An optimization method was proposed to solve the vibration issue of suspension system in both vertical and longitude directions, which is commonly caused by torque ripple of wheel-drive motor in electric vehicle. By flexibly modeling of the twist beam and realizing the high-frequency bushing model in the ADAMS, the elastic component was accurately described, then, the high-frequency applied suspension model was built. Based on ISIGHT software, the multi-objective optimization was conducted by integrating multi-disciplinary software. The optimization results showed that the weighted area of longitudinal vibration transfer function was optimized by 17.23% with slightly reducing the weighted area of vertical vibration transfer function. Within the target spectrum, the optimizing effect of vibration transfer is significant.

A brake master cylinder dynamic model for brake pedal feeling is proposed, in which the return spring preload, system friction forces, valve port clearance and changes in the brake fluid bulk modulus are considered. Experiments on the characteristics of the brake master cylinder are carried out under the condition of different pushrod speed, and then the key parameters of model are identified. By analysis and comparison, it is found that the test results and the simulation results have a good consistency. Further, the key factors of the brake master cylinder for brake pedal feel are analyzed, and results show that the piston diameter, spring preload and stiffness, and the brake fluid gas content significantly influence the brake pedal feel and master cylinder performance.

Highway Intelligent Space-Vehicle Framework Theory (HIS-VFT) is proposed to research vehicle driving state when passing highway tunnel. First, the longitudinal and lateral dynamics models are built based on the road driving characteristics to study the longitudinal and lateral driving states. Then, model predictive control method is used to predict relative vehicle distance and lateral deviation. Combined with prediction control characteristic, the variable conditions are set. Finally, MATLAB/CarSim is used to simulate the process. The simulation results show that using HIS-VFT can achieve vehicle safety driving when the traffic accident happens on the road ahead.

The effects of drawbead corner radius and metal flow direction on drawbead restraining force are studied by experiment testing and regression analysis. Experimental data are acquired respectively by a specially designed experiment device, in which the variable is the corner radius or metal flow direction. An approximate linear relationship exists between the corner radius and drawbead restraining force, while the relationship between the metal flow direction and drawbead restraining force is nearly parabolic. An approximate analytical model of drawbead restraining force is established based on the influence law and regression analysis. Finally, by T test of regression coefficient, it is concluded that the metal flow direction has stronger impact on drawbead restraining force than that of the drawbead corner radius. Besides, the coupling effect of metal flow direction and corner radius is insignificant.

The stochastic nature of traffic flow in stochastic user equilibrium networks was investigated. Based on the random utility theory, a stochastic traffic assignment model that can reflect the probability distribution of traffic flow was constructed using Baye's theorem. Driver's route choice behavior and the uncertainty in road capacity were taken into account in the model as the sources of fluctuation in traffic flow. A Markov chain Monte Carlo algorithm was also developed in order to estimate the characteristics of traffic flows from the probability distribution. The proposed model was tested on a medium-sized network. Results show that the model can be used to evaluate the effects of route choice behavior and uncertainty in road capacity on the fluctuations in traffic flows.

According to the traffic characteristics, the psychological and physiological activities of the driver in decision making process are analyzed. The main factors affecting driving decisions, such as the driver's range of perception and experience etc, were identified. Then, the psychological activities of the driver were considered and the driver perception threshold and emergency response threshold were determined. The method of fuzzy clustering analysis was used to classify a sample and the degree of membership of the sample in different driving decision-making was gotten. It was judged that the driver should accelerate, decelerate, uniformly drive or emergency brake. Finally, the driver perception threshold and emergency response threshold were analyzed on different grade roads. The results show that the driver's range of perception was not just the visual distance, and the driver's psychology would change a lot under different driving speeds. The proposed method is easy and simple to use. It can provide a theoretical basis for traffic simulation and traffic safety.

Considering the slip between steel girder and concrete slab, experiments of fundamental dynamic characteristics and dynamic response of steel-concrete composite beam under moving load are carried out. Then comparative analyses between the experimental, theoretical and numerical results are conducted, showing good agreement. The results show that due to the flexible shear connectors, the relative slip occurs between the steel girder and concrete slab of the composite beams under external loads, which significantly influences the dynamic response of the composite beams.

Six common lightweight aggregates are chosen as the concrete materials to investigate the anti-vibration performance of the shear wall. Software ANSYS is employed to construct nonlinear finite element model of the lightweight aggregate concrete shear wall. Results show that the hysteresis curve obtained by the finite element model is in excellent agreement with the experimental hysteresis curve, which verifies that this method can effectively deal with practical engineering problems. On this basis, the main factors and their influences on the seismic performance of the lightweight aggregate concrete shear wall are analyzed by experiment. This research may provide the guidance to construction worker for better practical architectural design.

Two tests were carried out respectively in this research, one to investigate the variation of the compressive strength of concrete, and one to investigate the seismic performance under high temperature. Comparing the unfired specimens and the fired specimens, through the study of concrete failure modes, using SPSS statistical analysis software to process the changes of temperature and compressive strength, the relationship between temperature standing time and compressive strength was obtained. Through the analysis of the skeleton curve, stiffness, ductility and bearing capacity of fired specimens, the following conclusions can be drawn: after high temperature, the yield load, peak load and ultimate load of the specimens are significantly reduced; while the shear load increases with the axial compression ratio, however when the axial compression ratio proceeds exceeds certain limit, the shear load decreases. Furthermore, a shear capacity formula was proposed, which can be used to identify the extent of building damage after a fire hazard.

To promote the realization of direct-start mode without using starter, the influences of active and passive factors on the first-cycle ignition and engine rotation characteristics were investigated on a four-cylinder GDI engine. Under different passive factors, the control ranges of the active factors, which assure the reliable first-cycle ignition, and optimized control making engine to gain higher speak speed are analyzed. Experimental results indicate that the reasonable control of ignition timing and excess air coefficient can obtain higher combustion pressure and consistency for the first-cycle. Under different passive factors including coolant temperature, initial piston position and fuel pressure in oil rail, certain differences exist in the ranges of ignition timing and excess air coefficient that assure the reliable first-cycle ignition. Meanwhile, the excess air coefficient and ignition timing should be selected separately to get better rotation characteristics and promote the realization of direct-start mode for different passive factors.

I order to study the effect of preload on cylinder head thermal fatigue, a thermal fatigue test bench for cylinder head is established, which takes the preload into consideration. The comparative tests on one certain type of cylinder head with and without preload are carried out. The results show that, with and without preload, at the cylinder head thermal test points the relative error of the maximum thermal stress is higher than 20%, and an average relative error of higher than 10% under these two conditions. In summary, the results indicate that mechanical constraint is essential in the simulation test of cylinder head thermal fatigue.

The hardware circuit implementation scheme of TCU of the electric control system for a certain type heavy vehicle with hydrodynamic automotive transmission was designed. In this scheme various control procedures were cured, including shift control, lockup clutch control, filling and emptying oil of hydrodynamic retarter control, shift quality control and so on. The software of the electronic control system was developed by layered and sub-module method. Automatic shift strategy with two parameters was formulated and combined with the bench test, which finally realized the functional verification. The testing results show that the designed hardware and software of the electronic control system can realize the integrated control of the automatic transmission and improve the shift quality.

Traditional internal force coupling control for shaking table is complex and shows poor control performance because of the same control parameter numbers of the cylinders. In order to stabilize the internal force, a novel Internal Force Coupling (IFC) control strategy is proposed based on deformation displacement and force spaces. Nonlinear equations of electro-hydraulic servo system and single rigid body dynamic model of mechanical part of the hyper-redundant shaking table are established with hydraulic and parallel mechanism theory respectively. Under the analysis of internal force space, internal forces calculated by force synthesis matrix are compensated in closed loop and decomposed to the inputs of the servovalves through the redundant deformation matrix. Simulation results show that the IFC control strategy effectively reduces cylinder forces and internal coupling forces of the hyper-redundant multi-axis hydraulic shaking table.

The cylinder stress state of the balanced two-ring axial piston pump is analyzed, and a calculation formula of the overturning moment of the cylinder is deduced. Also the calculation formula to balance the axial hydraulic pressure of the cylinder and the port plate is obtained and program is given. When the numbers of the pistons inside and outside are both even or both odd, the dynamic curve of the overturning moment of the cylinder as a function of its turning angle is drawn using software Matlab. The results show that, when the numbers of pistons inside and outside are both even or both odd, the inside and outside overturning moments of the balanced two-ring axial piston pump can partially or even completely counteract each other. So the overturning moment of the cylinder will significantly reduce and its fluctuation period will remain the same. When the numbers of pistons inside and outside are both odd, the overturning moment will be smaller than that the numbers of pistons inside and outside are both even.

A new type mechanical structure for intelligent lower limb sit-to-stand rehabilitation training system is designed. Mechanical analysis of the structure is conducted using finite element method. The sizes of the structure are optimized using software Hyperworks to improve the design. Results show that the presented mechanical structure not only meets the design requirements of the system, but also is simple for manufacture and assemble, and convenient to transport with low cost and lightweight.

The uncontrollable factors and controllable factors of the profile of three-dimensional flexible stretch bending forming are analyzed. Based on the analysis, a comprehensive quality evaluation method of three-dimensional stretch bending and a robust optimization method are proposed. The uncontrollable factors and controllable factors are analyzed based on orthogonal experiments by finite element simulations, and the factors greatly influencing the forming quality are identified. The comprehensive quality evaluation method is established using the loss of comprehensive quality as the optimization objective. The objective includes reducing the springback and controlling the change rate of profile thickness. The response surface model is adopted to carry out the fitting function. Genetic algorithm is applied for robust optimization, and the optimization results are validated. The results show that the robust optimization method can effectively reduce the comprehensive quality loss while the constraints are satisfied.

The mechanism of mechanical properties of Self-piercing Riveted (SPR) joints of aluminum alloy 5052 sheet with three rivets was investigated by experiments. The static and fatigue experimental data of m-o SPR joint (m-o joint) and m-I SPR joint (m-I joint) were obtained. The effectiveness of the experimental data was tested by normal distribution and two-parameter Weibull distribution. The best-fitting straight line of fatigue lift and two-parameter power function equations of fatigue load-fatigue life curves were acquired by least square method. Results show that the static fatigue failure modes of the SPR joint are the pull-out and fracture of the lower sheet respectively. The static strength, capacity of deformation and energy absorption, fatigue life and their stability of SPR joints with multiple rivets are all influenced by the distribution structure of the rivets. The m-I joints have favorable static property, but the m-o joints have excellent fatigue property overall. During the fatigue test, the sequences of the primary and secondary load bearing are generated by the distribution structure of rivets, and the fatigue fracture surface is the location of the primary load bearing sequence.

In order to improve the performance of wind turbine blades, a bionic design method was proposed based on the good aerodynamic characteristics of the swallow wing. The performance of the airfoil imitating the swallow wing was simulated and compared with the standard wind turbine airfoil using computational fluid dynamics technique. The results show that the bionic airfoil has higher lift coefficient and lift-drag ratio than the standard airfoil. The bionic blade was designed using the airfoil of the swallow. Numerical simulation and experiment for the wind turbine blade were carried out. Results show that the airfoil of the bionic blade has bigger pressure difference between the upper and lower surfaces, which can provide stronger lift. The efficiency of the bionic blade increases by 25% compared with that of the standard blade.

To enhance the capability of fleet and smooth navigation of quadruped robot in unknown and complex terrain, a control method of quadruped robot with straight legs in stance phase is proposed. First, in terms of the motion condition on the complex terrain, the kinematics and dynamics models of the quadruped robot with straight legs are established. Then, on the basis of the dynamic characteristics of the Spring Loaded Inverted Pendulum (SLIP), virtual elastic and damping elements are introduced to the multi-DoFs of the torso to realize the control of these DoFs during stance phase, especially, in the forward movement. Finally, the control method is verified in simulation of the robot navigating on flat or complex ground. The performances of two control strategies developed are discussed in simulation to improve the synthesized performance of quadruped under different conditions.

A kinematic model of the leg is established and the workspace of the foot is analyzed based on the walking leg designed for a hydraulically actuated hexapod robot. The performance testing system of the leg is established, the hardware of the controller of the testing system is designed, and the software for motion testing is developed. Experimental results show that the movement tracking error of single joint and the coordinate motion tracking error of joints are less than one degree. Dynamic characteristics of the joint meet that amplitude attenuation is less than -3 dB when the frequency is 0.5 Hz. It is verified that the performance of the waling leg meets the motion requirements of the hexapod robot.

To solve the problem of line-grasping control when an inspection robot crosses obstacles automatically, this paper presents an automatic line-grasping control method. First, a pose estimation method for transmission line is designed based on its textural features and integral projection algorithm. Then, based on the pose error of the transmission line, a human simulating intelligent controller is designed. Using the offset distance, declination, width of the line, and combing with the obliquity of the robot, the automatic line-grasping control is realized. Experiment results with different backgrounds and illuminations show that the poses of the transmission lines can be estimated availably by the proposed method, and the automatic line-grasping task can be achieved reliably and precisely.

The conventional sensorless algorithm for electric vehicle based on Extended Kalman Filter (EKF) can be influenced by the error of the rotor flux and shows considerable steady-state speed and position deviations. To take rotor flux as a variable in the EKF may increase the order of the operational matrix and the computation load of the real system. To solve this problem and still keep the advantage of the original algorithm, a novel Kalman control strategy using Nonlinear Two-stage EKF (NTSEKF) is proposed. The main idea is to decouple the Kalman filter into two parallel low order filters using coordinate transformation. In theory, the total number of multiplication and addition operations for EKF and NTSEKF is calculated respectively. Simulation and experimental results show that the novel strategy can reduce the number of arithmetic operations compared to EKF meanwhile still keep good identification performance.

OPRA_m is the classical model in the study of two-dimensional direction relationship. However, to solve practical problems, the multi-directional space relationship should be considered. In order to express the relative direction in three-dimensional space, the relative direction of the xoy plane is combined with the relative direction of the z-axis. A three-dimensional relative direction model, 3DOPRA_m, is proposed by extending the model OPRA_m for two-dimensional direction relationship, and two new relative angles are defined. Moreover, the formal definition of 3DOPRA_m multi-granularity, its reverse relationship and the conceptual neighborhood of 3DOPRA_m are given. Finally, the conceptual neighborhood based on the action of 3DOPRA_m model is applied to describe the car navigation problems. Results show that, compared with existing two-dimensional model, the proposed model is more expressive and can describe more relative directions in three-dimensional space, which has great potential for practical applications.

A synchronizing strategy of solving the diagnosability in distributed systems is proposed, which is used to reduce the costs of computation and communication. Under the condition of assurance diagnosability, the synchronization operation is decreased and the method is simplified. The trajectories, which are irrelevant with the diagnosability, are deleted, thus, the complexity synchronization operation is reduced due to the absence of these trajectories. The diagnosability of the distributed systems is discussed quantitatively; the synchronization operation and the cost of the synchronization are used as parameters to analyze how the structure of distributed system affects the diagnosability.

To achieve efficient Resource Description Framework (RDF) querying, an approach of translation of Simple Protocol and RDF Query Language (SPARQL) into Structured Query Language (SQL) is proposed. The translation rules of SPARQL graph patterns to SQL queries, and SPARQL queries to SQL queries are provided. An experimental system called SPARQLtoSQL is implemented. The experimental results are analyzed, which demonstrate the feasibility, correctness and effectiveness of the system.

A GPU-oriented performance portable parallel reduction algorithm for extreme value optimization and a global memory access optimization algorithm are presented to resolve the issues of deficient parallel granularity and global memory access conflict on different GPUs. Experimental results show that the presented optimization algorithms can obtain high performance on a variety of NVIDIA and AMD GPU devices, gaining a speedup of 3 to 6 times than existing methods.

A process calculus called L-π calculus was proposed to solve the issue of formal specification and verification for Wireless Sensor Network (WSN) routing protocol. The calculus can capture the characteristics of WSN, including the limit transmission range and broadcast communication. The defined syntax of the L-π calculus describes the network from the perspective of node. Structural congruence defined in L-π describes the non-sequential of node expression. Transitional rules of L-π calculus in terms of labeled transition systems are also defined to provide the function of dynamic deduction. After the definition of L-π calculus, a leader election protocol is described to demonstrate the description ability of the calculus. Finally, the calculus is illustrated by developing and analyzing formal models of the leader election protocol for WSN.

Designing of high-performance software architectures for streaming media applications on network processors is studied. The VPL (Virtual Pipeline) model is presented as an enhancement to the commonly used the HPL (Hybrid Pipeline) multi-core architecture model. VPL overcomes the load imbalance problem of HPL. Besides, VPL separates application logic design from resource mapping which improves flexibility of software development. VPL supports a variety of data flow mapping strategies. Two of such strategies are proposed. Performance of these strategies was analyzed theoretically and measured on real hardware to explore when to use each. As an example application of VPL, HiliMG, a streaming media gateway application, was designed and implemented on a Cavium OCTEON network processor. Experimental results on HiliMG further confirmed that VPL performs better than HPL in term of data throughput by at least 13%.

To solve the problem of file-sharing rate reduction and waste of upstream bandwidth caused by free-riding behavior in BitTorrent (BT) network, a new model of reputation-based trust management is proposed. This model iterated "Upload Trust" and "Evaluation Trust" to estimate the performance of each code, which combined the behaviors of upload and download together. It avoids the nodes cheating in one estimate standard and uses more parameters to identify dishonest nodes. So the nodes with free-riding behavior have an only a poor chance to get the resource. Through experiments over PeerSim, the results show that this model can effectively prevent the free-riding behavior and improve the file-sharing rate in BT network. It also has positive effect on dealing with Sybil attack and Collusion attack.

In this paper, the static and mobile sink models of wireless sensor networks are analyzed; then based on the application layer delay-tolerant level, a mobile-sink delay tolerance strategy is proposed. Using mobile agent node instead of static sink node, this strategy moves to the nearby resource node to send and receive data, thus alleviating “energy hole” problem. Using the application layer delay-tolerant level, the delay tolerance model under the environment of the mobile-sink nodes can reduce the energy consumption and greatly prolong the network lifetime.

The community structure of complex networks has attracted much attention. However, previous methods can not investigate the multi-scale property of the community. To address this problem, by analyzing conduct function C and average agglomerate probability, a local heuristic heteromorphosis strategy is proposed; then, spectral analysis of complex networks is combined with genetic algorithm; finally a multi-scale community detection algorithm HGASA (Heuristic genetic algorithm with spectral analysis) is proposed. Extensive tests on artificial networks and real world networks justify the superiority of the HGASA algorithm.

For the application of video monitor in long distance, a network video monitoring system is proposed based on dynamic strategy. The system consists of two parts: one part is the network video monitoring system server based on X264 encoder and Live555 streaming media server for Linux platform; and the other part is the network video monitoring system client based on VLC player for Windows platform. To maximize the encoding efficiency and the network utilization, two strategies are proposed: the dynamic selection algorithm strategy of intra prediction mode and the rate adaptive strategy based on network conditions. Experimental results indicate that this solution is feasible and has good monitoring effect.

Based on the haptic force dynamics feature, a method of Correlation-basis Combination Feature Optimization (CCFO) is proposed. First, the regional features of haptic force are extracted using mathematical morphology. Meanwhile, image features are extracted, such as the improved ratio of length to width, the contrast, the correlation, and the entropy. Then, the haptic force dynamics feature is optimized by criterion of correlation, and the optimal haptic force image feature is kept through analyzing the correlation coefficient. Finally, the haptic force feature set is obtained by linear summation. The experimental data are taken from the ITCSH Gait II database. The stability of the features is studied by calculating the intraclass correlation coefficient and coefficient of variation. The results show that the CCFO method can effectively reduce the number of features and hence improve the recognition rate in identity recognition.

In order to reduce the energy consumption and increase the response speed of the mobile stations in Long Term Evolution (LTE) communication technology, a novel power saving strategy is proposed in Discontinuous Reception (DRX) with sleep-delay mechanism. Considering that the system buffer is finite in practice, a multiple-vacation queuing model with a sleep-delay period and a waking-up procedure is built. Using the methods of embedded Markov chain and supplemental variable, the formulas of the system block ratio, the energy saving ratio and the latency of data frames are given. Combining the numerical experiments with analysis and simulation, the trade-off between different performance measures is investigated. Finally, the time length of the sleep-delay period and the buffer capacity in DRX are optimized. The results of this research have potential applications in the improvement and optimization of the energy saving strategy in wireless communication networks.

A novel arbitration technique for memory access packets is proposed, which is based on the round-trip latency prediction. First, the congestion information in the subsequent path of memory access packets is used to predict the waiting latencies of the memory access packets in the future, and then the round-trip latencies are calculated. Second, the predicted round-trip latencies are used to decide the arbitration for the memory access packets contending for the same link. The proposed technique is designed and implemented in the routers of mesh-based NoCs. Experimental results show that, under different network sizes and packet injection rates, compared with the classic Round-Robin arbitration mechanism, the proposed technique can greatly reduce the maximum latency, the average latency and the latency standard deviation of on-chip memory accesses, and it is proved to achieve better latency equalization of memory access.

The measurement of traditional image crosstalk degrees needs precision optical instruments, which has a high requirement on the environment light and also causes a complex data processing. This paper proposes a new binocular image crosstalk degree detection method based on the human eye observation combined with the brightness scale. First, the test pattern is generated and used for crosstalk degrees detection. Then, the equivalent visual brightness principles are used to select the best observation position. Finally, the selective matching between the crosstalk area and the brightness scale are used to estimate the crosstalk degrees. The experiment results show that the test results of the proposed method is the same as the test results of HIS color model. It can accurately describe the degrees of crosstalk of the grating type auto-stereoscopic display system. At the same time, the test results of the proposed method is not affected by the light of the experiment, which proves that the proposed method is strong robust for the testing environment.

In order to resolve the interference caused by the deployment of Femtocells in Macrocell, a co-tier interference and cross-tier interference coexistent model is proposed. By means of interference factor, the whole interference network is parted into co-tier-dominated, cross-tier-dominated and balanced areas. According to interference model and reasonable interference factor, the co-tier interference or cross-tier interference can be considered alone without impairing the performance of the Femtocell users. The probability distribution of the three areas is studied, which can provide theoretical guidance for practical project when laying out Femtocells in Macrocell. It can reduce the complicity of spectrum allocation, power control and access management.

Based on the principle of binary wavelet transform, a method of reconstruction of wavelet modulus maximum value for the noise in Magnetic Resonance Sound (MRS) signals is proposed. Combined with adaptive notch filter, the method is used to filter out noise. By analysis of the data of the initial amplitude, the average decay time and phase etc posterior the fitting of envelope signals, it is known that after several times of filtering processing of the noise signals, the obtained fitting error of the characteristic parameters is minimum, and the Signal to Noise Ratio (SNR) is obviously improved. Real data were measured and noise filtering was carried out using the proposed method in this paper. The obtained results are in good agreement with the drilling results, which verifies the effectiveness of the proposed method.

Cable-less storage seismographs are the direction of developing high-density and large-channel seismic exploration instrument in the future. The problem in remote monitoring data quality is the obstacle of the cable-less seismograph application. Based on the analysis of the requirements and characteristics of current communication technologies, a design scheme is proposed for a remote monitoring system of cable-less storage seismographs based on Beidou satellite communication technology. The hardware and software designs for the seismic data acquisition terminal subscriber unit are completed using the AT91RM9200 module and Linux embedded system. The main control center monitoring management software is designed by selecting Microsoft Visual C++6.0 as the development tool to realize the stable and remote monitoring for cable-less storage seismograph. This research provides an effective method for remotely monitoring the data quality of cable-less storage seismographs, and increasing the flexibility and reliability of cable-less storage seismographs during field operation.

An Amplify-and-Forward (AF) relay beam-forming and cooperation jamming scheme is proposed in the circumstances that the eavesdropper's channel state information cannot be obtained and the direct link is present. The beam-forming weights of relays are optimized using convex optimization theory, and the results of the optimization are global and unique. Simulation results show that the proposed scheme can effectively prevent eavesdropping, reduce information leakage, and consequently enhance secure transmission rate.

Considering the shortcoming of instability and low SNR in existing monaural voiced speech separation algorithms, a new voiced speech separation algorithm based on signal energy is proposed, which introduces the signal energy as another important voiced speech feature. This new algorithm is based on the improvement of the classical Hu-Wang algorithm, applying energy feature to the auditory reorganization part. It further improves the reorganization performance of the target speech auditory stream as well as reduces the influence of noise and improves its stability by applying the energy feature. The experiment results show that compared with Hu-Wang algorithm, this improved algorithm can improve the segmental SNR of the target speech segmentation and improve the separation performance obviously.

The existing traditional perception models can not characterize the real perception ability of wireless multimedia sensor node accurately. To overcome this shortcoming a novel three-dimensional directional perception model is proposed. On this basis, a target coverage algorithm is designed based on the optimization of tilt angle and deviation angle instead of adjusting the main perception directions to solve the all targets coverage problem. The title angle and deviation angle are optimized based on differential evolution algorithm in order to use fewer sensors to cover all targets. Simulation results show that the proposed method requires less number of sensors to cover all targets than random deployment.

Utilizing static sink to collect data from Wireless Sensor Networks (WSNs) may result in high traffic load in the vicinity of the static sink. Meanwhile, the nodes located near the static sink may be more requested than other nodes in network, so these nodes may consume more energy and trigger off route hole problem. To solve these questions, with the aid of mobile sink, a novel Mobile Data Collecting Algorithm (MDCA) is proposed. MDCA adopts the rule of packet intercept, so that the intermediate nodes can intercept data packets coming from distant nodes, which do not belong to propagation tree, and forward these data packets through its bypass leading to the mobile sink nearby. This rule can efficiently reduce the transmitted distance of data packet; hence improve the security of data packets. This strategy performs better in improving the success rate of packet transmission.

A machine vision based method is proposed to detect handheld phone use behavior. In this method, first, Active Shape Model (ASM) algorithm is used to segment the skin-color region of the driver face, and the region is tracked by F-B error algorithm when the head moves. Second, an adaptive skin-color range extraction algorithm is proposed to detect the skin-color pixels and determine the candidates of hand images. Third, Co-occurrence Matrix of Oriented Gradients (CMOG) is proposed, which combined with Histogram Oriented Gradient (HOG) feature is used to detect the real hand image from the candidates. Finally, three modes logic discriminating method is used to identify the driver state. Experiment results show that the proposed machine vision based method can detect the handheld phone use behavior effectively.

In order to handle the problem that current observation information is not used in the particle sampling stage of the Particle Filter (PF), a novel particle filter algorithm based on Second-Prediction (SP-PF) is proposed and implemented on the Graphics Processing Unit (GPU). First, sampling from the state transition function is performed to obtain the predicted particles, and by least square estimation, a new generator of sampling particles is constructed. Then, on the basis of the predicted particles, the current observation information is introduced into the secondary particle prediction. After double predictions, each new particle is an unbiased estimate of the current state. Finally, the current state is estimated by means of weighting these particles, and after the completion of state estimation, resampling is conducted. Experimental results demonstrate that the proposed algorithm improves the precision of estimation compared with Standard Particle Filter (SPF), Auxiliary Particle Filter (APF) and Unscented Particle Filter (UPF). Furthermore, the processing efficiency of SP-PF is greatly raised by using GPU.

For gamut mapping distortion type images, a new image assessment method was proposed based on traditional algorithms. First, each image feature extraction objective and extraction method were elaborated, and a final assessment formula, named GMSIM, was established based on these features. Then, applying three existing gamut mapping image sets and the TID2008 image library, the parameters of the GMSIM algorithm were fitted, and the prediction accuracy of the algorithm was analyzed. For gamut mapping image sets, hit rate was used as indication of the algorithm performance. Experiment results show that, compared with traditional SSIM method, the proposed GMSIM algorithm prediction accuracy of subjective perceptions for gamut mapping distortion images was improved by about 4.5%, although there is still much room for improvement. The prediction performance for TID2008 image library is also improved to some extent.

A bottom-up data driven algorithm is proposed to extract image salient region by utilizing visual saliency prior knowledge such as regional contrast and spatial location etc. In this algorithm, first, the foreground region in the image is extracted. Then, the regional contrast and spatial position feature functions are constructed. Finally, the saliency map is acquired by fusing these features. The model well establishes the relationship between the spatial space and regional contrast, and generates saliency maps with precise details. The results demonstrate that the proposed method consistently outperforms existing saliency detection algorithms, suppresses non-significant regional disturbance, and yields higher consistent saliency map, when evaluated using one of the large publicly available data sets MSRA-1000. Meanwhile, the extracted saliency map is applied to automatically segment saliency region and better segmentation results are obtained.

In order to extract eye feature accurately, a novel automatic detection localization method of iris and canthus is proposed. First, the face is recognized in a complicated background by using AdaBoost algorithm. Variance integral projection is used to extract eye area. Then, the deformation circular template and the improved optimization matching function are designed to locate the center of the iris and to calculate the iris radius. On this basis, according to the characteristics of the eye structure, both inner and outer canthus direction is calculated using linear template, which is relative to the center of the iris, and then Harris corner detection algorithm is used to set canthus best position along this direction. Finally, this method is validated by IMM face library, and the experimental results show that this method can locate the human iris and inner and outer canthus accurately.

Considering the problem of source number estimation in the presence of unknown spatially nonumiform noise and the limitation of the received data, a now method based on Bootstrap technique and characteristic subspace projection is proposed. The estimation projects covariance matrix estimate of array signal into signal eigen subspace and noise eigen subspace respectively. A sequential hypothesis test is formulated based on characteristic subspace projection. No assumption is made in the distribution of the data and the Bootstrap is used to estimate the distribution under the null of the proposed test statistics. Simulation results show that the proposed method has good performances when the signal power is equal or different and under the condition of space-nonuniform noise environment, especially in the low signal-to-noise ratio and small snapshots.

In order to achieve the selective focusing on different targets in electromagnetic plane wave, a decomposition method of the time reversal operator based on the time reversal invariance of wave equation is proposed. In this method, the transfer matrix and time reversal operator are obtained by emitting and receiving the signal for detecting targets with the antenna array. Then, the time reversal operator is used to perform eigenvalue decomposition. In eigenvalue decomposition, the number of nonzero main eigenvalues is the same as the number of targets, and the main eigenvectors corresponding with these nonzero main eigenvalues contain the direction information of the targets. According to these eigenvectors of targets to calculate the objective function of every detection point in the given target area can be obtained the objective function value distribution of the target area, and achieve the selective focusing on different targets. Finally, the simulations are performed and the results show that the selective focusing method with decomposition of the time reversal operator in electromagnetic plane wave is effective.