In this paper, SAR ATR (Synthetic Aperture Radar Automatic Target Recogni-tion) approach based on KPCA (Kernel Principal Component Analysis) is proposed. KPCA first maps the input data into some feature space using kernel functions and then performs lin-ear PCA on the mapped data. It takes the principal components in nonlinear space as sample features, then SVM classifier is used to classify targets. Experimental results with MSTAR SAR, data sets provided by the US DARPA/AFRL (Defense Advanced Research Projects Agency/Air Force Research Laboratory) show a better performance of classification and generalization.

This paper presents a new SAR raw data simulation method that is used to test and analyze the algorithm of moving target detection. The simulated SAR raw data are generated by superimposing the simulation moving target echo to a real background reflection data. The compound data are tested with the R-D imaging algorithm and MTI algorithm .The result validates the simulation method.

Based on the forward-squint-looking SAR mode of SAR system, this paper gives the mathematical model of SAR echo signal, and carefully analyses the instantaneous azimuth frequency , the instantaneous Doppler frequency component of the azimuth frequency and the impact of range chirp on azimuth frequency, then combining with the common chirp scal-ing algorithms, proposes an improved chirp scaling algorithm for forward-squint-looking SAR, imaging, and takes the simulated test with the algorithm. The results of theory and simula-tion show that the way is more reasonable for forward-squint-looking SAR geometric relation and can effectively complete range cell migration correction, range compression and azimuth focusing, and improve images’ quality.

The simulation of radio holographic signal for moving objects acts an important role in the study of the target imaging and recognition in the research of the SISAR. A gener-alized expression is induced on the basis of analyzing the relation between the shadow of the moving target and its radio holographic signal. Further, each component in the expression has explicit physical signification. The validity of the method is shown by the comparision of the simulation results with the experimental results of two air moving targets in the same condition.

Bayesian network is a new inference and express method of uncertain knowledge. It is proposed an inference and express technique for remote sensing imagery data which has complexity and uncertainty based on Bayesian Network Model(BNM). In the paper, the LU data, TSP and LST/Albedo data of AVHR.R time-sequence imagery which get from the project of China-Japan Asian dust storm in 2002 are used to analyze the dust storm and at the same time BNM is used to describe the knowledge and information inference. The satisfied results are given in the paper with the method.

A new method for analyzing nonstationary signal designated as local wave analysis is introduced. It stems from the concept of instantaneous frequency, and it can describe accu-rately the local properties of dynamic data in the time-frequency plane. The basis of local wave decomposition is not fixed but changeable with the changing wave shape of the dynamic data. The decomposition is adaptive, and therefore highly efficient. Compared with the available time-frequency analysis methods, this method has many advantages. It is commonly used for analyzing nonstationary signal. The application of this method is discussed, and the related problems in this field that need further study are pointed out, too.

In this paper, a fast recursive algorithm for the 2D sliding rectangular window FFT is proposed, based on the relationship of neighbor windows and the transformation property of the FFT. Further, the fast recursive algorithm is extended into 2D Gabor transformation of sliding rectangular windows. An analysis and a comparison on the computational complexity between the recursive algorithm and the traditional direct method are given. Both the theoret-ical analysis and the experimental result show that the new recursive algorithm can reduce the computation cost greatly, especially in the case of bigger image size or/and bigger window size. The proposed recursive algorithm can improve the computation efficiency of image analysis using the slide rectangular window FFT or Gabor transform.

Most existing content-based image retrieval systems using low-level features that could not describe high-level semantics thoroughly and accurately. In this paper, a novel system for content-based image retrieval is designed and created, which combines image semantics based on a multi-level model for image description. In this image description model, image contents could be analyzed and represented through different levels and the transition from low-level features to high-level semantics is thus achieved. Corresponding querying mechanism and feedback are also proposed based on this image model. Aiming at object semantics in image, this querying mechanism is much closer to human beings’ understanding of image contents so that it provides a convenient and effective querying procedure. The feedback used in the system is a self-adaptive relevance feedback based on object descriptions, it permits to propose different querying schemes according to the different demands raised by various users, and thus optimal results could be refined.

A digital watermark is an invisible mark embedded in digital images,video or audio documentSjWhich may be used for a number of different purposes including copyright protec-tion,authentication and captioning, In this paper, an effective adaptive watermarking algorithm for still images is proposed. This is a blind watermarking which can confirm the copyright with-out original image. The original image is transformed using wavelet transform and the water-mark is embedded in selected subband of frequency domain according to neighboring symbol’s mean value and odd-even adjugement rule. Experimental results and attack analysis show that the watermark algorithm proposed in this paper can produce the watermarked image which has high fidelity and is robust against, some image processing operations, such as JPEG lossy compression,median filtering, additive noise, scaling, and cropping attacks.

To reduce the incorrect association rate using NN (Nearest’, Neighbor) algorithm in complex environment in clutter, a new plot-track association algorithm-Approximate Multi-Sensor multi-target Joint Probabilistic Data Association (AMSJPDA) is presented in the pa-per. It uses all the measurements in the tracking gate and every measurement has its own power, Added the measurements multiplied by their power the near optimal track estimation is achieved. AMSJPDA, based on the Approximate probabilistic Computing (AC) and Direct probabilistic Computing (DC) brought forward by B. Zhou, is the amelioration of MS JPDA and demands less time than MSJPDA. It meets the need of large scale plates and the real-time performance of data fusion system. At the end of the paper the comparison result of AMSJPDA and the NN is given.

In this paper, the existence, uniqueness and global exponential stability of the equilibrium point of a class of Hopfield neural networks are investigated by using Liapunov direct method. Lipschitz continuous activations are considered and the limit on self-feedback is released. It can be nonlinear instead of linear in this paper. And under this condition, some sufficient conditions and convergence estimate for global exponential stability of neural networks are obtained.

This paper presents a novel method based on Tabu algorithm to design the fi-nite wordlength FIR digital niters, which is a deterministic paralell optimization algorithm simulating human intelligence. Firstly, Tabu algorithm is detailedly described and then Tabu optimization algorithm is used to design finite wordlength FIR digital filters. Finally, the sim-ulation results are given.

A channel estimation scheme, using measured autocorrelation coefficients as weighting coefficients of the channel processor, and its error performance in equation are presented in this paper. Error performance of moving averaging channel estimation in equation is provided as well. Numerical and simulation results show that the proposed scheme can trace channel accurately and adaptively and outperforms moving averaging methods under not only fixed but also variant vehicle speeds.

SIR-based preamble detection is investigated under multipath fading condition. Two-antenna receiver is designed for the random access channel and common packet channel in WCDMA uplink. The preamble detection performance is derived and closed-form expressions are given. Three utilization methods of antennas’ signals are compared by theoretical analysis and computer simulations.

The performance of Turbo Product Coded Pilot Symbol Assisted Modulation (TPC-PSAM) system is studied and a novel method to improve the performance of TPC-PSAM in Rayleigh fading channels by clipping the soft input information of decoder is presented in this paper. From the numerical simulation in different Rayleigh channels, it is concluded that the TPC-PSAM system with clipping soft information autperforms from 5dB to 8dB over the TPC-PSAM system at the bit error rate of 10^-4.

This paper presents the effect of carrier phase error on MC-CDMA performance in downlink mobile communications. Signal-to-Noise Ratio (SNR) and Bit-Error-Rate (BER.) is analyzed taking into account the effect, of carrier phase error. It is shown that the MC-CDMA system is very sensitive to a carrier frequency offset, and the system performance rapidly degrades and strongly depends on the number of carriers. For a maximal load, the degradation caused by carrier phase jitter is independent of the number of the carriers.

This paper presents the probability model of the control-choice cryptographic logic. A necessery and sufficient condition is gained which satisfies that the mutual information is zero between control and output sequences of the control-choice cryptographic logic. By using the information leak between control and output sequences, a method of analysing the control-choice cryptographic logic is given.

A static routing algorithm which applied in traffic ungiiiooring environment is put, forward. This algorithm calculates and configures the path of LSP by taking the1 current network resource into account. When needing to configure more than one LSPs in a, priority, Genetic Algorithm(GA) is used to find the optimal or sub-optimal configuration for all this LSPs and make the maximal link bandwidth usage lower than the value defined by the administrator. This algorithm can rationally distribute the network resource. Besides, an improved Dijkstra algorithm is put forward to calculate the shortest path for LSP.

According to the management requirements of connection service in multitech-nology networks environment, a connection service management framework based on Telecom Operations Map(TOM) is proposed, which achieves the automation of the configuration and fault, managements of connection service. Next, the management interfaces between integrated multitechnology network management system and the dedicated network management systems in the proposed framework are discussed. Network management requirements, information model, and the implementation based on CORBA about the management interfaces are pro-posed, which cover the configuration and fault management functional areas.

The dispersion equation of the axial symmetric hybrid mode inside the helix with a partially filled plasma column is derived, its dispersion curves are calculated, and the conditions to form that mode are illustrated. The effect of confining magnetic field on the radial profile of axial electric field is studied. In the situation that the strength of confining magnetic field has been entrusted a practical low value, the calculated radial profile shows much stronger coupling of the filed to the electron beam. In the situation without confining magnetic field, a Trivelpiece-Gould mode propagating along the surface of the plasma column presents.

Wide swath SAR is a hotspot of the research of SAR (Synthetic Aperture Radar). In many fields of SAR application which require global and dynamic surveillance or frequent surveillance of a certain part of the earth surface, it is expected that the swath of SAR is much more wider. A new method of wide swath SAR based on multiple modulation mode is presented in this paper. The selection of modulation mode, the solution to the problem of blind zone, and the data processing method of this system are discussed.

A new symmetrical formulation without O(1/R²) singularity is presented to cal-culate the impedance matrix elements as long as the rooftop basis functions have the character of local continuity. With this new formulation, one needs only take care about the O(1/R) singularity that is very easy to resolve. Several different types of rooftop basis functions are analyzed and compared, on which our formulation is applied. Numerical results are presented which have proved the validity, the efnciency and the universality of the new formulation.

With same circular radiating antenna, same center frequency pulse, the analytical and numerical results of axial energy of several typical waveform pulse are offered by electromagnetic theory. The results indicate: in the slow decaying area, the axial energy decaying property of isosceles trapezium electromagnetic wave pulse is better than that of sinusoidal electromagnetic waveform while its increasing time is shorter than one twentieth of the waveform width.