In modern multimedia teaching system, the students just obtain image information through the big screen's slider, but due to the factors such as distance, they can't get the data clearly and store the content of the teacher's lecture automatically, in order to meet this demand, the embedded processor, image acquisition module and WiFi have been combined, the JPEG (Joint Photographic Experts Group) hardware encoding and PP (Post Processor) supporting images caling and format conversion technology have been introduced, the embedded wireless image transmission and processing system based on JPEG hardware encoding
was designed. For the students, the image from teacher can be decoded, displayed and stored automatically improving the learning efficiency and realizing the image information interaction. Finally, in the Linux environment, the procedures of Ad Hoc and self communication were programmed to test the program, the experimental data verified the feasibility and completeness of the system architecture, network protocols, and the core algorithm. The conn
ecting speed of the terminals was realtime and efficient, the speed of image a
cquisition, compression, display, and wireless transmission can reach the intell
igent standards.

In order to determine the location of the noise and the type trouble of the vehicle, a method of orientation identification of vehicle accelerating noise sources is proposed, based on acoustic vector sensor. The combustion noise of the vehicle acceleration noise could boil down to a Chirp signal within a certain speed range. Using the acoustic vector sensor combined with traditional omni directional pressure sensor and dipole directional particle velocity sensor simultaneously and colo
cately measures orthogonal components of particle velocity and pressure at single point in space, and with MUSIC (Multiple Signal Classification) can selectively extract the interested components of the noise sources and present the exact location information. Simulation results verify the performance of the method.

With regard to the issue of energy consumption while reallocating resources for service image placement system in MCC (Mobile Cloud Computing), it adds a wireless channel condition Monitoring module on the basis of the existing service image placement system, then proposes an improved scheme for reallocating resources. According to the users moving situation, the system doesn't reallocate resources when users quality of service is in the accepted scope, thus saving energy for system. It determines the accepted scope according to the simulation result. Compared with the newest simulation result, the improved algorithm can save energy consumption of system by 30%.

In order to meet the demands for improving IP phone services problems, we proposed a design for  embedded wireless audio real-time interactive and recording terminal with good communication quality. ARM Linux embedded system is used as the communication platform. Streaming media protocol is used to implement the digital audio real-time communication under WiFi network surroundings. Speex codec is used to implement the audio coding and decoding. And in order to improve the communication quality, the preprocessor and echo cancelation functions of Speex are utilized. The communication distance is less than 20 m, the system have good real-time feature, and the speech is fluent with good audio quality. The most outstanding advantage is that the echo is canceled and the noise is deeply depressed in communications.

In order to process huge amount of data generated in the social network with efficiency and scalability, we designed the distributed TF-IDF (Term Frequency-Inverse Document Frequency) algorithm suitable for MapReduce, and implemented this algorithm on Hadoop. This algorithm extracts key words in user's weibo, in this way user's interest could be found, and the corresponding service could be recommended to the user. In order to verify the validity and scalability of the distributed TFIDF algorithm, the results of the distributed TF-IDF algorithm and TextRank algorithm was compared. The experimental results show that key words extracted by the distributed TF-IDF algorithm could represent characteristics of the user more accurately. By Contrasting the response time, it could be seen that the distributed TF
-IDF algorithm has a good scalability.

MUSIC (Multiple Signal Classification) algorithm and ESPRIT (Estimated Signal Parameters via Rotational Invariance Technique) algorithm can not effectively achieve DOA (Direction of Anival) estimation of the coherent sources. Based on MMUSIC (Modified MUSIC), introducing the transformation matrix and considering the autocorrelation and cross-autocorrelation of the array output and the corresponding transformed matrix, and combined with TLS-ESPRIT (Total Least-Squares ESPRIT) algorithm, an IM-ESPRIT (Improved ESPRIT) algorithm that is suitable for the two cases of the non-coherent and coherent sources has been proposed. Compared with the conventional CC-ESPRIT algorithm (Cross ESPRIT algorithm), when the angle of coherent source interval is 3° and signal-to-noise ratio of 0, DOA estimation is realized. It is shown by the numerical simulation that IM-ESPRIT algorithm has a better estimation accuracy under the condition of the small angle interval of the coherent sources and lower SNR.

In order to reduce the loss of fluorinated polyimide AWG (Arrayed Waveguide Grating) wavelength division multiplexer, guide the AWG experimental preparation, which can meet the demand of low loss communication system, a detailed analysis of influence of structural parameters on fluorinated polyimide AWG multiplexer was done. Array aperture is an important design parameter of the phased array waveguide grating multi/demultiplexer. The transmission characteristics of a fluorizated polyimide AWG multiplexer are analyzed. The diffraction efficiency of the device in the I/O (Input and Output ) slab waveguides is calculated. The effects of some structural parameters including waveguide aperture on the transmission characteristics are investigated and discussed. The AWG structure parameters were optimized and the diffraction efficiency is about 100%.

Because singlelayer curved compound eye with large numbers of ommatidium, are difficult to process and small FOV(Field Of View), a multilayer curved compound eye is proposed. When the angle between optical axis of ommatidia is 66.8°, and the FOV of ommatidium is 74.5°, the FOV of the multilayer curved compound eye can reach 180° by seven ommatidia without the blind area of imaging. Then optimal design works on ommatidia were conducted, which make the imaging quality meet the application requirements. Imaging experiments were carried out, verifying that the multilayer curved compound eye can achieve a 180°
 imaging without blind area.

In order to meet demands of highresolution and high-refresh-rate on LCD (Liquid Crystal Display) screen in a taxi sharecar system, a design scheme of a
high-speed LCD controller integrated in FPGA  （Field-Programmable Gate Array）was proposed. The basic structure and function of the controller was presented. We put forward the design methods to improve working speed and working efficiency of LCD controller, for example, adopting of self-defined commands of graphic processing, using high-speed SDRAM (Synchronous Dynamic Random Access Memory) as memory, taking full advantage of spare time, DMA (Direct Memory Access) operation, more buffer frame synchronized and switched by hardware. Experiment result shows that the LCD controller could refresh the picture cyclically, the output picture was not distorted, the refresh frequency could reach 60 frames per second, the theoretical pixel fill rate could reach 88 million pixels per second. The flexibility of the system was greatly improved. The performance of the controller in share car system was good.

In order to achieve the goal of saving electricity, the intelligent electric meter suitable for colleges is designed, based on the MSP430 MCU (Micro Control Unit). Current is sampled by electricity solenoid. The IC card identification, keyboard and display the human computer interaction function make it not only a general intelligent electric meter, but also a meter with many users. The charging of period of time are suitable for the measures of restriction of electricity at college, which has a positive application prospect.

According to the diversity of launch conditions a neural network PID (Proportion Integration Differentiation) controller is designed through the clutch speed tracking control. Keeping the engine running, the controller makes the launch process of the AMT (Automatic Manual Transmission) vehicle smooth and fast. In order to verify the neural network PID controller's effectiveness, a powertrain model of a medium-sized truck about AMT vehicles is established. The model includes the engine, the clutch, the transmission, the main reducing gear and wheels. And simulations are carried out in various working conditions according to the actual situation. Comparing the performance to the traditional PID controller's and analyzing the results of simulations, the neural network PID controller provides the good performance during AMT vehicles launch progress and has good adaptability in various working conditions.

In order to accurately identify the student, monitor the visitors flow rate of classroom, prevent students from cheating, we design an intelligent classroom attendance system based on the Internet of things technology, RFID (Radio Frequency Identification technology), photoelectric detection technology, wireless transmission technology. The system collects information through the RFID module and two alignment type infrared sensors, the data is sent through a serial port to embedded microcontroller, which will show its data processing results and cheating buzzer hints. At last the master controller sends data through the wireless devices to PC system which carries out information statistics and storage. Stream of 51 people is used for system testing, and people who come into the classroom include 3 people who hold a card (card of test people 2), 1 person that hold 2 cards (cards of test people 6,7), the other 47 people who hold their own cards. Test results showed that the test people 2,6 and 7 were appeared to cheat, and the total number that comes into the classroom is fifty. Through the test, the system can identify student accurately, monitor the visitors flow rate of the classroom, prevent cheating especially for a person who holds multi-card and many people who hold a card. It is practical and universal.

To enhance test achievement evaluation in learning and teaching, we have conducted multiple linear regression analysis in multiple linear statistical analysis. A regression model of test achievement evaluation was established after the index was quantitated. The relationship between national cet4 test scores and ordinary test of students was analysed. A mathematical model of multiple linear regression is established based on the performance evaluation, and the regression coefficient values are obtained. We checked the regression model and the coefficients. The results show that the model is suited for the case. The model can evaluate the given data, and can predict the students band 4 examination results according to student's final exam reasonably when some conditions are known.

In order to improve the life quality of older persons and to improve physical function of the impaired health status, we propose a new game based on somatosensory. Game system is designed by the CMOS （Complementory Metal Oxide Semiconductor） sensor, color detection and motion control components. After microcontroller configures OV9650 camera, FPGA (Field Programmable Gate Array) module reads the acquired image data of the users hand. We sample image into Matlab simulation in order to determine color identification range. The air moving hand controlling skateboard in the bottom of the screen, the going red ball rebound collision the bricks located at the top in the game. Experimental results show that in suitable indoor light, the game runs smooth, and have a good human computer interaction.

In order to overcome the risk of misrecognition caused by a single guide line, an intelligent vehicle with the ability to automatically recognize and track path is designed. Freescale MC9S12XS128 and COMS (Complementary Metal Oxide Semiconductor) camera are chosen respectively as core controller and information collection device. Path information is indentified through the process of binary image, image denoising and edge detection, and the runway shape is identified, which are served as the control input of steering gear and motor to achieve a stable and fast running process of intelligent vehicle. Combining driving state and the road information, and using the PID (Proportional Integral Differential) control strategies, the optimization of various typical track processing is realized, the ability of steering adjustment and a quick response of acceleration signal are guaranteed. It can run on a runway with white ground and black line on both sides, overcoming the risk of misrecognition caused by a single guide line.

Tool radius is an important factor affecting machining accuracy. Studying on tool and radius compensation based on milling mache, we put forward a radius compensation algorithm by judging the adapter type of two adjacent curves. The detail and the programming implement of the algorithm are proposed. The experiments result show that the algorithm is suitable for all the radius compensation type, for example adjacent straight line and straight line, straight line and arc, arc and straight line, arc and arc etc,  and the algorithm has small measurement error and high efficiency.

The principalagent relationship can be found between coreenterprise and its supplier. When the core enterprise owns more than one supplier running the same business, the influence among suppliers constrains the efficiency of the supply chain, the same problem can also be generated by asymmetric information, moral hazard and factors like those. Therefore, simply treating suppliers as a whole definitely, the most effective incentive mechanism can not be designed. Under the frame of rank-order tournaments, with the factors of interrelationship among suppliers taken into consideration, a model with one principal and three agents is analyzed. Because suppliers in different ranking have different influence over others, the weights of their interrelationship are also considered. Through the research, the rationality and effectiveness of incenting suppliers via classification and gradation are discussed.

We propose a method of camera calibration based on genetic algorithms to improve the accuracy of camera calibration. The method extracts corners of chess board images and calculates the coordinates of the reference corners, then estimates intrinsic and external parameters of camera. Genetic algorithm is used to optimize the parameters extracted by the proposed method. Through evolution, selection and variation among generations, optimization is achieved by convergence. The result shows that the average projection deviation with the method is 0.255 9 pixel, and the standard deviation 0.139 3 pixel.

To solve the problem of security vulnerabilities and information congestion in course selection management system, we used UML (Unified Modeling Language) modeling approach and combined with the main process of the online courses and the actual needs of users of the system to analyze and design course selection system by Rational Rose 2003. And we built the systems use case diagram, class diagram, activity diagram, sequence diagram, component diagram and deployment diagram. UML modeling application shows that this method is easy to understand, to change and to upgrade. And this method has achieved a good development effect, overcome the low scalability of the traditional method.