high working speed. It can be used in designing ternary DYL circuits.
By analyzing the switching characteristic of the multiple- transistor and combining its features of emit-input, emit-output and high working speed, a ternary double-edge-triggered flip-flops is designed in the paper. Computer simulation shows that this design has correct logic function and high working speed. It can be used in designing ternary DYL circuits.
2000, 22(5): 847-853.
Abstract:
A fast generalized algorithm for S parameter extraction of 3-D microwave structures is presented based on the 3-D FDFD (Frequency Domain Finite Difference) method in conjunction with the asymptotic waveform evaluation (AWE) technique.In the AWE technique,the electric field strength at the input and output ports are respectively expanded in a Taylors series around a frequency within the desired frequency band,and the coefficients of the Taylors series are matched via the Pad approximation to a rational function and the robust Singular Value Decomposition (SVD) technique is employed while solving the Pad equation.Using the rational function,the electric field strengths at the input and output are obtained at any frequency within the desired frequency band,which is used to calculated the S parameter.A good agreement between the numerical results and the reported data is observed.Compared with the traditional FDFD,the computational efficiency of the present method is considerably improved.
A fast generalized algorithm for S parameter extraction of 3-D microwave structures is presented based on the 3-D FDFD (Frequency Domain Finite Difference) method in conjunction with the asymptotic waveform evaluation (AWE) technique.In the AWE technique,the electric field strength at the input and output ports are respectively expanded in a Taylors series around a frequency within the desired frequency band,and the coefficients of the Taylors series are matched via the Pad approximation to a rational function and the robust Singular Value Decomposition (SVD) technique is employed while solving the Pad equation.Using the rational function,the electric field strengths at the input and output are obtained at any frequency within the desired frequency band,which is used to calculated the S parameter.A good agreement between the numerical results and the reported data is observed.Compared with the traditional FDFD,the computational efficiency of the present method is considerably improved.
2000, 22(5): 854-857.
Abstract:
The scattering matrix, which describes the scattering property of practical scatters, is generally a function of time. So, the scattered electromagnetic wave is no longer a completely polarized wave, but a partially polarized wave, although the incident wave is completely polarized. This paper mainly focuses on the optimal reception of partially polarized electromagnetic wave, and derives the method for optimizing the intensity received by the antenna.
The scattering matrix, which describes the scattering property of practical scatters, is generally a function of time. So, the scattered electromagnetic wave is no longer a completely polarized wave, but a partially polarized wave, although the incident wave is completely polarized. This paper mainly focuses on the optimal reception of partially polarized electromagnetic wave, and derives the method for optimizing the intensity received by the antenna.
2000, 22(5): 865-870.
Abstract:
Since self-similarity is a common property of traffic in the broadband networks,it is crucial to consider its impact on network management and traffic control.Nowadays,research on self-similar traffic control has been widely noticed.Some results prove that self-similar traffic exhibits different queueing features from the traditional models.Considering on the quality of service and traffic control mechanism,this paper investigates the impact of self-similar traffic on some traffic control mechanisms such as traffic shaping,Usage Parameter Control and Connection Admission Control,and obtains some important conclusions.
Since self-similarity is a common property of traffic in the broadband networks,it is crucial to consider its impact on network management and traffic control.Nowadays,research on self-similar traffic control has been widely noticed.Some results prove that self-similar traffic exhibits different queueing features from the traditional models.Considering on the quality of service and traffic control mechanism,this paper investigates the impact of self-similar traffic on some traffic control mechanisms such as traffic shaping,Usage Parameter Control and Connection Admission Control,and obtains some important conclusions.
2000, 22(5): 871-874.
Abstract:
The influences of supporting dielectric layers on the transmission properties of FSS with Y slot array embeded in the dielectric are studied in this paper. The frequency properties (resonance frequency, transmission bandwidth and transmission loss are included) of the FSS with different dielectric thickness are discussed. Calculated results show that dielectric losses are main sources of the base of FSS structure.
The influences of supporting dielectric layers on the transmission properties of FSS with Y slot array embeded in the dielectric are studied in this paper. The frequency properties (resonance frequency, transmission bandwidth and transmission loss are included) of the FSS with different dielectric thickness are discussed. Calculated results show that dielectric losses are main sources of the base of FSS structure.
2000, 22(5): 875-880.
Abstract:
The problem of scattering of a Gaussian beam by a rough dielectric surface is studied. By using the plane wave spectrum expansion and Kirchhoff approximation theory, the formulas of scattered field and the incoherent scattering cross-section for Gaussian beam illumination of a rough surface are derived. And the numerical results for incoherent back scattering cross-section are given. Finally, the above results are analyzed and discussed.
The problem of scattering of a Gaussian beam by a rough dielectric surface is studied. By using the plane wave spectrum expansion and Kirchhoff approximation theory, the formulas of scattered field and the incoherent scattering cross-section for Gaussian beam illumination of a rough surface are derived. And the numerical results for incoherent back scattering cross-section are given. Finally, the above results are analyzed and discussed.
2000, 22(5): 858-864.
Abstract:
Qi Ming and others recently analyzed and discussed the security of two kinds of the ElGamal signature schemes and the password authentication scheme based on two kinds of the signature schemes,and proposed two kinds of improved schemes.This paper first points out that the first p type signature scheme proposed by Qi is not secure,since attackers can forge signature for any message.Then this paper shows that the generalized ElGamal signature schemes can not resist the substitution attack.Finally this paper shows that two kinds of the improved schemes proposed by Qi can not resist the homomorphism attack,and does not have the security as Qi said.
Qi Ming and others recently analyzed and discussed the security of two kinds of the ElGamal signature schemes and the password authentication scheme based on two kinds of the signature schemes,and proposed two kinds of improved schemes.This paper first points out that the first p type signature scheme proposed by Qi is not secure,since attackers can forge signature for any message.Then this paper shows that the generalized ElGamal signature schemes can not resist the substitution attack.Finally this paper shows that two kinds of the improved schemes proposed by Qi can not resist the homomorphism attack,and does not have the security as Qi said.
