The issue of real-time coding and transmission of video over WLAN is discussed in this paper. First, a modified hybrid ARQ/FEC algorithm, based on the characteristic of video coding, which improves the performance of video transmission, is proposed. To prevent the temporal error propagation, an error resilient video algorithm, with select intact reference frames or intra coding according to the backward channel signaling, is proposed as well.

A new Global Rate Distortion(GRD) optimal macroblock coding mode decision scheme is introduced for the newly H.26L video compression standard. Based on the Internet packet loss model of Bernoulli and Gilbert, this R-D mode decision approach can result in higher error robust than classical method. Furthermore, our experimental results also demonstrate its superior adaptive error resilient ability and practicability.

Future wireless networks promise multimedia communication and computing services for both fixed and mobile users. One of the most important challenges of wireless multimedia service for mobile users is maintaining a guaranteed quality of service over end-to-end connections. The use of micro/pico-cellular architectures is one of the promising ways to improve the capacity of mobile multimedia networks with reuse of the scarce radio spectrum. However, the handoff rate is increased with decrease of the cell size. An efficient handoff supporting which can satisfy various QoS requirements and give high resource utilization is one of the critical issues for mobile multimedia networks. In this paper, a handoff scheme that called dynamic channel reservation based adaptive quality of service handoff algorithm is introduced. Performance comparisons are made to the performance of other handoff schemes.

Based on the property of OFDM signal, a new carrier frequency acquisition algorithm is proposed, which employs only one block of identical time-domain training sym-bols. The proposed algorithm has good precision, low computations and large acquisition range. It is intrinsically robust to multipath channel and timing error.

By analyzing the OFDM signal, a carrier frequency offset estimation algorithm for uplink of OFDM systems is proposed. The algorithm with high accuracy can be imple-mented easily, and the performance of the algorithm is almost not affected by the sub-carrier number per user, which makes the algorithm suit for the multi-user OFDM system. The simulation shows that the algorithm has a better performance than the algorithm based on cyclic prefix.

In this paper, a sub-carrier allocation scheme is proposed for the future mul-tiuser mobile multimedia communication systems, which, is based on Orthogonal Frequency Division Multiplexing(OFDM) using adaptive modulation. This scheme uses an algorithm based on a rapacious optimal algorithm. Simulation result shows that this allocation algo-rithm is close to the optimum at effectiveness and efficiency with much fewer computations.

In this paper, a carrier offset estimation algorithm for OFDM-TDMA system is presented by taking advantage of the inherent orthogonality among OFDM subchannels. The simulation results illustrate that the algorithm offers a higher accuracy in the frequency-selected fading channel and fits the carrier traking in time-variant environments.

This paper presents a combined symbol and sampling clock synchronization method for OFDM systems. Especially, the timing correction is obtained by interpolation or decimation over sampled signal. This method utilizes the differentially modulated subcarriers in transmitter to obtain the synchronization error signal. The error signal is smoothed by loop filter, then it is used to change the control register of interpolator and decimator: The proposed algorithm can track the frequency shift of sampling clock. The algorithm is simulated in both AWGN and multipath fading channel. Simulation results show that the proposed algorithm is very effective on recovering timing errors of OFDM systems.

In this paper, an efficient and fast fault restoration mechanism based on m : n architecture with bandwidth constraints is proposed while three evaluating parameters are selected and formulated. The comparison of restoration performance, which is based on different protection range such as local restoration, partial path restoration arid path restoration, is given. Also the proposed analytical algorithm can be applied to evaluate the restoration performance of MPLS networks when based on 1+1, 1:1 or 1 : n architecture.

In Wavelength Division Multiplexing (WDM) optical networks, lightpaths are high-capacity all-optical channels. A single fiber failure can lead to the failure.of all the lightpaths traversing the fiber, and results in significant loss of traffic. Therefore, effective survivability mechanism is needed to minimize the data loss. This paper proposes an Adaptive Sub-Path Protection (ASPP) algorithm for WDM mesh networks with traffic grooming capabilities. The idea of ASPP is to segment each working path into several equal length sub-paths, and to compute their protection paths respectively. Based on dynamic traffic with different load, the performance of ASPP has been compared with traditional protection approaches via simulations. The results show that ASPP can make the tradeoffs between the resource utilization and recovery times.

To obtain efficient survivability, this paper presents a novel restorable lightpath selection algorithm based on the automatically switched optical network frame. A numerical simulation is made in term of this routing selection algorithm. The result shows that the traffic has an obvious balancing trend; the block probability of connection requests has been reduced. At the same time, this algorithm has an excellent performance on bandwidth utility and sharing of the restoration resources. This paper also discusses the mesh network service lightpath and restoration lightpath setting up strategy. In the automatically switched optical network frame, a parallel restorable lightpath setup mechanism is proposed based on the Generalized Multi-Protocol Label Switching (GMPLS)principle.

This paper proves that the longer the key of Twofish is,the smaller the differ-ential probabilities of the key-dependent S-boxes are. A new method is proposed to produce the key-dependent S-boxes. Theory and simulation results show that the XOR differential probability and modular addition differential probability of the modified S-boxes are smaller than those of the original S-boxes.

As the erratic performance of Extended Kalman Filter(EKF) method in passive locaiton, a new filtering method called Modified coVariance EKF(MVEKF) is put forward. It is also compared with several filtering methods mostly used in passive location and tracking,’ such as EKF, MGEKF, IEKF. Computer simulation shows that this method is robust and has short convergence time. Because it is not required by this method that the measurement equation is linear in MGEKF, it can be used in other non-linear filtering applications.

Basing on studying the principle of active contour model and some solutions to it, a new partial optimal dynamic programming contour detection algorithm is presented and the results comparable to the dynamic programming algorithm are given. Experiment shows such new algorithm is much faster.

In this paper, a novel approach is presented for the multiuser channel identification. The approach makes use of the modulation-induced cyclostationarity and its capable of identifing individual channels of different users. By measns of the complex cepstrum of cyclic autocorrelaton, the blind estimation of the channelcan be achieved. The appraoch is rather simple, with considerable advantages. The identifiability condition and its proof are also concluded in the paper. Finally, the simulaton of identification algorithm demonstrates its convergence.

The Arithmetic Fourier Transform (AFT) is widely used because of its simple computational structure and little multiplications. But over-sampling is often needed for the implementation of AFT to meet the accuracy requirements in real applications and this is one of the main drawbacks of AFT and limits its application. In this paper, with the fact that linear interpolation implementation can gain very high accuracy, a new implementation is presented based on linear interpolation and over-sampling. This implementation can get accuracy close to that by over-sampling, thus the over-sampling problem of AFT is overcome.

In this paper, the differential operators are introduced firstly. Then based on the characteristics of Haar wavelet transforms and the differential operators, a transmissibility algorithm is proposed and applied to the singularity measuring of 2D signal. Secondly, a new algorithm called the Eno-haar (Essentially non-oscillatory-haar) wavelet transforms algorithm based on the differential operators is presented. And it is proved by experiments that this algorithm is effective and feasible to image compression.

An efficient iterative lagrange multiplier approach is proposed for the design two-channel low delay perfect reconstruction filter banks. This approach has high design efficiency and can be used to design both equal-length and unequal-length filter banks. A design example is presented and compared with other methods. It can be shown that two-channel perfect reconstruction linear phase filter banks with higher stopband attenuation can be obtained using the new method.

The traditional formula computing the synthetic aperture length of spotlight SAR and its localization are presented in this paper firstly. Then the new formula to com puting the synthetic aperture length of spotlight SAR is deducted on the basis of airborne and spaceborne geometry models. Especially in spaceborne model, the earth speed is considered. Finally, the data simulation results validate that it is necessary to use the newformula to compute the synthetic aperture length. This paper provides the reference for deagnmg synthetic aperture length of spotlight SAR.

In this paper, a method of detecting moving targets in SAR using the match FourieT transform is introduced. The signal model and the detecting method are discussed and simulations are performed. The results of the simulations and the practices of detecting moving targets in SAR data show that when the velocity of the radar platform is compen-sated and pulses are compressed, the moving targets are easily detected using match Fourier transform, and under certain conditions, the moving targets may be detected depending on the relative position of strong points in contiguous SAR image.

Combining the idea of the real-time input estimation algorithm (Feng Xinxi, et al, 1996), a novel maneuvering target model is presented based on the "current" statistics model. At the same times, an adaptive tracking algorithm is developed using the relationship between the variance. A large number of simulation results show that this new maneuvering target model can characterize all kinds of maneuvering motion of target and that the adaptive tracking algorithm has good performance. So it is worthy to be applied widely in practice.

The EM algorithm, as an iterative numerical tool for computing maximum likelihood (or MAP) parameter estimates for incomplete data problem, has been used in area of target tracking, particularly in passive tracking and scenario in which real-time processing is unnecessary. As an extension of .Johnston’s recent work, a recursive algorithm for tracking maneuvering targets in clutter, which based on AECM algorithm, is developed in this paper. In this algoritlim, model posterior probability and data association probability are computed via HMM filter respectively. Computer simulation indicates that performance of the algorithm is comparable with that of IMM-PDA, and the algorithm is valid.

Error model and its corresponding pre-distortion calibration method for trans-mitting array are presented in this paper. According to the characteristics of transmitting array, element-pattern-based and synthesized-beam-based methods are proposed to calculate mutual coupling matrix in transmitting mode. Simulation results verify the effectiveness of present methods. The second method shows better performances than the first one for the case where the transmitting noises exist in the channels.

This paper proposes a novel modulation mode named Pulse-cycle Skip Modu-lation(PSM) in switching converter and analyzes its principle and characteristics. The state equations, modulation factor, response speed and the efficiency of PSM are also described. Compared with (PWM) control mode, PSM is a better control method in switching con-verter.

In this paper a novel adaptive semi-blind beamforming algorithm is proposed for OFDM-MIMO systems. First the nonlinear neural networks combined with maximized mutual information rule are used for estimating the coarse weights of beamformer, then the beamforming algorithm adaptively adjusts the vector weights by means of the known pilot tones inserted into the frame structure, thus the channel equalization can be further im-proved. The proposed algorithm is compared to two other channel estimators under various multipath fading environments through simulations and experimental data. The results show that the semi-blind beamforming algorithm has much better overall performance in terms of BER, moreover has high aperture gains with very good directivity in the beamformer pattern.

In this paper, based on the characteristics of Turbo codes and Trellis Coded Modulation (TCM), an improved method Turbo Trellis Coded Modulation (TTCM) is pre-sented and analyzed over flat Rayleigh fading channels in image transmission with two typical mapping strategies. The selection of different mapping strategies of TTCM in Rayleigh fad-ing channels for image transmission is analyzed and investigated. Furthermore, taking 8PSK modulation of TTCM as the example, two different mapping strategies are compared. Com-puter simulation results and PSNR that can evaluate the quality of the transmitted image are given.

Segment dropping is a novel contention resolution proposed recently. Firstly, the drawback of tail-dropping contention resolution is analyzed in this paper. Then a mod-ified head-dropping policy is proposed. This policy drops the head of contending burst only if the overlapping region of the two bursts is less than the whole contending burst size, oth-erwise drops the whole contending burst. In order to have a better support of differentiated service and alleviate the disorder resulted from head-dropping, a new burst assembly scheme, priority-based mixed burst assembly, is proposed. Such a burst consists of several classes segments, each of which is composed of one or several same class packets and separate head information. The priority of segments is arranged in ascending order in a burst from head to tail. Simulation results show that the proposed scheme performs very well in terms of performance metrics such as the packet loss probability and differentiated services.

A dynamic secret sharing scheme based on discrete logarithms and factorization is proposed in this paper. It has the following advanced properties: (1) The dealer can renew system secrets without renewing the shadows of the participants; (2) When the system accepts a new participant or fires a participant, the shadows of other participants would not change; (3) When some participants’ shadows are revealed ,they can be renewed without any effect on the others; (4) It can detect the cheater; (5) Comunication is reduced and work efficiency is improved.