The concern here is retrieval of multiple tone harmonics observed in complex-valued multiplicative noises with nonzero mean. Cyclic mean statistics have proved to be useful for harmonic retrieval in the presence of complex-valued multiplicative noises with nonzero mean of arbitrary colors and distributions. Performance analysis of cyclic estimators is carried through and large sample variance expressions of the cyclic estimators are derived. Simulations validate the large sample performance analysis.

A new quadrature sampling technique for arbitrary bandpass signal within baseband sampling rate is presented. The input bandpass signal whose carrier frequency lies in the A/D baseband sampling rate is first decimated by factor 2 and modulated by (-1)ⁿ, and then is interpolated by a linear phase FIR all-pass filter, finally the modulated complex envelope of bandpass signal can be produced.

A method is presented for real-time estimation of the frequency, azimuth and elevation angles of the signal incident on uniform circular arrays(UCA) over a very wide frequency band(1~18GHz) in DFT beamspace. The method provides unambiguous frequency estimation under temporal undersampling and unambiguous azimuth and elevation estimation using integer search algorithm under spatial undersampling. The estimations of azimuth and elevation angles are automatically paired with corresponding frequency estimates. The variance of the frequency estimates of indirect method is 2～3 magnitudes less than that of the direct method.Simulations are presented demonstrating the capabilities of the method.

The sampling theorem for bandpass signal is the theoretical basis for design of all digital quadrature detectors. There are various expressions for this theorem in previous references, which has greatly affected the optimum design of all digital quadrature detectors. The correct expression is introduced in this paper, and the optimum design of all digital quadrature detectors is discussed.

This paper describes the implementation of the corrector of I & Q errors in coherent processor. The correcting network consists of combination logic circuit and it can correct the gain and phase imbalances and the bias errors of the in-phase and quadrature channels in coherent signal processing. The correcting coefficients are computed with DSP using a test signal. The image level without correction is about -27dB if the errors of gain and phase in the coherent processor are 0.1dB and 5° respectively. The experimental results show that the image level is reduced to -52dB from -27dB after correcting the errors.

Potential is an important conception in Gibbs random field (GRF) model. In this paper, an approach for texture analysis based on maximum potential is presented. Experimental results prove that this approach can be an effective means for texture analysis.

The problems of parameter selections for potential function used to initialize cluster centers are discussed, and two methods are given for determining these parameters. Then a new density function to initialize cluster centers is also given which is computational effective. Finally, a set of compared experiments is presented to show the effectiveness of the proposed methods.

In this paper, an efficient engineering classification of ship noises based on a local adaptive wavelet and Gauss neural network synthesis classification system is presented. The classification systems combine two methods of feature extraction and classification, which are adaptive wavelet neural network and adaptive Gauss neural network. It is capable of learning new types of signals and not destroying the learned network. The classification system is used to extract automatically feature from and classify for noises radiated from actual three types of ships. The classified results are encouraging, and this method is proved to be superior and efficient engineering application in the future.

Some new results about estimation of attraction domain of memory patterns and exponential convergence rate of the network trajectories to memory patterns for Hopfield continuous associative memory are obtained by using of some new technics and Lyapunov method. These results can be applied to evaluate the error-correction capability of Hopfield continuous feedback associative memory and to synthesze the procedures of Hopfield continuous associative memory neural networks.

A new method of adaptively controlling the sampling rate of tracking is proposed in this paper. The analytic relationship between the sampling interval and the accuracy and model parameters of filter is applied.The problem of deciding the amount of the sampling interval is thus solved when the sampling rate is required to be controlled. It is shown by the simulation tests that the default of heavy computation budern in the methods of varying sampling rate based on multiple models is overcome meanwhile the prime advantages of those methods is conserved.

The designing of step frequency signal is concerned.Stochastic sequence step frequency signal is advanced to overcome the ordered step frequency signal s default of range-velocity coupling. The advanced signal is better than the ordered step frequency in clutter rejection and range-velocity resolution.

A new motion estimation algorithm based on spatio-temporal cohenence is presented. In this proposed algorithm, an initial motion vector of the block which has good coherence with its neighbors or that of the previous frame is directly predicated. A more precise motin vector is obtained based on the initial one by full-searching block matching algorithm (FBMA) in a small range. As to blocks with poor coherence, FBMA with ordinary range is used, and their motion vectors are at last adjusted by the coherence criterion given in the paper. Experiments show satisfying results can be reached with the proposed algorithm.

Aiming at the problems existing in the present frequency planing, this article proposes a method to overcome the obstacle of handwork frequency planning and high demand of electronic map, using expert system and genetic algorithm. So that it gives a more precise frequency planning scheme based on expert experience, knowledge and genetic algorithm even lacking of accurate electronic map.

A method (VP-PRMA) based on the users demand to improve the performance of voice service in PRMA system is proposed. In the paper, a new system measurement based on relative penalty factor is used. After the analysis of PRMA, the new method, using variable contend probability according to the actual condition, is presented. According to the simulation, the proposed method improves the voice performance as expectated.

The principle of synchronization in FD/SSMA (Frequency Diversity/Spread Spectrum Multiple Access) system with DFT (Discrete Fourier Transform) implementation is studied. It is shown that both byte-timing synchronization (i.e. the time-window of DFT transform) and the transmitter carrier synchronization are needed in this FD/SSMA system. The capturing, locking and tracing of timing-synchronization, the carrier s synchronization about its frequency and phase will be finished in order by the recognizing signals of timing-synchronization and three error signals, of which the timing-error signal, the frequency and phase error signals of the system’ carrier are included, where their implementation may be carried out easier due to the DSP (Digital Signal Processing) techniques to be applied.

In this paper, Malvar wavelet transform is implemented in MCM (Multicarrier Modulation) system. It is proposed that assigning the bandwidth nonuniformly using Malvar wavelet transforming is to combat with the multi-tone noises existing simultaneously in MCM system and to improve the efficiency of utilizing the bandwidth.

This paper uses factoring method for computing rooted communication reliability of networks, i.e., the proability that there are operating paths from the root vertex to all other vertices. Two new reliability-preserving reductions and an edge-selection strategy are presented by using the topological structure of acyclic directed networks. Based on that, a factoring algorithm is developed. It is very efficient for networks which are not very dense.

This paper presents a synthesis methodology for ECL circuits based on a mixed voltage-current signal representation and operation denned on the voltage and current signals. The ideas presented in this paper are then demonstrated on the design of an ECL 1-bit full adder. The paper concludes by presenting an algebra system which is suitable for current signal representation and operation on currents.

The determination of the current distribution on the dipoles of microstrip antennas and the coupling between elements of arrays are especially important for the analysis of microstip antennas and arrays.Unfortunately,the calculation is tedious and time-consuming because of the existing of the Sommerfeld integrals.A new method is proposed in this paper to calculate the current distribution by using Chebyshev decomposition to deal with the sommerfeld integrals.The calculation results show that the presented method greatly improves the computing efficiency while retaining high accuracy.

A new method of truncation MR artifact reduction is proposed in this paper. In the method, any digital signal spectrum can be expressed as the weighted sum of spectrum functions related to the signal singularity locations. The singularity locations and weight coefficients can be extracted from MR truncation spectrum data. Thus we can gain the high spectrum using the sum and reduce the truncation artifact in MR image. Experiment and simulation results show that this method gives much better reconstructed image than available methods.

This paper describes the characterization of texture properties at multiple scales and orientations using the wavelet transform, and introduces a new wavelet feature suitable for textured image classification. It is pointed out that the new feature is superior to conventional energy measurement by analyzing its stability and its visual proterty in detail. Finally, nine kinds of natural images are classified successfully based on wavelet feature using BP neural network. The results demonstrate natural textured images can be classified without error and done at higher correct classification rate under white noise.

A new method for choosing biorthogonal wavelet bases which is suitable for image coding is presented. A thoroughly comparative study is made and biorthogonal wavelet bases suitable for image coding are selected by using several criteria of evaluation proven to be closely related to image coding. It is verified by coding experiment.

This paper analyzes the reason for low convergence rate of BP network standard backpropagation algorithm to nonblanced training set, researches how to improve the learning algorithm in order to increase training speed, and through two experiments tests and verifies above theory.

In this paper, a new method of communications using chaotic signal is presented. This method is that the slope of the non-linear resistance in Chua s circuit is modulated with the input signal. At the receiving end, the signal is demodulated using the inverse course. Finally, the basic circuit and simulation result are given.

The paper presents an algorithm for detecting interference in adaptive filter, because interference can deteriorate adaptive filter performance. The algorithm can improve the performance of adaptive filter under interference environment. The computer simulating results are given in this paper.

FDTD coupled with Matrix Pencil method was applied for computation of the resonant frequency and external quality factor of waveguide loaded cavities. With the new method, the computing time is reduced considerably and the accuracy of numerical results are improved compared to the combined approach of FDTD and DFT. Compared with conventional used Prony s method, the advantages of MP technique are its more computational efficiency and less sensitivity to noise. Numerical results of a 3-D rectangular cavity are given for verification of the method.

A scheme of FDTD near to far zone transformation for ground wave scattering from objects above a flat earth is presented in this paper. With the equivalent principle, the equivalent electric and magnetic currents can be obtained from FDTD near field output boundaries. The fields radiated by these currents in the presence of a flat earth are then calculated using the reciprocity theorem and the known lateral wave formula of a vertical dipole(King, et al. 1992). Numerical results show excellent agreement between this scheme and analytical method. The feasibility of another approximation approach (Q.L.Li, et al. 1997) is also tested by this scheme.