Email alert
1997 Vol. 19, No. 4
column
Display Method:
1997, 19(4): 433-438.
Abstract:
This paper presents a novel theorem on the multi-dimensional function approximation ability of feed forward multi-layer neural networks (FFMLNN), which states that the function approximation ability of FFMLNN is independent of the dimension of the function to be approximated when the number of the hidden units is sufficiently large. This theorem simplifies greatly the analysis of the function approximation ability of FFMLNN because one needs only to study the one dimensional function approximation ability of FFMLNN. An application of the proposed theorem is given.
This paper presents a novel theorem on the multi-dimensional function approximation ability of feed forward multi-layer neural networks (FFMLNN), which states that the function approximation ability of FFMLNN is independent of the dimension of the function to be approximated when the number of the hidden units is sufficiently large. This theorem simplifies greatly the analysis of the function approximation ability of FFMLNN because one needs only to study the one dimensional function approximation ability of FFMLNN. An application of the proposed theorem is given.
1997, 19(4): 439-444.
Abstract:
Bidirectional Associative Memory(BAM) models are two-layer heteroassciative networks. In this paper, conditions for the existence of equilibrium points and global stability are emphatically disscussed for BAM with axonal signal transmission delays. The correctness of obtained conclusions is verified by use of some examples. The disscussed methods are more general. The obtained results have important leading significance in the design and application of BAM.
Bidirectional Associative Memory(BAM) models are two-layer heteroassciative networks. In this paper, conditions for the existence of equilibrium points and global stability are emphatically disscussed for BAM with axonal signal transmission delays. The correctness of obtained conclusions is verified by use of some examples. The disscussed methods are more general. The obtained results have important leading significance in the design and application of BAM.
1997, 19(4): 445-450.
Abstract:
A circuit of artificial neural network for implementing stable adaptive IIR filter is proposed. The circuit parameters effect the speed and stability of the adaptive process. The formula for the parameters selection is also given. So this kind of systems is suitable to be used processing the fast changing signals.
A circuit of artificial neural network for implementing stable adaptive IIR filter is proposed. The circuit parameters effect the speed and stability of the adaptive process. The formula for the parameters selection is also given. So this kind of systems is suitable to be used processing the fast changing signals.
1997, 19(4): 451-458.
Abstract:
This paper studies Synthetic Aperture Radar(SAR) images with the multiresolution analysis of wavelet transform. This study indicates that the influences of the coherent speckle noise on different frequency components of the SAR images are different, and that the manners of singularity of the SAR images and the speckle noise are different, too. So, this paper presents a denoising method of wavelet analysis to reduce the coherent speckle noise. In the meanwhile, a lot of experiments approve that this method not only suppresses the speckle noise effectively, but also preserves as many target characteristics of original images as possible. It shows that this denoising method of wavelet analysis offers a very attractive alternative to suppress the coherent speckle noise of the SAR images.
This paper studies Synthetic Aperture Radar(SAR) images with the multiresolution analysis of wavelet transform. This study indicates that the influences of the coherent speckle noise on different frequency components of the SAR images are different, and that the manners of singularity of the SAR images and the speckle noise are different, too. So, this paper presents a denoising method of wavelet analysis to reduce the coherent speckle noise. In the meanwhile, a lot of experiments approve that this method not only suppresses the speckle noise effectively, but also preserves as many target characteristics of original images as possible. It shows that this denoising method of wavelet analysis offers a very attractive alternative to suppress the coherent speckle noise of the SAR images.
1997, 19(4): 459-463.
Abstract:
This paper studies intensively the bilinear interpolation and 2D-IFFT method of reconstructing image from the unevenly spaced data in the spatial frequency domain. The basic ideas are to decide an evenly grid, interpolate the unevenly spaced data on a uniform grid and take the 2D-IFFT of the data on the uniform grid. The paper presents the principle of deciding the grid, discusses bilinear interpolation and analyses some important problems of 2D-IFFT. The method has been applied to synthetic aperture radar(SAR) imaging and the result is satisfactory.
This paper studies intensively the bilinear interpolation and 2D-IFFT method of reconstructing image from the unevenly spaced data in the spatial frequency domain. The basic ideas are to decide an evenly grid, interpolate the unevenly spaced data on a uniform grid and take the 2D-IFFT of the data on the uniform grid. The paper presents the principle of deciding the grid, discusses bilinear interpolation and analyses some important problems of 2D-IFFT. The method has been applied to synthetic aperture radar(SAR) imaging and the result is satisfactory.
1997, 19(4): 464-470.
Abstract:
A new method for the motion compensation of multiple moving targets imaging with SAR/ISAR based on Wigner-Ville and Hough transform is presented and studied. The main steps for motion detection, estimation and compensation are given. The simulated experimental results with computer and microwave anechoic chamber show that the new method is effective for suppressing the cross-terms and noise, so it is clear the new method is an improvement on a unitary WVD method, and the new method can be used to improve the detection and estimation capability of single or multiple moving targets in noisy enviroment.
A new method for the motion compensation of multiple moving targets imaging with SAR/ISAR based on Wigner-Ville and Hough transform is presented and studied. The main steps for motion detection, estimation and compensation are given. The simulated experimental results with computer and microwave anechoic chamber show that the new method is effective for suppressing the cross-terms and noise, so it is clear the new method is an improvement on a unitary WVD method, and the new method can be used to improve the detection and estimation capability of single or multiple moving targets in noisy enviroment.
1997, 19(4): 471-475.
Abstract:
This paper combines the least-square method and iteration method to get the fundamental matrix and develops a new evaluation function based on the epipolar geometry. During the iteration, with the evaluation function as a measurement, the points which bring larger noise are deleted, and the points with smaller noise are retained, thus the precision of our method is increased. The experimental results indicate the new method is precise in calculation, stable in performance and resistant to noise.
This paper combines the least-square method and iteration method to get the fundamental matrix and develops a new evaluation function based on the epipolar geometry. During the iteration, with the evaluation function as a measurement, the points which bring larger noise are deleted, and the points with smaller noise are retained, thus the precision of our method is increased. The experimental results indicate the new method is precise in calculation, stable in performance and resistant to noise.
1997, 19(4): 476-482.
Abstract:
In reference of the theory of adaptive IIR filters, the paper puts forward the recursive structure and quasi-recursive structure of Adaptive Volterra Filters(AVF), and discusses their characteristics and areas of applications. The introduction of recursive structure can remarkably reduce the parameters and computational cost of AVF. The on-line identification algorithm of quasi-recursive structure AVF with its application in non-linear system identification and the filtering algorithm of recursive structure AVF with its application in non-linear correlated noise cancellation are also given. In simulations, the above algorithms are compared with multi-layered perceptron and non-recursive AVF. The results show the algorithms of the paper have obvious advantages both in performance and in computational cost.
In reference of the theory of adaptive IIR filters, the paper puts forward the recursive structure and quasi-recursive structure of Adaptive Volterra Filters(AVF), and discusses their characteristics and areas of applications. The introduction of recursive structure can remarkably reduce the parameters and computational cost of AVF. The on-line identification algorithm of quasi-recursive structure AVF with its application in non-linear system identification and the filtering algorithm of recursive structure AVF with its application in non-linear correlated noise cancellation are also given. In simulations, the above algorithms are compared with multi-layered perceptron and non-recursive AVF. The results show the algorithms of the paper have obvious advantages both in performance and in computational cost.
1997, 19(4): 483-488.
Abstract:
In this paper, a modified Block Matching Algorithm(BMA) with motion correlation constraint is proposed at first. Then a novel motion estimation algorithm, which computes motion vector for each pixel by interpolating motion vectors is presented. In order to increase inter-frame prediction gain and decrease the computational complexity, a motion vector optimizing algorithm for local image region is described at last. Experimental results show that the proposed algorithm can improve the prediction performance obviously with a moderate increased complexity compared with the conventional BMA.
In this paper, a modified Block Matching Algorithm(BMA) with motion correlation constraint is proposed at first. Then a novel motion estimation algorithm, which computes motion vector for each pixel by interpolating motion vectors is presented. In order to increase inter-frame prediction gain and decrease the computational complexity, a motion vector optimizing algorithm for local image region is described at last. Experimental results show that the proposed algorithm can improve the prediction performance obviously with a moderate increased complexity compared with the conventional BMA.
1997, 19(4): 489-493.
Abstract:
Using the secret-sharing schemes, especialy the (k,n)-threshold schemes as tools, one method of constructing A2-Code from the general A-Code was given. It was shown that the new codes have all the good properties of the old ones, such as the known information-theoretical and combinatorial bounds of the latters, although asking for some data bits extension. Clearly, this method shows a new dimension of how to design the secret and efficient A2-code.
Using the secret-sharing schemes, especialy the (k,n)-threshold schemes as tools, one method of constructing A2-Code from the general A-Code was given. It was shown that the new codes have all the good properties of the old ones, such as the known information-theoretical and combinatorial bounds of the latters, although asking for some data bits extension. Clearly, this method shows a new dimension of how to design the secret and efficient A2-code.
1997, 19(4): 494-502.
Abstract:
The periodic cell stream is a very important member Among the input traffic sources in ATM networks. In this paper, a finite-buffered ATM multiplexer with traffic sources composed of a periodic cell stream, multiple i. i. d Bernoulli cell streams and bursty two-state Markov Modulated Brenoulli Process(MMBP) cell streams is exactly analyzed. The probability mass function of queueing delay, the autocorrelation and power spectrum of delay jitter for this periodic cell stream are derived. The analysis is used to expose the behavior of delay jitter for a periodic cell stream through an ATM multiplexer in a bursty traffic environment. The simulation results indicate that the analytical results are accurate.
The periodic cell stream is a very important member Among the input traffic sources in ATM networks. In this paper, a finite-buffered ATM multiplexer with traffic sources composed of a periodic cell stream, multiple i. i. d Bernoulli cell streams and bursty two-state Markov Modulated Brenoulli Process(MMBP) cell streams is exactly analyzed. The probability mass function of queueing delay, the autocorrelation and power spectrum of delay jitter for this periodic cell stream are derived. The analysis is used to expose the behavior of delay jitter for a periodic cell stream through an ATM multiplexer in a bursty traffic environment. The simulation results indicate that the analytical results are accurate.
1997, 19(4): 503-509.
Abstract:
The nonlinear coupled differential equations of dimensionless beam width parameters of two Gaussian optical beams with orthogonal polarization in a family of Kerr-law nonlinear reversal parabolic graded-index rod lenses are derived by using a variational approach based on the scalar Maxwell s equations and then solved. The characteristics of fields in the rod lens and object-image transformation through the rod lens of two orthogonally polarized optical beams are investigated.
The nonlinear coupled differential equations of dimensionless beam width parameters of two Gaussian optical beams with orthogonal polarization in a family of Kerr-law nonlinear reversal parabolic graded-index rod lenses are derived by using a variational approach based on the scalar Maxwell s equations and then solved. The characteristics of fields in the rod lens and object-image transformation through the rod lens of two orthogonally polarized optical beams are investigated.
1997, 19(4): 510-515.
Abstract:
High speed Multi-Chip Module (MCM) has been widely adopted in the complex system, where Simultaneous Switching Noise is one of the key factors affecting the system function. In this paper, two dimensional electromagnetic model is used to simulate power/ground plane simultaneous switching noise in MCM. A novel numerical method for the time-domain electromagnetic problems, named Method of Characteristics, is presented and used to solve the above problem. The results obtained is in accordance with that by the method of FD-TD and that reported by literature.
High speed Multi-Chip Module (MCM) has been widely adopted in the complex system, where Simultaneous Switching Noise is one of the key factors affecting the system function. In this paper, two dimensional electromagnetic model is used to simulate power/ground plane simultaneous switching noise in MCM. A novel numerical method for the time-domain electromagnetic problems, named Method of Characteristics, is presented and used to solve the above problem. The results obtained is in accordance with that by the method of FD-TD and that reported by literature.
1997, 19(4): 516-521.
Abstract:
Based on Fresnel-Kirchhoff theory, the electromagnetic scattering from fractal gratings, which have structure on uniform Cantor sets or Siepinski carpet, is discussed. The scattering intensities of 1-dimensional and 2-dimensional fractal grating with different fractal dimension are calculated and their properties on spatial frequency distrubution are then analysed. The numerical results illustrate the fractal characteristics and frequency selectivities.
Based on Fresnel-Kirchhoff theory, the electromagnetic scattering from fractal gratings, which have structure on uniform Cantor sets or Siepinski carpet, is discussed. The scattering intensities of 1-dimensional and 2-dimensional fractal grating with different fractal dimension are calculated and their properties on spatial frequency distrubution are then analysed. The numerical results illustrate the fractal characteristics and frequency selectivities.
1997, 19(4): 522-526.
Abstract:
Following the strictly solving the continuity equation of the photo-induced carriers, the FDTD analysis for millimeter wave propagation in an optically controlled rectangular dielectric waveguide was presented. Excellent agreements betweeen numerical results and experiments were found.
Following the strictly solving the continuity equation of the photo-induced carriers, the FDTD analysis for millimeter wave propagation in an optically controlled rectangular dielectric waveguide was presented. Excellent agreements betweeen numerical results and experiments were found.
1997, 19(4): 527-531.
Abstract:
A generalized spectral Green s function formulation of N-layer superstrata structure is given with three dimension sources, which is a set of closed form formulas. A full-wave analysis model for rectangular microstrip antennas covered with N-layer dielectric has been established by using the above spectral Green s function. The unknown surface current density on the microstrip patch for such structure is found as the solution of an integral equation. The input VSWR and radiation patterns of antenna are also obtained. The numerical results have been verified by the experimental results.
A generalized spectral Green s function formulation of N-layer superstrata structure is given with three dimension sources, which is a set of closed form formulas. A full-wave analysis model for rectangular microstrip antennas covered with N-layer dielectric has been established by using the above spectral Green s function. The unknown surface current density on the microstrip patch for such structure is found as the solution of an integral equation. The input VSWR and radiation patterns of antenna are also obtained. The numerical results have been verified by the experimental results.
1997, 19(4): 532-537.
Abstract:
It is very effective that use BDD to describe the synchronous circuits. This paper has proposed the reducing way for BDD in order to collapse the number of inputs, routes and states. Based on the features of circuit, several heuristic methods that speed up verification are presented.
It is very effective that use BDD to describe the synchronous circuits. This paper has proposed the reducing way for BDD in order to collapse the number of inputs, routes and states. Based on the features of circuit, several heuristic methods that speed up verification are presented.
1997, 19(4): 549-552.
Abstract:
The concept of dyadic addition family is presented. The relations between the families of dyadic complementary binary sequences and dyadic addition family, and the existing condition of the families of dyadic complementary binary sequences are studied. The method of searching the families of dyadic complementary binary sequences by computer is given.
The concept of dyadic addition family is presented. The relations between the families of dyadic complementary binary sequences and dyadic addition family, and the existing condition of the families of dyadic complementary binary sequences are studied. The method of searching the families of dyadic complementary binary sequences by computer is given.
1997, 19(4): 553-557.
Abstract:
This paper generalizes the definition of generalized Hamming weights for linear codes to nonlinear codes, derives a generalized Elias bound, and gives the weight hierarchy of linear constant weight codes.
This paper generalizes the definition of generalized Hamming weights for linear codes to nonlinear codes, derives a generalized Elias bound, and gives the weight hierarchy of linear constant weight codes.
1997, 19(4): 558-561.
Abstract:
This paper presents a new conception-the brain function state diagram (BFSD) or brain fucntion state mapping (BFSM) which describes the function cases of brain and gives the method for finding BFSD or BFSM from the brain electrical wave (BEW). In this paper the brain function state diagrams of rabbits are normal and epileptic fit. The results indicate that BFSD which shows the informations on epileptic fit are sufficient, marked, immediate and simple.
This paper presents a new conception-the brain function state diagram (BFSD) or brain fucntion state mapping (BFSM) which describes the function cases of brain and gives the method for finding BFSD or BFSM from the brain electrical wave (BEW). In this paper the brain function state diagrams of rabbits are normal and epileptic fit. The results indicate that BFSD which shows the informations on epileptic fit are sufficient, marked, immediate and simple.
1997, 19(4): 562-565.
Abstract:
In this paper, the method of the fractal interpolation is studied and is used for the image enlargement and compression coding. Some examples show the good results and high computing efficiency.
In this paper, the method of the fractal interpolation is studied and is used for the image enlargement and compression coding. Some examples show the good results and high computing efficiency.
1997, 19(4): 566-569.
Abstract:
In microwave remote sensing quantitative study, it needs to know dielectric properties of remote sensing objects precisely. The free-space method was used to in situ measure typical objects permittivities in the fields for obtaining precise data of fitting for remote sensing real objects. The basis and calibration method of free-space method were given. The permittvities of black soil, meadow, and ice with the method were presented. The results were analysed.
In microwave remote sensing quantitative study, it needs to know dielectric properties of remote sensing objects precisely. The free-space method was used to in situ measure typical objects permittivities in the fields for obtaining precise data of fitting for remote sensing real objects. The basis and calibration method of free-space method were given. The permittvities of black soil, meadow, and ice with the method were presented. The results were analysed.
1997, 19(4): 570-573.
Abstract:
A switched-current(SI) modified universal building-block(MUB) is proposed with low sensitivities as well as positive and negative outputs. Various basic operations of sampled-data signal processing can be performed by the MUB. Based on this building-block, a new SI high-Q biquadratic filter is generated from the switched-capacitor prototype by the means of signal flow graphes, which can obtain low-pass, high-pass, bandpass and band-reject filtering functions.
A switched-current(SI) modified universal building-block(MUB) is proposed with low sensitivities as well as positive and negative outputs. Various basic operations of sampled-data signal processing can be performed by the MUB. Based on this building-block, a new SI high-Q biquadratic filter is generated from the switched-capacitor prototype by the means of signal flow graphes, which can obtain low-pass, high-pass, bandpass and band-reject filtering functions.
1997, 19(4): 574-576.
Abstract:
In this paper the design and implementation of knowledge based geographic information system are discussed. An object oriented relation frame semantic data model and new physical data structure are proposed. In the system advanced distributed tactics is used to implement data distributed query and update between different nodes. Also the system possesses the functions of the computer aided decision based on knowledge and models and the friendly adaptive human computer interface. The test shows that the system has the advantages, such as high intelligence, good distributed functions and convenient for use.
In this paper the design and implementation of knowledge based geographic information system are discussed. An object oriented relation frame semantic data model and new physical data structure are proposed. In the system advanced distributed tactics is used to implement data distributed query and update between different nodes. Also the system possesses the functions of the computer aided decision based on knowledge and models and the friendly adaptive human computer interface. The test shows that the system has the advantages, such as high intelligence, good distributed functions and convenient for use.
1997, 19(4): 538-543.
Abstract:
First, the significance and the importance of SAR raw data compression is discussed. Then the brief history of its developmant is reviewed. Two schemes of data compression namely block adaptive quantization (BAQ) and vector quantization (VQ) is described. Finaly, the prospects for synthetic aperture radar raw data compression is looked forward.
First, the significance and the importance of SAR raw data compression is discussed. Then the brief history of its developmant is reviewed. Two schemes of data compression namely block adaptive quantization (BAQ) and vector quantization (VQ) is described. Finaly, the prospects for synthetic aperture radar raw data compression is looked forward.
1997, 19(4): 544-548.
Abstract:
The idea about undulating propagation of electromagnetic wave is presented. And some characters and operating principle to form undulating propagation of electromagnetic wave by multifrequency radiation source with initial twisty phase angle continually varying versus the time are introduced. The electromagnetic wave propagation in the undulating pattern along the propagating path will profit many fields of the electromagnetic wave application, such as point to point communication, electromagnetic energy conveyance,detection and ranging etc.
The idea about undulating propagation of electromagnetic wave is presented. And some characters and operating principle to form undulating propagation of electromagnetic wave by multifrequency radiation source with initial twisty phase angle continually varying versus the time are introduced. The electromagnetic wave propagation in the undulating pattern along the propagating path will profit many fields of the electromagnetic wave application, such as point to point communication, electromagnetic energy conveyance,detection and ranging etc.
