Idle spectrum sensing is the key that cognitive radio can be implemented. Energy detection is frequently applied as a idle spectrum sensing technology for easy principle and easy implement. However, the detection performance may have abrupt change in the bad external communication environment (SNR:Signal to Noise Ratio is low). Firstly this paper analyzed the performances of energy detection under different SNRs, and then denoised the signal before spectrum sensing improving the performances of detection under low SNR. The improved energy detection algorithm is applied to cooperate spectrum sensing to ensure certain reliability of each CR(cognitive radio) node detection performance. So that the detection performances of cooperate sensing are improved. The result of simulation indicates that comparing to traditional cooperate sensing algorithm, the ROC（Receive Operating Characteristic） curve has a marked improvement when cooperate the detection results after the process of wavelet denoise from each node.

 In order to find a way to improve the performance of image compression, the characteristic and effect of 2D-DCT（Discrete Cosine Transform）, twodimensional tensor product wavelet，and the latest geometrical wavelet are compared. The same image is decomposed with different geometrical wavelet and then reconstructed with the same number of significant coefficients, thus the sparse representation capability of wavelet can be compared, the result shows that Bandelet and DCT are more efficient when require high compression ratio, while wavelet is still the most efficient tool when require high image quality.

 Aiming at a lower detective probability of the existing radar signal detection method to phasedarray radar signal, this paper presents a detection method based on wavelet denoising and extremum sequence algorithm. The method wavelet threshold denoising algorithm to pretreat the received signal at first, which can ameliorate the SNR（Signal Noise Ratio） of signal. Then it uses extremum sequence algorithm to detect the radar signal accurately and estimate its parameters accurately, which has a higher detection precision in a higher SNR. Simulation results show that compared with the existing method, the proposed one has 21.8% improvement in detective probability to phasedarray radar signal in a low SNR. 

In order to find a gesture recognition method with fast and high recognition rate, a gesture recognition algorithm is presented based on an improved probabilistic neural network. The improved algorithm uses K-W method to filter the most representative features of sEMG（SurfaceMyoelectrogram Gestures） features, and utilizes particle swarm optimization method to optimize the transmission rate. In the experiment of identifying seven kinds of hand gestures, the average correct recognition rates of improved probabilistic neural network are more than 90%, while the traditional BP algorithm was only 85.% correct. Simulation results show that the improved neural network algorithm has much shorter training time and stronger classification ability.

Aimed at the problems of long propagation delay and high energy consumption in underwater acoustic sensor networks, an ARBDT（Adaptive Routing Protocol Based on Distance Threshold) is presented. It decreases routing overheads by delivering data packets hop by hop in single path. In addition, ARBDT can adaptively adjust distance threshold according to network density, which restrains unnecessary responses from redundent nodes to reduce the collision probability of control packets, and makes the sender select next forwarder rapdily. Compared with the SBR（SectorBased Routing) protocol, simulation tests show that the proposed protocol can achieve better performance on delay and energy consumption.

To identify the target for solving the problem of feature extraction, a pulse coupled neural network combined with shape image mixed feature extraction methods is presented. This method using pulse coupled neural network, airspace signals are converted to image signal characteristics, combining the object shape of the image is greyscale information, the unity and shape description, the experimental results prove that the method of object in a certain extent of deformation, translation, zoom not sensitive to the target recognition system is a kind of very good feature extraction method.

Aiming at the autonomic network management problem in composite functional network system, on the basis of MHBS(Meteorological and Hydrological Business System), indepth researching of device status prediction and DENng policy model and the combination of above two mechanisms was done. The NARBP（Nonlinear Auto RegressiveBack Propagation） prediction model was proposed and some experiments of the model were done on the basis of real measure data. The result indicated that the SSE（Sum Squared Error）, MSE（Mean Square Error） and MSPE（Mean Square Percent Error） of NARBP prediction model was less than existing linear AR(Auto Regressive) model about 4/5. Finally the management decision components was expanded based on DENng policy model for the adaptability of system status prediction on application level providing an instance of policydriven data communication management with multiple channels to show the feasibility of expansion DEN-ng policy model in autonomic network management.

Aiming at the problem in the application of the RBF（Radial Basis Function） neural network to the credit score, we propose a RBF neural network classification under the Linex loss function, and produce the test results on the Germany  credit score data sets among the UCI（University of California Irvine）. The results illuminate that the method proposed in this paper may effectively solve the problem in the application of the RBF neural network to the credit score.

In China, the demand for high yield spandex. To solve this problem, based on the polymerization reaction sequence, flow charts are analyzed and improved, and the method of using computer Internet hardware and improvement of polymerization methods to improve the production efficiency and achieve the best modification of production. Then, using method of DCS（Distributed Control System） control of production lines of integrated optimization is studied. Aiming at LiaoYuan have company production status, based on the research and development of modern integrated manufacturing system, integrated automation technology and the optimal process scheduling techniques of spandex fiber production process of comprehensive control system of computer, to reduce production cost and increase the efficiency.

 In order to quickly and accurately verify the EMU air conditioning system simulation results,the author presents a program of data acquisition and analysis system.By comparing measurement and calculation.This system help designers to evaluate thermal comfort inside, establish new passenger car airconditioning effect of test mode. System consists of wireless data acquisition subsystem, airconditioning performance analysis software and a database composed of airconditioning design. Wireless sensor networks based on ZigBee protocol, node hardware application JN5139 module design, the tree algorithm used Cskip network address allocation policy, the nonlinear compensation of temperature and humidity sensor; CFD（Computational Fluid Dynamics） calculations using the FLUENT software; with ADPI（Air Diffusion Performance Index） Evaluate thermal comfort. Smallscale simulation experiments show that the system is reliable, accurate realization of the temperature and humidity and other parameters of the realtime measurement.

We propose a novel algorithm for extracting rules from artificial neural network. After the network is trained and pruned successfully, the activation values at the hidden unit are clustered into discrete values. In the cluster phase, the cluster number of weights can be adjusted dynamically according to activation values of their corresponding hidden units. The experiment results have shown its feasibility and accuracy, and prove that this algorithm decreases the complexity of rules extraction and makes the number of the extracted rules small.

In order to overcome the problem in the assessment of the investment project, determining the index weight, subjectivity is proposed through the least-square method determining the index weight.It is the optimal method to have preference information and objective information entropy weights of output, with least-square method for tools, set up the index weight determination model. This method can be applied to the evaluation decision more accurate objective, realistic effectively. This method overcomes the evaluation process determines the index weight, while the subjective judgment to improve the accuracy and scientific and practical value.

For improving the generality Siemens numerical control technology,taking SINUMERIK840D NC system as the development platform, using the OEM（Original Equipment Manufacture） package provided by Siemens, and giving full play to the interface superiority of VB(Visual Basic) and VC++(Visual C++), the VB program and the DLL(Dynamic Link Library) file under VC++are organically integrated, then the algorithm design of cam profile under the given data is analyzed,the wheel center trajectory equation is discussed in depth and the mathematical model between the actual grinding wheel feed and the cam rotation angle is built, and eventually the OEM software satisfying the special camshaft grinder processing technology is developed.

To project high-dimensional data into low-dimensional space can be effectively utilized to visualize and explore properties of data. An approach of NLPCA (NonLinear Principal Component Analysis)and SOM(Self-Organizing Map) neural network is presented for clustering and visualization of gene expression data. The experiment results show that the performance of clustering gene expression data based on the SOM network is efficient.

 For the purpose of building a digital model linking the reservoir geology and reservoir development,the concept of digital reservoir based on GIS （Geographical Information System）is discussed.A GIS based digital reservoir is proposed to display gelogical study and carry out development dynamic analysis. A framework of digital reservoir based on 2D GIS is presented, and the feasibility of integrating reservoir modeling and simulation using DEM（Digital Elevation Model） and 3D block grid together with other models to built digital reservoir system is analyzed. The functionalities of 3D GIS based digital reservoir model are also discussed.

 On the basis of the basic concepts of active voice and passive voice of verbs, the characteristics of dynamic verbs in passive voice structure is described emphatically through practical examples from English Sci-tech papers.some correct forms of passive voice which are often easy to be mistaken are introduced, and some suggestions for using active voice and passive voice exactly are put forward in polishing written language of scitech papers.

How to make the swarm reach the desired speed with the intervention of soft control is studied to achieve speed control for the swarm flocking system. Soft control is introduced in the distributed system based on local information interaction to interfere the movement of group, so that speed control for swarm flocking can be realized without changing the local information interaction rule of the swarm. Algebraic graph theory and LaSalle invariance principle are used to analyse the stability of the system with soft control. Simulation experiments have been done to verify the designed control strategy and the results show that the proposed control strategy enables the swarm robots system to reach flocking behavior.