To resolve the question of controlling the in-vehicle MOST（Media Oriented System Transport） device node, design and realize a in-vehicle MOST control node. Comparing to the traditional connection methods, MOST bus has the advantages of anti-interference, low quality, high speed,low cost, and etc. With OS81050 and ATmega128, optical fibers are used, control node send message to MOST device nodes. All device nodes can work synchronously. The control node can monitor the status of device nodes, adjust MOST network when the status of a device node is changed.With the research on the principle of MOST network, it becomes foundation and theory to apply MOST technology to in-vehicle electronic technology, smart homes.

Dynamic spectrum allocation is a key technigue of a CR(Cognitive Radio) system. The existing spectrum allocation method focuses on the conflictfree allocation to maximize spectrum utilization, etc. There is no full account of the statistical information of dynamic spectrum, users of different business do not need different bandwidth and realtimenonrealtime needs. The dynamic spectrum allocation for CR based on statistical information and multiqueue has been studied. Compared the exclusive allocation methods with spectrum pool sharing methods, the simulation results show that in the case of centralized control, spectrum pool sharing ensures better fairness and spectrum switching provides higher throughput of the system.

To study on the superiority and flexibility of watermarking algorithms in multiwavelet domain, on base of detailed analysis of the decomposed coefficients of BOR(Balanced Optimal-Reconstruction) multiwavelet,  the “layer” concept of an image in MWD(Multi-Wavelet Domain) is proposed. For this charater, data can be hidden in different layer subimages. An adaptive blind watermarking algorithm is put forward to embed a binary image into the first layer subimage according to above trait. Experimental results show that compared with the algorithm in Haar domain, the given method has better imperceptibility and higher robustness against image processing attacks.

According to the problem of whether the information transmission capacity of communications following the Shannon formula limit, the information transmission capacity of signals is studied. The symbol is defined as the key contacting signals and information, and the symboltransmission rate of signals is defined as the importance index describing the information transmission capacity of signals, and maximal symboltransmission rate of signals of electronic communication systems is worked out. The results are the information transmission capacity of communications essentially lies on information carrier and information transmission mode,and on nowadays communication modes the information carries only by use of the space degree of freedom of signals and the information transmits only by use of the undulatory property of signals. The Shannon formula shows the information transmission capacity of signals. And finally, the information transmission capacity of high speed signals, microcosmic signals and mesoscopic signals are likely to break through the Shannon formula limit.

With the growing number of the in-vehicle electronic devices, the weight of vehicle increases, wiring becomes more difficult, EMI (Electronic Magnetism Interfere) affects the quality of transmission.The software and hardware design of MOST-based in-vehicle audio system are introduced. Audio devices can be controlled, stream data can be transmitted, and functions related to transmission can be explored. The experiment indicates that MOST bus has the advantages of anti-interference, low quality, high speed and low cost. So it becomes foundation and theory for applying MOST to in-vehicle multimedia, smart homes, or add other in-vehicle devices to MOST bus afterwards.

 To improve the precision and speed of fitting hyperbola, and to overcome the influence of random errors between the measurement of laser beam quality factor M ², the method for fitting hyperbola with NE(Normal Equations) is adopted. The fast numerical algorithm of solving the normal equations with triangular decomposition is presented. It is shown by experiments that the M ² factors obtained by the normal equations method is consistent with that solved with SVD(Singular Value Decomposition) method, but the first one is faster and more accurate, the operation speed can be improved by one order of magnitude.

 Allocation mechanisms of spreading-spectrum rate in wireless networks with low signal-to-noise ratio due to strong interference are investigated for the purpose of ensuring relatively reliable communications. Based on DS-TDMA(Direct Spreading-Time  Division Multiple Access), a novel ASMAC（Adaptive Spreading Multiple Access Control）protocol, is present on the basis of adaptive frequency-spreading techniques, whose major innovation lies in that each mobile terminal is configured to piggyback its current SNR(Signal-to-Noise Ratio) and its expected traffic rate with its access request to the BS(Base Station). Before the beginning of each frame, BS will consider the overall requirements from different mobile terminals and allocate spreading-spectrum rate among them according to specially designed policy. Simulation results show that this algorithm is capable of providing an average connection rate of 8.6% even under an environment with -12 dB SNR.

For better signal processing and feature extracting of nonstationary signal, timefrequency analysis is one of the top interests and more and more research has been put on this topic. For the limitations of the conventional methods of timerequency analysis, a novel method of timefrequency analysis is analyzed which is applicable to nonstabilization and nonlinear signal. Empirical mode decomposition algorithm which is used to extract features of signals through utilizing the empirical mode decomposition and hilbert transformation are proposed and its characteristics are analyzed. The analysis method and steps of empirical mode decomposition timefrequency characteristic is presented. To demonstrate the validity and unique virtue of empirical mode decomposition timefrequency analysis method, the transient signal features abstraction and signal tendency abstraction are studied by the method. Simulation results have shown the feasibility and validity of the method. Lastly, timefrequency analytic method based on empirical mode decomposition is used to extract of signals trend component and to demonstrate the efficiency and superiority of this new timefrequency analytic method.

In order to deal with the optimization of complex electromagnetic problems, beginning with the theory of particle swarm optimization, by analyzing the convergence of the algorithm and its limitations, the performance of PSO (Particle Swarm Optimization )was improved, parameters tracking was combined, and PSO-PTS (Particle Swarm Optimization Parameters Tracking Strategies)hybrid algorithm was proposed. The method can effectively reduce the search region of PSO algorithm, ensure the uniformity of solution, enhance the computing speed and the accuracy of solution. Basic theory, math model and predict step of the algorithm was introduced. E-shaped dualfrequency microstrip antenna by using PSO-PTS hybrid algorithm was simulated. The results show that the antenna designed by this metheod can sufficiently realize the demand of miniaturization. 

 For lacking of multifinger fingerprint segmentation algorithm to induct multifinger recognition engineering applications at present, it will increase the difficulties of finger recognition application. To satisfy the need of engineering applications, a new multi-finger fingerprint segmentation algorithm is proloosed.Distributions of X-projection, Y-projection and segmentation of projection of edge image are counted to get the coarse each finger position, and then the direction of main axis of each finger, X-projection, Y-projection and segmentation of projection of each region are computed to locate each fingerprint region. In experiment, 600 multifinger images, including different background light, hand gesture are tested by the proposed algorithm. Experimental results show that the algorithm can meet the need of engineering applications.   

In order to achieve distributed task allocation dynamically and effectively in multi-robot systems, multi-robot fire-disaster response was presented.The mathematical model of multi-robot cooperative fire-disaster response had been build, and marketbased multirobot task allocation for fire-disaster response was proposed. The distance factor and the fire behavior and time factors are taken into consideration, its bidding expressions accord with request of dynamic task allocation for fire-disaster response. Simulation experiment is performed in the selfdevelop platform, and the simulation results are analyzed. The experimental results show that this strategy can achieve multirobot dynamic task allocation for firedisaster response in different situation.

For efficiently eliminating morbid samples and improving the generalization ability of neural networks, we present an approach to eliminate morbid samples in forward neural networks based on the search thought and the Hamming distance, through developing the thought and introducing the distance. The approach can directly carry out searching and eliminating morbid samples, and do not consider prior knowledge, forms of samples, etc. Hence, its applicability is stronger. The results of numerical demonstration analysis show that the approach is scientific, effective and can effectively find out morbid samples of learning samples,it has obvious application value to solve problems of real world.

In order to avoid the shortcoming of non-dynamic base vectors and specific parameters in existing feature extraction based on linear transformation,after analyzing iris geometry features and recognition principles,an iris feature extraction using independent component analysis method is proposed. The algorithm eliminates the iris feature space redundancy furthest, and overcomes the flaw of non-dynamic feature base vectors in traditional linear transformation.The iris classification with BP neural network achieves lower dimension and effective feature expression. Experiments  in small sample space  using self-made iris database JLU-IRIS indicate the accuracy and validity of the algorithm through three groups of different recognitions rate 100%, 96.5%, 92.5%.

Because it is difficult to analyze the OPN(Object-oriented Petri Net) model of equipment repair support system which is enormous and complicated, the CPN(Colored Petri Net) tools simulation software in simulating and analyzing its OPN model was put forward. According to the thread of simulating and concrete operation method, it was successful to construct the CPN model of the example OPN model of equipment repair support system. By the process of running and the result of simulating that can conclude that there are no conflict and deadlock CPN model. Then, it concludes that the method of simulation analysis for the OPN model is rational and effective. 

In order to address some issues of the routing protocols in the Ad Hoc network,such as poor validity and large control overhead,we use the OPNET simulation tool for the performance of AODV(Ad Hoc On-Demand Distance Vector) routing protocol simulation. From the network transmission delay, network throughput, network load and other key parameters of the simulation results,we can see that AODV routing protocol performance is relatively stable.That is,on-demand routing protocols are more suitable for Ad Hoc networks than the active routing protocols. The results of this study have practical significance to the researchers in the Ad Hoc network.

 According to the actual circumstance and problems of low technological level and long product development cycle that exist in manufacturing enterprises, the rapid process preparation system structure which includes interface level, core level and database level was established. The rapid process preparation information modeling was established by using district sorting algorithm and fuzzy intelligent query algorithm, the process configuration sorting sequence could be made and the optimized process route could be chosen through these algorithms,the sorting sequence of characteristic information of parts could be done quickly.

P2P (Peer-to-Peer) traffic has been a major consumer of bandwidth. The identification of P2P applicationsand the monitoring of P2P traffic are the key points of concern for carriers. The traffic collection, traffic analysis and traffic control techniques and methods of P2P applications are studied. A test has been made in the MAN(Metropolitan Area Network) to identify and control P2P traffic by using a traffic flow monitoring and controlling equipment, which proves the validity of the P2P traffic analysis and management. 

In order to improve the efficiency and quality and reduce the cost of training, a networked real time platform-type flight simulator is designed and implemented. We adopt a completely datadriven technique to design an effective aerodynamic math model using stability coefficients. Furthermore, object oriented method is used to design and implement the integrated simulation of flight instruments, threedimensional scene, flight equation and console of faculty man. The multifunctional platformtype flight simulator is provided with favorable function of manmachine communication and composed of real joystick, accelerator stick, pedal and sixdegreeoffreedom head tracking system basedon light and implements computation of flight dynamic model, management of graphics, real time data transmission and setting of flight condition. In practice, it satisfys the needs brought forward and reduces the cost of training,sueh as human power,fuel and so on,improve the effciency and quality of training. 