Because Westwood algorithm can not distinguish wireless random packet loss and network congestion packet loss, we propose an improved algorithm Westwood-c, add the packet loss distinguish mechanism and improve the additive increase mechanism on the basis of the original algorithm, The paper sets congestion window Pcwnd by comparing the backlog beta message number N and setting value β of threshold making the network throughput maintain a high level, improving the network resource utilization.In NS-2 (Network Simulatorversion-2) simulation environment, comparing with the throughput of Westwood-c algorithm and Westwood algorithm under different scenarios. The results show that the Westwood-c algorithm throughput has increased significantly.

The optimal waveform design based on SINR(Signal to Interference plus Noise Ratio) needs to search for the Lagrange multiplier which leads to a large amount of computation. To solve the problem, a method of maximum energy distribution based on SINR is presented. It distributes the discrete energy reasonably to maximize SINR of the echo. The optimal transmitted waveform of cognitive radar can be obtained with this method.The simulation results show that the achieved optimal cognitive radar waveform based on the SINR maximum energy distribution algorithm is close to the waveform based on the Lagrange multiplier algorithm and this method does not need to search for the multiplier. So the amount of computation can be reduced, the computation rate can be improved.

In order to solve the maximum likelihood DOA(Direction of Arrival) and to estimate shortcoming of large calculated amount in multidimensional nonlinear search, this work combines the cuckoo search algorithm and the maximum likelihood algorithm, using the advantage of cuckoo search algorithm which has a fast searching speed, to optimize multidimensional nonlinear spectral function of maximum likelihood estimation of DOA. This work improves the position iteration algorithm speeding up the convergence, and preserves the main idea of Cuckoo search algorithm at the same time. The simulation results show that the modified cuckoo search algorithm has better convergence in DOA estimation, and the estimate performance is better.

Signals reliability of SNMR(Surface Nuclear Magnetic Resonance) is seriously influenced spiky noise caused by operating engineering machinery, in order to solve the problem an algorithm based on the median of standard deviation is developed and operated on artificial SNMR signals. This technique is designed to conduct threshold comparison to recognize and delete those spiky noise of strong randomness and high amplitude. By operating this method on artificial and field SNMR signals, it can be found that many spiky noise with amplitude of 10 times larger than real SNMR signal are removed. The decay shape of real SNMR signal are well restored.

Effectively detecting the band frequency and power of airspace interference and cluster is the prerequisite for improving the accuracy and reliability of EES(Electromagnetic Environmental Sensory) and MIMO(Multiple-Input Multiple-Output) radar. This paper presents a new electromagnetic environment percep
tion algorithm based on TLS(Total Least Squares) and SVD(Singular Value Decomposition). TLS-SVD algorithm uses cross\|spectrum AR(Auto Regressive) model estimation parameters and considers the disturbance estimation error of cross-correlation function matrix, then takes into account the TLS perturbation method for vector error estimate on the equation side so the cross-correlation function can be realized. The simulation results show that the proposed algorithm has a good performance of spectral estimation. It can sense the clutter band of the environment accurately, and provide favorable guarantee for sending and receiving waveform of the EES-MIMO radar using the remaining “clean” band.

In order to obtain the double motor fourwheel driving electric vehicles maximum driving force in different road surface, this paper proposes a traction control strategy based on LuGre friction dynamic model and a SMO(Sliding Mode Observer). Using LuGre dynamic model's parameter θ that react pavement condition to realize the pavement conditions online identification. By designing the SMO initiatively estimate the road condition parameters θ with small tire slip ratio while the electric vehicles in the vertical driving conditions. With the estimation optimal slip ratio, we can achieve the maximum traction provided by the road. We design dynamic saturated nonlinear control strategy to limit the maximum output torque of motor and to ensure the electric vehiclesanti-skidding under the different pavement conditions. The simulation results show that the optimal slip ratio estimation method based on sliding mode observer can provide maximum traction for electric vehicles under the different road conditions. The strategy is important in active safety of the electric vehicles driving.

INS/BD constitutes an integrated navigation system with high precision and complementary advantages, but the usual Kalman filtering can not obtain a satisfying precision. So the robust H∞ filtering is used for the INS/BD integrated navigation system, overcoming the precision impact caused by the model uncertainty, and converting the problem of filter designing into a convex optimization problem subject to LMI（Linear Matrix Inequality) constraints. Simulation experiments show this algorithm can be used for integrated navigation system effectively.

In order to increase the transmission rate of visible light communication system, using OFDM(Orthogonal FrequencyDivision Multiplexing) modulation techniques, a transmitter of visible light communication system is designed based on DSP processor. The hardware system includes DSP electrical platform, D/A conversion circuit and Bias-tee/LED driver circuit. The software system includes 16PSK modulation scheme to encode OFDM sub-channels, to enhance antinoise capability of the system enhanced, so that it can not be affected when the channel characteristics changed. Experimental results show that OFDM transmits data frames may be implemented by the transmitter of VLC (Visible Light Communication) system based on OFDM. The effective data transfer rate of the sender is 18 kbit/s when the modulation method is 16PSK, IFFT(Inverse Fast Fourier Transform) is calculated as 16order radix2 fast Fourier transform algorithm. The load resistance is greater, the waveform distortion of OFDM output signal by the LED(Light Emitting Diode) driver circuit is less. Since the OFDM technology has been transplanted into a visible light communication system developed, the data transfer rate of the system is improved, effectively suppressing ISI(Inter-Symbol Interference) and ICI(Inter-Carrier Interference) caused by the channel fading and delay.

In order to overcome the disadvantage of traditional ECG(Electro Cardio Gram) monitoring system, which circuit structure is complex, the volume and weight is big, and is not easy to carry. A portable wireless ECG monitoring system was designed, which used the analog front-end SOC(System On Chip) chip ADS1298 as the core chip of ECG measurement, combined with high speed and low power consumption MCU STC12LE5A60S2, and a wireless transmission module. The ECG data was transmitted to the PC through wireless communication, in which the digital ECG signals was processed and displayed via an application software of LabVIEW. Results show that this system has some advantages such as simple circuit structure, small volume, strong anti-jamming capability, portable, high precision, etc. This system can play an important role in detecting and preventing pre-hospital cardiovascular disease, and reducing the incidences of sudden cardiac death.

In order to increase the accuracy of gradient correction of magnetic measurement and to improve the magnetometer's function, the mathematical feature for the IGRF（International Geomagnetic Reference Field） is described, the complete formula suitable for programming and a program based on Visual C++
is also given. The program can calculate the geomagnetic elements of any position in the world based on the IGRF11 model. Compared with the standard results, the total intensitys mean error of the program is 1.2 nT, which can meet the requirements of the gradient correction. The study also provides a theoretical basis for the development of automatic resonance and PC software.

In order to provide important data protection for flight safety, flight maintenance and flight training, a data recorder is designed and its embedded software is investigated. First, analyze the key factors which influence the real-time data recorder software and figure out the interrupt response delay requirements according to the records of frame length data and the interrupt frequency. Then measure the interrupt response delay under different addressing modes by use of software GPO（General Purpose Output） and oscilloscope. Finally optimize the addressing recording mode according to the relationship between delay and ping-pong buffer depth. A platform is built to test the data recorder software. Experimental results indicate that the optimized software has shorter interrupt response time interval than non-optimized one. The shortest interval can reach 56 microseconds. The recorder can record all the interrupt source data with no missing data. It can satisfy the system requirements of real time and rapid speed.

To classify the defects, a defect hierarchical prediction model is proposed. From analyzing the defects attributes, a defect attribute prediction method is built. By using this method we can make the similar attributes defects together. From the analysis of the relationships among the similar attributes defect clusters, the defect relationship prediction theory is proposed, which can group the defects with strong relationships. The results indicate that the system architecture defect hierarchical prediction is better than k-means especially on the cluster density, numbers and objectivity.

To save the cost of experiment, improve the work efficiency, we predict the Bcell epitope by computing method, and then use the results to direct the experiment further. We extract 10 epitope relevant amino acids propensities, and then classify antigen surface residues by using vector machine algorithm to predict the candidate epitopes. According to fifteen test cases, we validate the effectiveness of the method.

In order to meet the demands of safety monitoring on small private workplace, this paper designs and implements a remote automatic video monitoring system with real-time, convenience, operability, low-cost and high-reliability. The system starts from the functional theory of the monitoring system, combines with the development of the computer image processing technology, captures image through a CMOS(Complementary Metal-Oxide-Semiconductor Transistor) cam
era, implements face recognition on the basis of OpenCV, and generates the trigger condition according to the result of face recognition and detection of moving
 things in the videos. According to the trigger condition, the SIM900A module will send alarming message to the user specified phone via the GSM（Global System for Mobile Communications） network. As a typical application of the combination of computer image processing and GSM network, when the target breaks
 into the monitored area, the system can give an alarm to the users phone in a timely manner, with high reliability and practicability by effectively handling
 such interference as noise.

Bipartite graph of queries and concept is insufficient in concept grabbing and calculating weight of the query words, which cause the user interest model construction not reasonable. The improved bipartite graph of queries and concept will represent the users query intent more accurate by using tf×idf grasping concepts and calculating the weight of query query words through user's browsing time. A user interest modeling algorithm based on improved bipartite graph of queries and concept was proposed. And experiments verified that user interest model constructed by this algorithm is more reasonable.

In order to enhance investment information superiority of the venture capital firm. we collect joint venture investment data in China during year the 2000~2013, and empirically analyze the relationship between the network embeddedness of a venture capital firm and the formation of venture capital alliance at the initial stage of venture capital development. The results show that the leader faces adverse selection problem when choosing partners. In addition, the leader hardly relies on its industry influence to raise followers threshold of joining the alliance when opening up investment areas. Finally, the opportunity of the follower joining the alliance is independent of the scale of joint investment.

According to the needs of teaching and scientific research in the field of mobile application security, the experimental platform of mobile application security is developed and constructed. Android is selected as the experimental system, and the experimental unit is constructed through creating a virtual machine in the computer and combining with the intelligent device. And many experimental units are connected by the high-speed network to form the whole experimental platform. According to the characteristics of the mobile application security, the design strategies for the experiments are made. 24 specific experiments consisted of 8 modules are designed on the basis of these strategies, covering three levels of the mobile application security. Based on the application of specific experimental module, the application method of the platform is introduced. The construction and application of the experimental platform for mobile application security meets the needs of the teaching and scientific research, and achieves a good teaching effect.

By focusing on information processing features in the environment of distributed cluster processors currently used in aeronautic, vehicle and vessel electronics. The architecture of embedded software component formation and component\|based software platform is proposed. It improves the processing approaches of communication data between devices under environment of distributed cluster processors, relieves constraints of embedded device resources, difficulties in software integration and complexity from communication environment, which has already been implemented and verified through embedded real-time operating systems. This embedded component-based software platform enables quick software integration, effective monitoring and management of system resources.

Because fuzzy Cmeans clustering method is sensitive to initial cluster centers and easily trapped into local minima, we cant get precise classification result in network intrusion detection. To solve the problem, a network intrusion detection method based on GSO(Glowworm Swarm Optimization) algorithm is proposed. First, samples with label is used to get initial cluster center. Then, GSO is employed to optimize cluster center. Simulation result shows that the method is effective.

Comprehensive evaluation on construction of educational informatization in primary and secondary schools provides a basis for the development of primary and secondary schools themselves and the decision of educational management. This article uses set\|valued iteration to construct the evaluation index system to support system evaluation on the basis of the comprehensive analysis. First, construct the hierarchical model of evaluation index system which consisted of 6 primary indicators and 24 secondary indicators, and provide a brief description of the primary indicators. Then, set up weight calculation mathematical model about set-valued iteration, including the two aspects of basic mathematics model and iterative process with strategy. Finally, according to the solving process to calculate the weight of index system. The results show that the method using the set-valued iterative to calculate index weight algorithm is simple, but the results in weight are more objective and real when the number of the same index are more.