Email alert
2002 Vol. 24, No. 1
Display Method:
2002, 24(1): 1-5.
Abstract:
With respect to multicomponent chirp signals appeared in ISAR imaging, a new signal decomposition method based on CLEAN in frequency domain is proposed to suppress crossterms in Wigner-Ville Distribution (WVD), in which the multicomponents are first decomposed and then WVD of each decomposed one is computed. Compared with other chirp decomposition method available, the method presented here not only reduces the number of parameters to be estimated, which simplifies the computation, but also keeps the correct signal features. As shown in the experiment, a clearer time-frequency distribution with high time frequency resolution and greatly suppressed crossterms can e obtained, and the ISAR imaging quality is improved greatly.
With respect to multicomponent chirp signals appeared in ISAR imaging, a new signal decomposition method based on CLEAN in frequency domain is proposed to suppress crossterms in Wigner-Ville Distribution (WVD), in which the multicomponents are first decomposed and then WVD of each decomposed one is computed. Compared with other chirp decomposition method available, the method presented here not only reduces the number of parameters to be estimated, which simplifies the computation, but also keeps the correct signal features. As shown in the experiment, a clearer time-frequency distribution with high time frequency resolution and greatly suppressed crossterms can e obtained, and the ISAR imaging quality is improved greatly.
2002, 24(1): 6-11.
Abstract:
This paper analyses the airborne radar ground clutter suppression abilitv of DPCA by decomposing the clutter zone into small celis, and points out that DPCA is a shnple tinic-space processor. This paper also analyses the reason that lowers the DPCA performancc and proposes a DPCA method based on clutter frequency shift. Theoretical analysis and simulation show that this method s performance is good and it is easy to implement.
This paper analyses the airborne radar ground clutter suppression abilitv of DPCA by decomposing the clutter zone into small celis, and points out that DPCA is a shnple tinic-space processor. This paper also analyses the reason that lowers the DPCA performancc and proposes a DPCA method based on clutter frequency shift. Theoretical analysis and simulation show that this method s performance is good and it is easy to implement.
2002, 24(1): 12-18.
Abstract:
The motion compensation errors of the airborne Synthetic Aperture Radar (SAR,) using the typical aircraft Inertial Navigation System (INS) are analyzed. It is conduded that the navigation errors of INS are of low frequency and divergent. The INS errors cari be kept in a relativelv small.range at.the beginning stage. Therefore, the high resolution SAR, imaging can be achieved by combining INS with clutterlock and autofocus mothods.
The motion compensation errors of the airborne Synthetic Aperture Radar (SAR,) using the typical aircraft Inertial Navigation System (INS) are analyzed. It is conduded that the navigation errors of INS are of low frequency and divergent. The INS errors cari be kept in a relativelv small.range at.the beginning stage. Therefore, the high resolution SAR, imaging can be achieved by combining INS with clutterlock and autofocus mothods.
2002, 24(1): 19-24.
Abstract:
In this paper, a method of direct digital generation of intermediate frequency(IF) pulse compression signals is proposed, and the hardware architecture is also presented. Some key issues are analyzed quantitatively. Based on these, a generation system is implemented, which is applicable to generate six kinds of LFM, NLFM waveforms. The designed IF reaches 15MHz and bandwidth reaches 1MHz with timewidth equal to or less than 300s. The measuring results of SFDR and noise are all below-68dB, which show that the performances of the system can meet the needs of practical engineering application.
In this paper, a method of direct digital generation of intermediate frequency(IF) pulse compression signals is proposed, and the hardware architecture is also presented. Some key issues are analyzed quantitatively. Based on these, a generation system is implemented, which is applicable to generate six kinds of LFM, NLFM waveforms. The designed IF reaches 15MHz and bandwidth reaches 1MHz with timewidth equal to or less than 300s. The measuring results of SFDR and noise are all below-68dB, which show that the performances of the system can meet the needs of practical engineering application.
2002, 24(1): 25-30.
Abstract:
The resampling of the passband signal with variable frequency is likely implemented by interpolating and decimating the original sequence according to the converting ratio (P/Q). When the value of P is very large, the multiple interpolating filters will be required, and due to anti-image frequency filtering, the coefficient matrix of filters are very complex, which results in the resampling difficult to implement. This paper presents a method of the Polynomical Approximate Filtering (PAF). It uses a group of coefficient matrixed for low-order PAF, with efficient operation and variable interpolating delay. The computer simulation show that it is suitable to the high order interpolating filters with variable delay. In some cases, the tolerance is within the allowable range.
The resampling of the passband signal with variable frequency is likely implemented by interpolating and decimating the original sequence according to the converting ratio (P/Q). When the value of P is very large, the multiple interpolating filters will be required, and due to anti-image frequency filtering, the coefficient matrix of filters are very complex, which results in the resampling difficult to implement. This paper presents a method of the Polynomical Approximate Filtering (PAF). It uses a group of coefficient matrixed for low-order PAF, with efficient operation and variable interpolating delay. The computer simulation show that it is suitable to the high order interpolating filters with variable delay. In some cases, the tolerance is within the allowable range.
2002, 24(1): 31-36.
Abstract:
Dynamic Time Warping(DTW) has been widely used in speech recognition systems as a nonlinear time alignment technique. It uses the dynamic programming technique to search the optimal warping path for two time sequences. Although this algorithm needs less computation and shorter training and searching time, it is a local optimization algorithm. The Tabu Search (TS) algorithm is the generalized heuristic global search technique with short-time memory, and suitable for solving many nonlinear optimization problems. This paper applies this technique to speech recognition systems, and presents a new algorithm for optimizing time warping based on TS approach, which makes time warping functions optimized globally. Simulation results show that TSTW has better time warping performance than DTW and GTW.
Dynamic Time Warping(DTW) has been widely used in speech recognition systems as a nonlinear time alignment technique. It uses the dynamic programming technique to search the optimal warping path for two time sequences. Although this algorithm needs less computation and shorter training and searching time, it is a local optimization algorithm. The Tabu Search (TS) algorithm is the generalized heuristic global search technique with short-time memory, and suitable for solving many nonlinear optimization problems. This paper applies this technique to speech recognition systems, and presents a new algorithm for optimizing time warping based on TS approach, which makes time warping functions optimized globally. Simulation results show that TSTW has better time warping performance than DTW and GTW.
2002, 24(1): 37-44.
Abstract:
Chaotic signal source design based on neural networks trained by modified back propagation algorithm and synchronization by nonlinear feedback method have been proposed in this paper. This neural network source can very easily generate many chaotic signals by changing synaptic weights of neural networks. Nonlinear feedback term is used to synchronize chaotic signal sources designed by neural networks. Experimental results show that these chaotic sources can generate much more signal than other chaotic sources designed by the single chaotic map, and are very easily synchronized by the same nonlinear feedback method.
Chaotic signal source design based on neural networks trained by modified back propagation algorithm and synchronization by nonlinear feedback method have been proposed in this paper. This neural network source can very easily generate many chaotic signals by changing synaptic weights of neural networks. Nonlinear feedback term is used to synchronize chaotic signal sources designed by neural networks. Experimental results show that these chaotic sources can generate much more signal than other chaotic sources designed by the single chaotic map, and are very easily synchronized by the same nonlinear feedback method.
2002, 24(1): 45-53.
Abstract:
It is always difficult to train a two-layer Feedforward Neural Networks (FNN) based on the cumulant match criterion because cumulants are nonlinear implicit functions of the FNN parameters. In this work, two new cumulant-based training methods for two-layer FNN are developed. In the first method, the hidden units of two-layer FNN are approximated with multiple linear systems, and further total FNN is modeled with a mixture of experts (ME) architecture. With the ME model, FNN parameters are estimated with expectation-maximization (EM) algorithm. The second method, for simplifying the two-layer FNN statistical model, proposes a simplified two-level hierarchical ME to remodel the FNN, in which hidden variables are introduced such that training total FNN is changed to training a set of single neurons. Based on training single neuron, total FNN is estimated in a simplified version with a faster convergence speed.
It is always difficult to train a two-layer Feedforward Neural Networks (FNN) based on the cumulant match criterion because cumulants are nonlinear implicit functions of the FNN parameters. In this work, two new cumulant-based training methods for two-layer FNN are developed. In the first method, the hidden units of two-layer FNN are approximated with multiple linear systems, and further total FNN is modeled with a mixture of experts (ME) architecture. With the ME model, FNN parameters are estimated with expectation-maximization (EM) algorithm. The second method, for simplifying the two-layer FNN statistical model, proposes a simplified two-level hierarchical ME to remodel the FNN, in which hidden variables are introduced such that training total FNN is changed to training a set of single neurons. Based on training single neuron, total FNN is estimated in a simplified version with a faster convergence speed.
2002, 24(1): 54-59.
Abstract:
At certain measurement rate and error, the best estimation accuracy which may be achieved by single passive observer is decided by the velocity and trajectory and trajectory of the observer, independent of the estimation algorithm. In this paper, simulation experiments show the effects of measurement rate, error and observer velocity, trajectory on the error Cramer-Rao lower bounds of position estimate under two kinds of observer trajectories. Then the best Cramer-Rao lower bounds of three kinds of observer trajectories are compared. The experimental results are very useful for design of passive target tracking system.
At certain measurement rate and error, the best estimation accuracy which may be achieved by single passive observer is decided by the velocity and trajectory and trajectory of the observer, independent of the estimation algorithm. In this paper, simulation experiments show the effects of measurement rate, error and observer velocity, trajectory on the error Cramer-Rao lower bounds of position estimate under two kinds of observer trajectories. Then the best Cramer-Rao lower bounds of three kinds of observer trajectories are compared. The experimental results are very useful for design of passive target tracking system.
2002, 24(1): 60-67.
Abstract:
The estimation of local mean signal strength is the core of SSOA (Signal Strength Of Arrival) location. The primary source of error is fast fading and shadowing (slow fading). Signal strength averaging is an important approach to estimate the local mean. This paper suggests a double-averaging algorithm with dynamic widow-width to improve the estimation and thus enhances the location accuracy. The results of computer simulation proved that the methods improve location accuracy by 30% than traditional methods in fading channel. And, with the aid of dynamic window-width, sufficient location accuracy can be obtained while mobile station is in low mobility.
The estimation of local mean signal strength is the core of SSOA (Signal Strength Of Arrival) location. The primary source of error is fast fading and shadowing (slow fading). Signal strength averaging is an important approach to estimate the local mean. This paper suggests a double-averaging algorithm with dynamic widow-width to improve the estimation and thus enhances the location accuracy. The results of computer simulation proved that the methods improve location accuracy by 30% than traditional methods in fading channel. And, with the aid of dynamic window-width, sufficient location accuracy can be obtained while mobile station is in low mobility.
2002, 24(1): 68-76.
Abstract:
This paper explores the problem of non-coherent multiuser weight/phase optimal joint detection for QPSK signal in over-saturated multiple access communication. Firstly, non-coherent multiuser weight/phase estimator is given via theoretical analysis. Then, the low complexity non-coherent multiuser weight/phase optimal joint detection algorithm is proposed, which is fit with practical working-condition and is applicable to linear modulation methods with two-dimensional constellation. Finally, the performance of the non-coherent optimal receiver and the phase estimate error is analyzed via computer simulation and some useful conclusions are drawn.
This paper explores the problem of non-coherent multiuser weight/phase optimal joint detection for QPSK signal in over-saturated multiple access communication. Firstly, non-coherent multiuser weight/phase estimator is given via theoretical analysis. Then, the low complexity non-coherent multiuser weight/phase optimal joint detection algorithm is proposed, which is fit with practical working-condition and is applicable to linear modulation methods with two-dimensional constellation. Finally, the performance of the non-coherent optimal receiver and the phase estimate error is analyzed via computer simulation and some useful conclusions are drawn.
2002, 24(1): 77-82.
Abstract:
This paper proposes the mapping of spectral coefficients for normalized Haar transform-ha map, and gives the mapping transform methods between ha map and K-map. This transform process is shown by examples. This method has the feature of simplicity, intu-ition and precision.
This paper proposes the mapping of spectral coefficients for normalized Haar transform-ha map, and gives the mapping transform methods between ha map and K-map. This transform process is shown by examples. This method has the feature of simplicity, intu-ition and precision.
2002, 24(1): 83-89.
Abstract:
In this paper, the numerical method of combining surface parametric quadratic modeling with singularity orders reducing in moment method is analyzed. The objects are modeled by parametric quadratic surface. The singularity order of the integral is reduced from O(1/R2) to O(1/R). Compared with other methods, this method simplifies the complexity of self-impedance elements calculating and increases its calculating stability and accuracy. It is applied to calculating RCS of some objects. Good agreement with the results from some other methods is found.
In this paper, the numerical method of combining surface parametric quadratic modeling with singularity orders reducing in moment method is analyzed. The objects are modeled by parametric quadratic surface. The singularity order of the integral is reduced from O(1/R2) to O(1/R). Compared with other methods, this method simplifies the complexity of self-impedance elements calculating and increases its calculating stability and accuracy. It is applied to calculating RCS of some objects. Good agreement with the results from some other methods is found.
2002, 24(1): 90-96.
Abstract:
The cutoff frequency of heteromophic waveguide in special structure is solved with the matching of electric and magnetic field on the invented boundary by vector partial differen-tial operator theory. The method rnay be used to solving the cutoff frequency of almost all of the modes precisely and expediently. The method rnay also be used to calculating the trausmit-ting characteristics of ridged waveguide and other heteromophic waveguide loaded rectaugular metal.
The cutoff frequency of heteromophic waveguide in special structure is solved with the matching of electric and magnetic field on the invented boundary by vector partial differen-tial operator theory. The method rnay be used to solving the cutoff frequency of almost all of the modes precisely and expediently. The method rnay also be used to calculating the trausmit-ting characteristics of ridged waveguide and other heteromophic waveguide loaded rectaugular metal.
2002, 24(1): 97-101.
Abstract:
In this paper, the adaptive antenna array performance is given when the dual polarized jamming signals are considered. If the phases of two polarized elements are changed dnring the adaptive weight calculation, the amplitute and phase errors between the antenna elements will be also changed due to the inconsistaricy of the element cross-polarization. This kind of error can not be corrected and the signal to noise plus interference ratio will be rednced.
In this paper, the adaptive antenna array performance is given when the dual polarized jamming signals are considered. If the phases of two polarized elements are changed dnring the adaptive weight calculation, the amplitute and phase errors between the antenna elements will be also changed due to the inconsistaricy of the element cross-polarization. This kind of error can not be corrected and the signal to noise plus interference ratio will be rednced.
2002, 24(1): 102-107.
Abstract:
This paper presents observed results of spurious spectrum of TWTA used for satel-lite. The rnechanism of spurious spectrum formation is analysed and the rnethods of its sup-pression are discussed. It is shown from the study that in order to obtain pure spectrum, the measures rnust be taken for both design of TWT and EPC circuit. Based on this study, the aruplitude of spurious spectrum of spaceborne TWTA has been reduced 10-15dB
This paper presents observed results of spurious spectrum of TWTA used for satel-lite. The rnechanism of spurious spectrum formation is analysed and the rnethods of its sup-pression are discussed. It is shown from the study that in order to obtain pure spectrum, the measures rnust be taken for both design of TWT and EPC circuit. Based on this study, the aruplitude of spurious spectrum of spaceborne TWTA has been reduced 10-15dB
2002, 24(1): 108-114.
Abstract:
The degradation induced by interface state is one main reason for failure occurs in deep-sub-micron MOS devices. Based on the hydrodynamics energy transport model, the degradation induced by acceptor interface state is analyzed for deep-sub-micron grooved-gate and conventional planar PMOSFET with different channel doping density. The simulation results show that the degradation induced by the same interface state density in grooved-gate PMOSFET is larger than that in planar PMOSFET and in both structure devices, the impact of P type acceptor interface state on device performance is far larger than that of N type. It also manifests that the degradation is different for the device with different channel doping density.
The degradation induced by interface state is one main reason for failure occurs in deep-sub-micron MOS devices. Based on the hydrodynamics energy transport model, the degradation induced by acceptor interface state is analyzed for deep-sub-micron grooved-gate and conventional planar PMOSFET with different channel doping density. The simulation results show that the degradation induced by the same interface state density in grooved-gate PMOSFET is larger than that in planar PMOSFET and in both structure devices, the impact of P type acceptor interface state on device performance is far larger than that of N type. It also manifests that the degradation is different for the device with different channel doping density.
2002, 24(1): 115-119.
Abstract:
The impact of traffic dependence on queueing performance is of great significance to traffic control mechanism and resource allocation algorithm in high-speed networks. This article adopts a unique method to analyze the influence that the traffic dependence brings to the queueing performance under given network conditions. The results of simulation show that long-range dependence during a certain time scale will deteriorate the queueing performance, and that such an effect is insensitive to increasing the buffer size. Also the marginal distribution and Hurst exponent of traffic series are inadequate to determine the queueing performance.
The impact of traffic dependence on queueing performance is of great significance to traffic control mechanism and resource allocation algorithm in high-speed networks. This article adopts a unique method to analyze the influence that the traffic dependence brings to the queueing performance under given network conditions. The results of simulation show that long-range dependence during a certain time scale will deteriorate the queueing performance, and that such an effect is insensitive to increasing the buffer size. Also the marginal distribution and Hurst exponent of traffic series are inadequate to determine the queueing performance.
2002, 24(1): 120-124.
Abstract:
Rapid carrier and bit-timing synchronization are essential in burst mode communication systems. A new algorithm for fast synchronization in a bust mode OQPSK receiver is proposed in this paper. The algorithm is based on a specific bit-pattern preamble, which allows total decoupling of the carrier phase and bit-timing phase and thus permits direct, simultaneous estimating of the reference parameters by open-loop approach. It is shown that the algorithm achieves a significantly agile carrier and bit-timing recovery. The algorithm is applicable to satellite communication systems, mobile communication systems etc.
Rapid carrier and bit-timing synchronization are essential in burst mode communication systems. A new algorithm for fast synchronization in a bust mode OQPSK receiver is proposed in this paper. The algorithm is based on a specific bit-pattern preamble, which allows total decoupling of the carrier phase and bit-timing phase and thus permits direct, simultaneous estimating of the reference parameters by open-loop approach. It is shown that the algorithm achieves a significantly agile carrier and bit-timing recovery. The algorithm is applicable to satellite communication systems, mobile communication systems etc.
2002, 24(1): 125-130.
Abstract:
This paper studies the ways to determine the buffer size of the AAL2 packet voice multiplexer with bit dropping. A conclusion is drawn that the buffer size which corresponds to the delay budget of 9 millisecond is sufficient to meet the quality of service requirements of AAL2 packet voice. An algorithm for determining the buffer size and the bit dropping thresholds is also presented with relevant computation formula. The computer simulation results show that the algorithm presented here is good for determining the buffer size and the bit dropping thresholds with small packet loss probability, low packet delay and high mean bits per sample and can greatly improve the buffer efficiency.
This paper studies the ways to determine the buffer size of the AAL2 packet voice multiplexer with bit dropping. A conclusion is drawn that the buffer size which corresponds to the delay budget of 9 millisecond is sufficient to meet the quality of service requirements of AAL2 packet voice. An algorithm for determining the buffer size and the bit dropping thresholds is also presented with relevant computation formula. The computer simulation results show that the algorithm presented here is good for determining the buffer size and the bit dropping thresholds with small packet loss probability, low packet delay and high mean bits per sample and can greatly improve the buffer efficiency.
2002, 24(1): 131-135.
Abstract:
The design of Generation 3 mobile comrrmnication systern is required to support multimedia conimunication. Effective power control is necessary, as different QoS i.s required at reoeiving ends for each service. This paper presents SRA (Stepwise Removal Algorithrn) algorithm through introducing the idea of power equivalence. It is verified by ernulator analysis that this algorithrn can increase system capacity while ensuririg serving quality of each service.
The design of Generation 3 mobile comrrmnication systern is required to support multimedia conimunication. Effective power control is necessary, as different QoS i.s required at reoeiving ends for each service. This paper presents SRA (Stepwise Removal Algorithrn) algorithm through introducing the idea of power equivalence. It is verified by ernulator analysis that this algorithrn can increase system capacity while ensuririg serving quality of each service.
2002, 24(1): 136-139.
Abstract:
Thresholding is an important form of image segmentation and is used in the image processing for many applications. In this paper, an automatic technique for thresholding of digital images based on the maximum correlation criterion is presented. This method maximizes the correlations associated with the distributions of the background and object classes to obtain the optimal threshold. It is found that the proposed approach is more effective than that proposed by Yen, and its computational complexity is much lower compared with the methods of entropy.
Thresholding is an important form of image segmentation and is used in the image processing for many applications. In this paper, an automatic technique for thresholding of digital images based on the maximum correlation criterion is presented. This method maximizes the correlations associated with the distributions of the background and object classes to obtain the optimal threshold. It is found that the proposed approach is more effective than that proposed by Yen, and its computational complexity is much lower compared with the methods of entropy.
2002, 24(1): 140-144.
Abstract:
An effective binding energy and a formula f or evaporation fields in tlie configu-ration of a scarming tunneling microscope has been introduced. Based on which, evaporation fields for various tip/substrate combinations, especially those for non-identical tip/substrate combinations have been calculated. To do this a universal function for atomic potential of an adatom at different distances from a metai surface is proposed.
An effective binding energy and a formula f or evaporation fields in tlie configu-ration of a scarming tunneling microscope has been introduced. Based on which, evaporation fields for various tip/substrate combinations, especially those for non-identical tip/substrate combinations have been calculated. To do this a universal function for atomic potential of an adatom at different distances from a metai surface is proposed.
