Based on the recently developed chaos and multifractal theory,this paper introduces new methods for radar signal processing,radar targete characteristic analysing and radar target recognition.The Lyapunov exponents are calculated.With the correlationalintegral method,the multifractal dimensions of radar signal are further obtained.Then by the ART2 neural network,the radar target recognition is carred out.The results of this paper show that the chaos,multifracted and neural network theory have great potentials in the fields of radar signal processing and radar target recognition.

A new method for radar target recognition based on modular neural networks is reported in this paper.In this method,the response from an unknown target is first sent to several waveform predicators that the some BP neural networks trained by responses from known targets respectively.Then the predicator erros are inputted to a classfier using the rule of maximum a posteriori or the rule of modified minimum squared errorrs.The simulation of PC computer shows that the new method has a good performance on radar target recognition.The method also has other advantages such as easy realization,clear structure and easy expansion.

For linear and non-linear frequency modulation (LFM/NLFM) signals with large time-bandwidth product (TB),radar super-resolution in range domain at low SNR is investigated.A novel signal enhancement algorithm based on FFT for LFM/NLFM signals used for preprocessing,which significantly raises the effective signal-to-noise ratio(SNR),is presented.By use of decorrelation between echoes of different targets achieved by pulse-to-pulse carrier frequency hopping,a vvariety of approaches are suggested to further lower the SNR threshold for range super-resolution.Monte-Carlo simulation results are presented to illustrate the inproved performance.

In this paper, a PD radar signal is researched, which is modulated by randomly staggered pulse sequence whose phase is also modulated by random binary-phase code. The expression of its ambiguity function is firstly presented, the main derived method is also given, and how to generate this pulse sequence is discussed. Finally simulation results are analyzed.

This paper analyzed the principle of maximum entropy method in QCW radar, advanced the requirements of the maximum entropy spectral estimation in test radar system, designed and implemented the arithmetic unit by C31 chip, and accomplished real-time control for the maximum entropy spectral estimation circuit. Data transmission rate can reach 40Mb/s and data handling capacity arrives at IMb/s when the order of algorithm is 8. Researches this paper made will promote QCW radar from theory to application.

In this paper a mutual coupling calibration method suitable for uniform linear arrays is presented where only one known calibration source in the near field is necessary for effective calibration, which can be extended to the case of uniform circular arrays. The paper has also discussed the position error influence of the source on the calibration., and the optimum azimuth of the source is analized. Finally some simulation results are given to show the advantages of the new method.

In this paper, the problem of optimal reception of partially polarized signals is investigated by means of Stokes vector description of EM waves. The receiving antenna s global optimal polarizations are derived using Lagrange multiplier method with SINR(Signal-to-Interference-and-Noise-Ratio) and PDSI(Power-Difference-between-Signal-and-Interference) as the objective functions respectively. At the end of this paper, results are verified by a numerical example.

The performance analysis of quadrature correlation detection algorithm for DS-SS signal is presented. The algorithm can detect effectively both BPSK and QPSK spread spectrum signals. The analytical results are confirmed by computer simulation. The simulation results show that the algorithm can detect DS-SS signals with the input SNR of -15dB.

In this paper,the updating equations of FBLMS and UFBLMS algorithms are transformed into the time domain,then two algorithms(FOBA and UFOBA) are presented by employing time-varying convergence factors which are optimized in a least-square(LS) sense respectively.Although FOBA and UFOBA algorithms require a relatively modest increase in computation for each block iteration compared to FBLMS and UFBLMS algorithms respectively,it is shown by simulations that these algorithms improve convergence speed and accucary of adaptation.

A block feature matching (BFM) criterion function used for block motion estimation in some the international standards such as H.261, H.263, MPEG1, MPEG2, and HDTV codecs, is proposed in this paper. The complexity and VLSI implementation of real-time block matching motion estimation need to involve the following considerations: the motion searching complexity within a given search window, the basic matching complexity for each matching operation, and the required amount of data fetching from frame buffer to the motion estimation processor. Simulation indicated that the proposed BFM function is very simple and effective, thus the corresponding complexity of block matching computation and the time of data fetching are greatly reduced. Moreover, the BFM function can provide the potential of parallel VLSI implementation. Comparisons of the BFM with other criterion functions are also given in detail.

A motion adaptive frame interpolation algorithm for progressive frame rate conversion of digital video systems is proposed. Firstly a motion vector is searched for each pixel in the interpolated frame using the block-matching search method of linear prediction. Then the resulting motion field is median filtered and differentiated to remove inconsistent or wrong vectors. Finally averaging along the motion trajectory is used to get the interpolated pixel values. Simulation results indicate the proposed method with fast search and differentiation in a large area can solve the motion blurring of fast moving object and the scene overlapping of a scene change in the interpolated frame, and the corresponding objective quality of the interpolated frame is also improved.

In this paper, The properties of Tseng runs in RCF of a binary sequence are studied. The relation between Tseng runs and run numbers of different run lengths in sequence are derived, and some new conceptions: run catena, concatenated elements, etc., are introduced for proving the theorems in this paper. And these theorems not only clearly describe that the sidelobes of CF are influenced by the run numbers of different run lengths in sequence, but also discover the existence of the typed constant, which may be used to determine the type of code words.

In this paper, the definition and properties of Line Spectral Frequency(LSF) parameters are introduced briefly. An efficient solution method of obtaining LSF parameters is presented. A new Split Vector Quantizer(SVQ) of LSF parameters is designed in LPC speech coding. The objective and subjective evaluations show that the proposed scheme offers transparent quantization quality with 25 bit/frame.

The chaotic sequence is sensitively dependent on initial conditions and parameters of the chaotic system. This paper, firstly, presents a multiple access communications system model which is based on chaotic sequence; Secondly, introduces a chaotic circuit, and analyzes the multple access communication which is based on chaotic sequence; Finally, gives the simulation results which illustrated that chaotic sequence may become a good pseudo-random code to M sequence and gold sequence in the multiple access communications.

An exact matrix-geometric solutions using Geom/PH/1/K queueing model are presented for analysing a backpressure input/output queueing ATM (Asynchronous Transfer Mode) switch under uniform traffic. It shows valuable results for designing a broadband packet switch.

Hash algorithm is always with high operation speed such as MDx, SHA. A natural problem is that could one apply for a fast hash scheme to construct a cipher block algorithm. This paper denotes such a problem as dual problem of hash algorithm. Based on the known result, it is proved that several fast secure block cipher can be constructed if a fast secure hash algorithm is given.

A new algorithm for symbol timing recovery using baudrate sampling is presented, which does not demand strict requirements for the AGC circuit in the front part. Simulations show that estimate performance can be greatly improved using this algorithm.

This paper presents a systematic and rigorous discussion for the solution of maximum power transmission between two plannar aperture antennas using a Rayleigh quotient for the power transmission efficiency established in the paper. An analytical solution for the system consisting of two plannar circular antennas and related numerical results useful for engineering design have also been given.

Numerical models for the physical parameters and physical effects about numerical simulation of low-temperature bipolar transistor are discussed. The numerical approaches required for simulation of low-temperature behavior are presented. A simulation program has been given to investigate bipolar transistor behavior range of 77-300K. Finally, the characteristics of a typical bipolar transistor are simulated at 300K and 77K.

A new soft-output algorithm of decoding is derived from the Bahl algorithm. The complex operations of estimating posterior probabilities and difference metric in the Bahl algorithm are paralleled and well formulated into a set of simple matrix operations which are fit to design efficient VLSI circuits. The probability of u_k ≠ u_k is quantized with n_δ bits and is output together with u_k branch code X_k. Thus, the algorithm not only increases the decoding speed, but also provides a soft-output algorithm for multi-level concatenated code.

criterion for designing chaotic spreading sequences based on symbolic dynamics is proposed. A family of mutually independent Bernoulli sequences can be generated by this approach. A condition for the metric entropy to obtain q-ary Bernoulli sequences is proved.

A new vector quantization(VQ) approach based on genetic algorithm(GA) is presented in this paper. VQ codebook design is essentially a classification of training vectors. Because of GA s global optimum ability, using GA for codebook training can obstain a global optimum codebook, and overcome local optimum limitation of traditional algorithms. This method is also independent on initial conditions, more robust, highly regular and parallel in architecture.

A modified adaptive filter algorithm of smooth pseudo LMS(SPLMS) based on pseudo LMS(PLMS) is presented in this paper. Besides having advantages of PLMS, it has faster beginning convergence rate, lower misadjustment of weight, and better robustness aganist noise and disturbance. The theoretical analysis is confirmed by computer simulation results.

In order to overcome slow convergence rate of the standard BP algorithm, this paper presents an adaptive backpropagation algorithm which can update learning rate and birr factor automatically based on dynamic training error rate of change. Simulation result of the XOR problem shows much faster convergence rate can be obtained by this algorithm.

Sparse-array Synthetic Impulse and Aperture Radar (SIAR) can isotropically radiate so that the long-time coherent integration can be obtained. But there are some problems in the process. An effective method of the long-time coherent integration is presented, which includes simply implemented motion compensation and time-frequency analysis. The result shows that the novel radar (SIAR) system has better performance for weak target detection.