Rigidflexible coupling model of twist beam suspension was built, and the correctness of the model was verified by the experimental Kinematics & Compliance data. By studying the influence of the position of the cross section, the orientation of the cross section, the orientation of the bushing on the suspension characteristics, the relationship between these factors and the suspension characteristics were found, provides a basis for the twist beam suspension design and tuning.

A complex car model was built for ride comfort analysis using ADAMS/Car and simulation of the model was performed on the random road and pulse road.A virtual simulation test rig was developed to test the car body natural frequencies.Comparison between test results and calculation results show the virtual simulation test rig is effective.The test rig provide a way for vehicle model body natural frequencies test. Using the methods of frequency domain analysis and time domain analysis, the full car transfer characteristic was researched. Based on these, the full car has been simplified to twofreedom model and vibration model was established using ADAMS/view. The factors of influencing vertical vibration have been researched, including suspension stiffness, damping , tire stiffness and so on. Influence of these factors on ride evaluation parameter was analysed.The analysis show that the variational rules of suspension working space, dynamic tire load along with the changing of suspension stiffness, damping.

The adjustable shock absorber of semiactive suspension was developed. On the basis of the project analysis, the structure of adjustable shock absorber was presented. And the samples were manufactured and the bench test was made to obtain velocity characteristic. The absorber samples and the data acquisition system were fixed in the target vehicle, and the road tests were made. The results indicate that the adjustable shock absorbers can retrain body pitch vibration during urgent braking maneuver effetively.

A vehicles active suspension system by using the phase plane division control and LQG control strategies was proposed. A halfcar fourdegreeoffreedom active suspension mathematical model was built and simulated in the environment of Matlab/Simulink. The simulation results were analyzed by the timedomain and frequencydomain methods. The performance indices, including the body acceleration，the body pitch angle，the suspension working space and the tire dynamic deflection, were compared among  the integrated control active suspension, the LQG active suspension and the conventional passive suspension. The simulation results indicat that the vehicle body pitch angle is reduced evidently and the ride comfort is ameliorated by using the integrated control, which has a small influence on the LQG control effect.

Two and fourwheel drive electric vehicle models were built based on the EHB system to realize regenerative braking. In accordance with “under the premise of meeting the regulatory requirements, braking energy recovery maximization” principle, the optimal braking force distribution, parallel system of power distribution and a fixed ratio of braking force distribution algorithm were carried out, to compare braking energy recovery of frontwheel drive and fourwheel drive electric vehicles,respectively. Simulation results show that frontwheel drive system for electric vehicles using the optimal power allocation algorithm recover more braking energy; fourwheel drive electric vehicle using a fixed ratio of braking force distribution algorithm has the highest energy recovery rate; fourwheel drive and frontwheel drive electric vehicle with maximum energy recovery braking force distribution strategy, energy recovery rate is almost the same, economy is not significant different. However, fourwheel drive electric vehicle braking force distribution method is a fixed ratio, which is closer to the ideal braking force distribution curve than the precursor, due to that it is more secure.

The impact factors of electric power steering system stability were studied from the perspective of motor current control methods and boost gain. To improve system stability margin, motor control principles and torque signal processing methods were proposed. Through simulation and real vehicle experiment verification, the stability of electric power steering system can be improved significantly by current control for highfrequency gain attenuation and torque BiQuad filtering method.

The accuracy of electric drive system model can affect selections of vehicle control strategy and control parameters for hybrid electric bus. In the development process of hybrid commercial vehicles, a simulation model was established based on AVL CRUISE,taking the advantages of drum dynamometer to compare and calibrate the differences between simulation data and experimental data of available vehicles,so the accuracy of simulation model was further improved,and lays a foundation for the simulation of hybrid commercial vehicles development.

The basic principle of thermoelectric theory and configuration of thermoelectric generator were introduced,and a thermoelectric generator model was established. The key influence factors on fuel efficiency of thermoelectric generator were discussed,and series hybrid electric vehicle with thermoelectric generator was simulated based on software PSAT. Simulation results show that fuel efficiency can be improve approxiamtely 5% under different driving conditions.

Aiming to the lack of existing steering model, a steering model used for vehicle dynamic simulation was developed. This steering model established a complete dynamic system consist of wheels, axles, steering tie rod and rack, so the left and right wheels had the ability to coordinate with each other. Stickslip friction model was used to calculate the friction on steering kingpin, so the vehicle model had the ability to resist the small interference from outside. After verified the accuracy of the steering model, the typical operating conditions of external small disturbance were simulated, the simulation results showed that new model had the ability to resist the external small perturbation, and the fidelity of vehicle model was improved.

An eighteen degrees of freedom multiaxle vehicle dynamic model was established in Matlab on the basis of antiblock system，and vehicle stability control strategy was proposed. Using direct yaw moment control algorithm, joint control of side slip angle and yaw rate based on the dynamic model were proposed with fuzzy control strategy,which side slip angle and yaw rate were limited within a certain range,and the turning braking stability of vehicle was simulated according to the designed control algorithm. The results show that the joint control of side slip angle and yaw rate achieves better results than indivadual control and it will be better to improve vehicle turning braking stability.

A four degrees of freedom multiaxle vehicle steering model was established. Analysed timedomain and frequency domain characteristics of multiaxle vehicle steering performance based on front wheel steering control strategy, zero lateral angle control strategy and yaw rate tracking optimal control strategy. The results of the analysis indicate that in the case of zero lateral angle control strategy and yaw rate tracking optimal control strategy, the vehicle yaw rate, side angle, rolling angle and pitching angle are smaller than that in the case of the front wheel steering control strategy, and the system response fast for yaw rate and lateral angle, indicating the flexibility to be able to turn better; and they response lag for rolling angle and pitching angle relative to the front wheel steering control strategy, indicating a better steering stability. The results of the analysis also shows that the steering performance of yaw rate tracking optimal control strategy is better than zero lateral angle control strategy.

For rollover detection and control study, a simplified nonlinear rollover prediction model commercial vehicles was proposed. In order to obtain valid model as close to the target real vehicle as possible, a set of key vehicle parameters were calibrated from experiment data collected from real vehicle field operation tests. Calibration of the vehicle prediction model simulation results compared with the experimental results, it is shown that the prediction model can be accurately predicted the rollover dangerous state.

Design requirements for solenoid valve and hydraulic pump motor drive circuit were analyzed, the solenoid valve and hydraulic pump motor drive circuit was designed, and solenoid valve and hydraulic pump performance test was carried out.Test results show that the drive circuit is simple, practical and reliable; PWM which controls valves and hydraulic pump motor can be effectively regulated wheel cylinder pressure gradient, but there are some dead zones and saturation zones; using pulsefrequency conditioning can effectively adjust wheel cylinder pressure gradient. The results provide supportive basis for ESC system precise controlling solenoid valve and hydraulic pump motor.

Fully considereing the operating principle and pattern of ESP in real vehicles, used MATLAB/xPC realtime platform and CAN communication network to transmit all the signals needed by the ECU in the same way with real vehicles. Using CAN communication to monitor the working status of the ECU, the real vehicles were fully simulated and a closed loop system with drivervehicleenvironment was established. Compared with Signals of real vehicles, the test bench proved to have a high precision of simulation. Finally, the authors used the test bench to verify the ECU designed by themselves，the experiment and simulation indicates that the controller is qualified with satisfactory control effect.

Since that the existing SAE three-link leaf spring model represent hysteretic characteristic insufficiently, this paper, based on analyzing the mechanism of hysteretic characteristic, introduces the force element representing the dry friction between leafs into the SAE three-link model by using optimization method. The Comparison of the test and simulation results indicates that the SAE three-link model which imported the dry friction could simulate the hysteretic characteristic of leaf springs well.

An estimator is designed to estimate the articulation angle based on extended Kalman filter(EFK), where a 3 DOF tractor semitrailer vehicle model and nonlinear tire lateral force is employed, and the signals from sensors of electronic stability control (ESC) system for commercial vehicle are used. The estimator is validated on a 21 DOF nonlinear vehicle model and the results show that the articulation angle can be estimated accurately in both linear and nonlinear area of tire lateral force.

In order to meet the requirements of lateral velocity and longitudinal velocity in vehicle active safety control systems, a novel method based on unscented kalman filter (UKF) was proposed for estimation of lateral velocity and longitudinal velocity. This method was evaluated under a variety of driving conditions and the estimation values were compared with simulator values from CarSim. The results demonstrate that the proposed method is robust and can improve the estimation accuracy of lateral velocity and longitudinal velocity.

To compare with the results of automotive wind tunnel test, one virtual wind tunnel model was built. It has the same dimension and physical conditions with the actual automotive wind tunnel. By contrasting the test results of actual automotive wind tunnel and virtual wind tunnel，the feasibility of virtual wind tunnel was discussed. The results show that the virtual wind tunnel could get similar data of test to the actual wind tunnel test, and it could help us to observe the flow field and get more data we could not get in actual wind tunnel.

That the vehicle's crashworthiness is optimal designed in the conceptual design process based on simplified model can avoid repeatedly modified in the detailed design. A simplified model of the pendulum impact was built based on the detailed finite element model.The overall design objectives are first set and then they are broken down into several substructure performance curves with the simplified model, finally the section of the structure are determined aiming to theirs performance curves , which achieved CAE driven design process.

The material experiment about mechanical characteristics and microstructure were carried out. This is the base of application and side crash simulation of hot forming HSSS in car body. In order to improve the side crashworthiness of car body, the side structure of car body is substituted with hot forming high strength steel material. The side crash of changed car body is simulated to different combination of change. The results indicate that the side crashworthiness of car body is improved by application of hot forming high strength steel sheet.

To reduce carbody weight by using of magnesium  and aluminium alloy and ensure its structure performance simultaneously, the effect of sheet material and thickness on carbody structure performance and lightweight was discussed, and carbody NVH performance and vehicle sidecrash were simulated by finite element method. The result shows that using the lightweight material can reduce carbody weight, meanwhile ensure vehicle NVH and safety performance.

An approach of using artificial neural network (ANN) to evaluate vehicle nonsteady noise quality was introduced. The results of the interior noise quality subjective evaluation under the accelerating condition were taken as instructive samples, the inputs of ANN were loudness, roughness and level A, the subjective evaluation level of noise quality was the output of ANN. The Matlab neural network toolbox was used to build a back propagation (BP) network based on LevenbergMarguardt algorithm, compared with the multidimensional linear regression model, the results indicated that the neural network model was applied to evaluate nonsteady noise quality, whose performance index (PI) is 0.001, better than the multidimensional linear regression model (0.003).

For solving the problem of testing the automotive wheel alignment parameters noncontactly and accurately, a detection system of wheel alignment parameters based on binocular vision is proposed. The calibration method of the vision detection system is analyzed. The method based on distortion correction of lens and linear model is employed to calibrate the cameras. The calibration experiments are accomplished on the condition that the baseline distances of the vision system are 400 mm，600 mm，1000 mm, respectively. Experiments show that accuracy of the vision system increases as baseline distance increases. The mean error of 32 threedimensional points on the calibration board is 0.453 897 mm and the maximal error is 0.794 095 mm when the baseline distance is 1000 mm. The feasibility and accuracy of the test system are inspected by the experiments.

The objective of this investigation is to design sampling system of collecting exhaust emission by sampling bag and absorbing formaldehyde and unburned methanol in exhaust emission with pure water, to establish measurement system in which combine with gas chromatography and liquid chromatography to separate and measure formaldehyde and unburned methanol emissions, and to measure formaldehyde and unburned methanol emission during cold start by using this measurement method from a inlet port fuel injection sparkignition methanol engine by means of cyclebycycle control strategy. The results show that the combination method of gas chromatography and liquid chromatography may separate and measure the formaldehyde and unburned methanol emission effectively. The detection limits of formaldehyde and methanol for gas chromatography are 0.16 μg/mL and 0.18 μg/mL, respectively, and the detection limits of methanol is 0.15 μg/mL for liquid chromatography. This measurement method can satisfy measurement requirement for formaldehyde and unburned methanol of methanol engine.

According to engine parameters and lowspeed steadystate data of certain model Mitsubishi GDI, a nonlinear timedelay mean value idling mode was developed. The delay includes inductiontotorque delay under steady state of engine, includes delay between time of air input controlled by air throttle to actual time of this section into the cylinder under transient state of engine, also includes delay of the action of the air throttle. Aiming at torque interfere generated by auto accessories using of electricity under idling mode, the timedelay H∞ state feedback controller was designed. This not only guarantees closedloop control system asymptotic stability, but also meets suppression ability as  γ about the output to disturbance input. Finally, making use of the exact engine model established by en_DYNA software to verify the modeling correctness and validity of control method. Simulation results showed that: the control effect is improved, the range and frequency of motion of the control actuator is reduced by taking into account the inherent timedelay phenomenon of engine when designing controllers.

Different factors which influence urea dosing control have been studied and a control model was established. The control strategies were realized by RCP and were validated on engine bench. Ammonia slip amount is very low by NOx transient control strategy. For both ESC and ETC cycle, China IV emission regulation were fulfilled.

As the critical issue in emergency evacuation, aiming at the shortest evacuation time, we build a route selection model in emergency evacuation based on quasiuser optimum dynamic traffic assignment theory, and develop a constrained Kshortest paths algorithm within a dynamic restricted searching area. Efficiency and feasibility of the model and algorithm are proved by the simulation results.

A dynamic traffic assignment model was proposed based on emergency evacuation passageway under the condition of disaster. The traffic demand and the road impedance under the condition of disaster were analyzed. The precondition of model was described, the dynamic traffic assignment variational inequality model based on user optimal was established, and the algorithm of the model was designed. The example analysis showed that, evacuating the regional traffic influenced by the disaster with the model, could guarantee the smooth characteristic of the emergency evacuation passageway and the optimal of user's path, and avoid the increase of the evacuation time because of congestion.

For a better solution of the incoordination of regional control strategies in traditional traffic control system, this paper proposes a traffic coordination control strategy of regional boundary on the basis of traffic state identification and develops a traffic coordination control model based on fuzzy control and hierarchical coordination. The actual effect of this model is tested with the micro simulation tool VISSIM, and the result shows that the model could efficiently coordinates the regional boundary with its upper, lower intersection so that could assure the rapid evacuation of traffic congestion and increase the operation benefit of the whole system.

The queue process in the intersection was firstly analyzed, then from the view of avoiding the spillover of queue vehicles, a maximum cycle length optimization model was presented aiming at the queue length of the intersection was shortest over the whole duration of saturated traffic state, then the effectiveness of the model was validated based on Vissim environment. The results showed that when operating cycle length was smaller or larger than the optimal value, the queue length and average vehicle delay would increase. This model can improve the benefits of the intersection effectively.

In order to improve the critical volume conditions of pedestrian signal, the delay model of no signal branch road section under condition of the equal right of way for pedestrians and vehicles was built. Moreover, the partitioning standards of pedestrian group and the calculation method of conversion coefficient of pedestrian crowd and vehicles were put forward. Based on that, the arrival rate of pedestrian crowd was converted to the arrival rate of vehicles, and equilibrium phase delay formula was used to calculate the delay of signalcontrol branch road section. The critical volume curves of pedestrian signal for branch road section were obtained by comparing the delay under branch road section condition with or without traffic signal.

According to the problem that differences of running states between floating cars and traffic flow are neglected about traffic information collection technologies based on floating car data, which leads to traffic information obtained with low quality, a method for determining the quantitative relations between link average travel time of traffic flow and link average travel time of floating cars is designed using statistical techniques, and a method for estimating link average travel time of traffic flow is put forward by fusing measured data and historical data, then the methods are verified based on measured floating car data. The results show that the methods mentioned above can produce link average travel time with higher quality, which will provide more reliable decision information for traffic management departments and travelers.

In order to improve the efficiency and effectiveness of traffic organizational decisionmaking for emergency evacuation. On the basis of information centrality theory, the concepts of key nodes, efficient key nodes and connectivity key nodes was put forward, and the method of identifying each kind of key nodes and key routes for emergency evacuation was designed. A case study based on the actual road network and practical traffic data was presented to demonstrate the effectiveness and feasibility of the proposed method.

Main factors on taking measure of speed reduction were analyzed when one driver meets the dangerous situation of road in advance. Mechanism and constrains of psychology which induce the speed reduction behavior were discussed. To combine driver's risk avoidance behavior modes with behavior control model, the mechanism and the model of speed reduction behavior were built up based on capability of risk cognition and driving control. A case study was implemented by using data of survey of vehicle speed reduction behavior at curve section in twolane highway. Experiments and analysis results show that the degree and driver's psychology process of speed reduction is reflected distinctly by using the model when drivers are driving in highway with a speedlimit sign.

An evaluation index system of TDM was developed. Then an evaluation model system with universal applicability was developed, using Multinomial Logit model, Nested Logit model and activitybased transportation demand forecasting theory. With this model system, the acceptability of the public transport priority policy was predicted. The results indicate that the TDM policy can be evaluated comprehensively with the evaluation model developed in this paper under aspects of traffic impact, user satisfaction, etc. The study of this paper can provide decision method for government to carry out rational and effective TDM policy, and it is helpful for the popularization of TDM theory and the solving of urban traffic congestion problem.

In order to study the road landscape impact on the traffic safety by simulation, taking driver sight distance as a starting point, the driver model, vehicle model, road model and landscape model were built based on considering the main impact factors and the impact discipline. These models provides basis for estimating the effect of landscape impact on traffic safety in simulation.

In order to improve the management quality of hazardous materials transportation and identify the security management weak link of hazardous materials transportation process, a new method combining with fishbone diagram and fault tree was proposed to improve the hazardous materials transportation process management. Firstly, all the influence factors of hazardous materials transportation process management were systematically analyzed by fishbone diagram method. Then, according to the logic relation of these influence factors, the fishbone diagram was transformed into a fault tree analysis model. By calculating structure importance of the fault tree, the importance degree sort of the influence factors was gained and then the main influence factors such as transportation declaration system defects, multidepartment information communication blockade and cross enforcing law were determined. In the end, it provides the corresponding process optimization management strategies of hazardous materials transportation. The research conclusion shows that it can efficiently solve the outstanding problem in hazardous materials transportation process management and prevent hazardous materials transportation accident. In addition, it provides a new idea for using the quantity analysis model to the process optimization management.

In order to ensure the safety of Metro intelligent transport network and the reasonable utilization of resources, AAA（Authentication,Authorization,Accounting) mechanism was introduced to the intelligent transportation system, which is used to design the operation mode for the managent of Tansportaion. A plan was proposed, including the security access to the Metro intelligent transport network for  the  moving vehicles, seamless connection between access points, the accounting for the information of both network and intelligent transportation system, with the algorithm for authentication, authorization and accounting between realms. AAA algorithm was composed with C language，and tested by the data coming from the simulation using Vissim4.20.

A method on occupant detection under a single video camera was presented, which used a window classifier to explore occupant header in the image. The window classifier was constructed by histograms of oriented gradients and support vector machine classifier, and it was trained by occupant header and nonoccupant header. To get multiscale detection results, nonmaximum suppression was used to get the maxim point in the space of position and scale, which is occupant header region. In experiment, 200 occupant header samples and 200 nonoccupant header samples under occupant different pose condition were trained to get the classifier mode between occupant header and non occupant header, and occupant videos including different occupant poses were tested by the proposed method. The results show the method can satisfy the need of occupant detect in smart airbag application.

Compound modified asphalt (CMA) was created by using waste tire rubber powder, SBS, asphalt and aromatic oil. The absorption coefficient experiment of CMA was designed by orthogonal experiment scheme, the experimental data are analyzed by variance analysis. The test results show that the most optimum mixture ratio of mass fractions for CMA is 5% rubber powder, 3% SBS, 7% aromatic oil.

In order to enable the asphalt pavement of Highway Tunnel has a good performance, adjusted asphalt mixture in the above 4.75 millimeters thick aggregate gradation, and add nontoxic, low smoke and can improve the road performance of the composite fire retardant. Studied the changes oxygen index after adding flame retardant asphalt, increased in the above 4.75 millimeters the proportion of coarse aggregate and added flame retardants impact on the road with asphalt mixture performance. Test results show that flame retardant SMA asphalt mixture has good high temperature stability, low temperature crack resistance, water stability, bonding, and resilient modulus compressive strength and splitting. Moreover, the best flame retardant content is the the 7% of asphalt quality, at this time regardless of ambient temperature or high temperature, the oxygen index of asphalt can reach 33.

In order to study the damage diagnosis of concrete structures under ambient excitation, an effective diagnostic method of structural damage location and assessment degree of damage was discussed based on modal strain energy theory. Through improved the existing modal strain energy indicators, a damage diagnosis method based on crossmodal strain energy was proposed. Using the method the rapid damage diagnosis of structure was made with low order modals information. Through the establishment of concrete simple beam model of threedimensional solid, the finite element numerical simulation was performed to verify the effectiveness and feasibility of proposed method. The results of the present study were expected to be useful for the health monitoring technology of concrete structure and provides a theoretical basis.

To enable polypropylene fiber reinforced concrete has a good performance, silica fume and slag were added to raw materials in the concrete to improve compressive strength and splitting tensile strength. By orthogonal design and theoretical analysis, the effects of different content of polypropylene fibers, silica fume, slag on properties of concrete were studied. The experimental results show that the effect of silica fume on the properties of concrete is particularly significant;incorporation of silica fume concrete can enhance the compressive strength and splitting tensile strength; the use of polypropylene fiber composites, silica fume, slag can improve the early strength of concrete; the best amount of adding poly acrylic fiber, silica fume, slag in concrete is of 0.015%, 7%, 7% respectively.

Using the performance test method, the reliability of the reinforced concrete structural element was studied. According to deflection calculations, the average reliability index is 1.20, the failure probability is approximately 15%, the proportion of the live load to the reliability index is the most important factor. The greater the proportion of the live load, the greater the factors that can not be determined,the greater the correspoding reliable indicator. According to the result of anticrack analysis, the average reliable indicator of the first grade anticrack assessment is 3.8, the corresponding failure probability is approximately 0.03%; the average reliable indicator of the second grade anticrack assessment is 3.0, the corresponding failure probability is approximately 0.44%. The different anticrack coefficients will lead to the different reliable indicator, the larger the anticrack coefficient to the first grade structural element, the higher the reliable index.

In order to investigate the method of electromagnetic measurement of stress, an experimental research of the relation between stress and induction EMF (electromotive force) was performed for Q235A parallelchord steel truss. Two types of magnetic core sensors, planar or arc, was put respectively in the surface of chord and vertical web member flange, Ishaped steel of 12#, or in the surface of diagonal in diameter 25 mm. The magnetic line of force was made to penetrate the partial area of member and EMF was induced at the point of measurement. When the load was acted on the truss, the data of induction EMF and strain were measured at the same time. The experiment indicates that the change law of stress and induction EMF hold unanimity, within the range of loading. With the compressive stress increasing, EMF of upper chord and vertical web member decreases. With the tensile stress increasing, EMF of bottom chord increases, and the algebraic value of negative EMF of diagonal rises. The experimental datafitting result accords with the relation between load and EMF from the analysis of stressmagnetic moment deflection.

Hotforming process of UHSS is simulated, using ABAQUS software based on a thermomechanically coupled thermoelastoplastic finite element method, and specially some effects of blank holder force on problems such as thinning and springback for hotforming of Ushaped workpiece are analyzed. The results showed that the springback can be restrained effectively by increasing blank holder force, and the forming accuracy will be improved, but the thickness is reducing with the increasing blank holder force. The larger the bank holder force is, the severer the thinning will be. Obtained results can be applied to development of hotforming process of UHSS workpiece.

To improve the surface wear resistance of the ceramic coating fabricated by microarc oxidation on aluminium alloy, ceramic coatings embedded with SiC nanoparticles at different current frequency (500, 700 and 900 Hz) were produced by adding SiC nanoparticles into electrolyte during MAO process. The morphology and phase composition of the coatings were investigated by Field Emission Scanning Electron Microscopy (FESEM) and Xray diffraction (XRD). The surface roughness of the coatings was characterized by Confocal Laser Scanning Microscope (CLSM). Simultaneously, the tribological properties of the composite coatings were evaluated by a universal materials tester (UMT). The results showed that the wear resistance of the ceramic coatings with SiC nanoparticles was improved significantly compared to the ceramic coatings without SiC nanoparticles. Furthermore, the composite coating formed at current frequency 900Hz showed the better wear resistance than the others.

The advanced laser welding technology was used to connect the ultrahigh strength steels and highquality welded joints were acquired.The author observe the microstructure, test the hardness of the welded joint and analyze the fracture mode of the welded joint by using scanning electron microscope. The results show that good welding quality can be obtained by means of laser continues welding method，and the welded joint internal structure is dense and has no the defects such as crack and shrinkage hole. The shear fracture visage shows that the fracture mode belongs to both ductile fracture and brittle fracture.

Based on the analysis of tapered shaft forging by deformation of cross wedge rolling, using the platform of threedimensional finite element software DEFORM3D, cross wedge rolling process of conical shaft with different parameters was simulated,and cross section deformation strain field, stress field in the rolling process were obtained. The influence law of rolledpiece cone, section shrinkage rate on stress and strain during deformation were analyzed. The results show that stress and strain increase with the increase of section shrinkage rate,and internal strain decreases and stress almost remaines the same with the increase of cone angle.

Bending process of outside profile pillar assembled in highspeed railway vehicle was simulated and analyzed using finite element software PamStamp 2G for researching bending process of large scale profile with complicated nonsymmetrical hollow section. Methods for controlling bending defects such as cross section distortion, line segment curve and low longitudinal profile precision were proposed. The results show that line segment curve can be improved effectively by adjusting forming direction and using back load is. The larger the back loadis, the smaller the curve value of line segment becomes. A new mandrel was used for controlling cross section distortion and improving longitudinal profile precision, and the smaller gaps between mandrel and inside wall of profile yields smaller section distortion value and higher longitudinal profile precision. The rationality of optimized process was tested and verified by experiments.

Thin nanocrystalline Cu films are deposited on a glass substrate using an improved electroless plating technique and Cu films oxidized using heat oxidation technology, Nanocrystalline Cupric Oxide (CuO) films were fabricated on unetched glass substrate. The process of Cu films oxidized and crystalline structure and grains size were investigated by XRD、FTIR and XPS. The surface morphologies of CuO films with different thickness were characterized by FESEM. Experimental results show that nanocrystalline Cu films were directly oxidized to form nanocrystalline CuO films which have nanonano structure at 350 ℃ for 1 h. With the thickness of CuO films increasing from 54 nm to 113 nm, the average grains size increases from 12 nm to 23 nm, and the agglomerates size increases from 30~50 nm to 60~100 nm, and the optical transmittance ratio decreases from 74.7% to 65.17%, and the range of band gaps is from 2.83 eV to 2.94 eV.

Aiming at the problem of tracking and measuring moving object under complex background, a new method is proposed, which is based on machine vision, called “centeriteration”. This method takes advantage of image analysis technique to detect the profile of specified moving object and compute its center in each frame of the video from two camera. According to binocular theory, the moving object's position in the world coordinate is acquired. Therefore, the aim of tracking moving object is achieved. Experimental results show that the proposed method could track the moving object effectively under complex background.

The occurrence of CNC machine tool's failure is random, and the failure effect is fuzzy. Traditional criticality analysis is based on the arithmetic product of basic failure rate, failure mode probability value and failure effect, and the criticality analysis result can only reflect the average degree of criticality within a certain period. To solve this problem, a dynamical criticality analysis method based on random and fuzzy theory was proposed. So the randomness of failures and the fuzzy of failure effects can be comprehensive considered, the contribution of each factor to the analysis result can also be considered, and multiple effect factors analysis can be carried on. Taking the six subsystems of some certain CNC machine tool for example, proposed method was proved to be accurate and effective and has high application value in the reliability research of CNC machine tool.

Failure number reliability influence model, failure downtime reliability influence model as well as dynamic reliability influence model of NC machine tool were established. Some subsystems reliability influence for failure data of domestic NC machine tools were calculated and ranked. Finally, the three model's results were compared and analyzed through appropriate method,and the key subsystems were found and this result will be helpful for failure analysis and reliability improvement.

Against the situation that the manual diagnosis on the equipment operation condition is deficient, the application of extenics has been employed to design a diagnosis system on the equipment operation condition. The system integrated the data collection, display, store, analysis, diagnosis and warning into one, evaluated the extenics model and the extenics model of the equipment operation condition by considering the importance of the equipment. The system has an intelligent function on the condition diagnosis, could make diagnosis and identification on multilevels and multiangels based on the premonitory parameter and equipment information. So it provides a software tool for the application of the extension in the diagnosis system of the equipment operation condition.

Based on the experimental optimal design for the experimental scheme, the laser graphic texturing was applied to process the grid bionic surface micromorphology on the roller surface. The experiment of  antifatigue capability of the roller was performed. The results show that the pitting ratio of all the roller specimens with the grid micromorphology had been decreased, while that of the ordinary roller had been increased. The contact fatigue spalling appeared on the ordinary roller surface. The antifatigue capability of the ordinary roller is much lower than that of the roller with micromorphology. The antifatigue capability of the roller had been improved to 150% by the grid bionic surface micromorphology.

This paper reconstructs the wing and spanwise airfoils of owl using a threedimensional scanner. Referring to the flying characteristic of owl, the aerodynamic of spanwise airfoils are simulated at different angels of attack (AOA) and different Reynolds numbers. Calculation results show that the ratio of lift coefficient to drag coefficient is the largest when the AOA ranges from  3°to 7°,and the best aerodynamic performance airfoil on the whole at spanwise fraction from 40% to 50%.

Nonlinear time series analysis and its application in the quantification of dynamic stability were discussed,and VICON MX 3D motion capture system was used to collect the dynamic data of hip, knee and ankle when human were walking on treadmill at 40%, 60%, and 80% of the Froude velocity. Matlab was used to realize the calculation of delay time τ through CC method and the calculation of embedding dimension m through falsenearestneighbor analysis, and finally the largest Lyapunov exponents λmax under different speeds was calculated.Then λmax was adopted as a quantified index for stability to study the relationship between speed and dynamic stability. The smaller λmax valued, the better stability was. The results show that λmax increases linearly and stability decreases with the increase of speed.

Based on the fuzzy mathematics, elementfree Galerkin method (EFGM) was employed to deduce FEFGM equilibrium equations, and the algebraic properties of interval number was used to solve the FEFGM equilibrium equations. By the procedure, the possibility distributions of responses of fuzzy structures can be generated in terms of the membership functions of the input fuzzy numbers. FEFGM was used to compute the displacement fields of crack tips and direct displacement method was used to compute the stress intensity factors. Numerical examples demonstrate the efficiency and accuracy of the proposed method on calculating the stress intensity factors.

To overcome shortcomings of large amount of computational efforts and long computation time existing in Monte Carlo simulation for system reliability analysis，an improved method was proposed in the light of curve fitting. Reliability problem was reformed by depending on an optimal parameter, and the failure probability was treated as a function of this parameter by exploiting the regularity of it and fitting the reliability function by LevenbergMarquardt method. A tenbar truss structure numerical example was given to prove the feasibility of proposed method. Result indicates that the proposed method has high computational accuracy with less computational efforts and  computation time，providing an effective way to analyze reliability of large complex structures.

For highcomplexity of H.264 algorithm and low processing power of embedded, put forward a new motion estimation search algorithm MTS, so that H.264 is more suitable for an embedded platform. MTS, a new motion estimation search algorithm, uses continuous transformation triangle regularly, and with crosssearch mode to search. The results of experiment based on ARM9 S3c2440 embedded platform, show that the using MTS algorithm can improve H.264 compression speed nearly doubled, under the premise of the video quality and rate, which is better for ARM9 embedded platform.

For the capabilities limitations of mobile video system based on embedded, designed the application software on the thread in acquisition side and monitor side, to improve the video processing speed. Designed the mobile video system based on ARM9 S3C2440 hardware platform, and do the research the thread of the application software. Change the single thread to double thread on the collect side, monitoring side is designed with one thread, which coordinated video collection, codecs and player. Experimental results show that the speed of video processing increase effectively, improving two frame per second or so, which is more to meet the needs of mobile video applications with only 400 MHz frequency of the platform.

In order to improve the convergence of BGP (Border gateway protocol) routing effectively, an improved algorithm was presented based on the model of SPVP (Simple path vector protocol), and the Tdown convergence boundary value of the improved algorithm was discussed, which can quickly detect and identify the key node of AS(Autonomous system) link failure and reduce quantity of invalid BGP routing update messages and increase speed of BGP routing convergence. The validity of the proposed algorithm was demonstrated by simulation software of SSFNet in two different network topologies.

Problems always appear as P2P is applied to instant messaging systems, such as long latency, frequent and unwonted leaving online, bad performance of locating and looking up object nodes and so on. Based on hybrid P2P, DHT and the Kademlia algorithm, a new P2P model was provided, the important process of constructing, such as looking up object nodes, nodes joining and leaving the P2P networks was described, and carrying out of each process was given in detail. The testing result shows that the model is better than hybrid P2P in looking up object nodes and latency, and the application result in realtime communication system improves that the model is correct, efficient and stable.

A new method for resampling of pointbased models was presented. A novel point primitive called "spherical surface point" was used to approximate the vicinity of a model point. Resampling was accomplished by inserting appropriate points between resampling related points and their neignbor points. Ray tracing was used to render models. Experimental results show that the proposed resampling method effectively solves the undersampling problem of pointbased models and achieves satisfied visual quality.

A new hazard evaluation model was proposed based on TAN classifier，this model adopts probabilistic method. First, historic hazard evaluation data set by experts was built, and then using these data TAN model was built, and finally this model was used to evaluate new hazard in fireworks production area and the results is reasonable,proving the effectiveness of proposed method.

Based on Markov chain，a new method was proposed to predict the RNA secondary structure.Firstly，prefix and suffix matching algorithms were applied to find all the possible (including the pseudoknot) stem zones quickly, then free energy of stemloop stems was calculated, and finally RNAML was established to find the secondary structure with the minimum free energy. The experimental results show that proposed method reduces the computational complexity and improves specificity and sensitivity of prediction with high accuracy, so can predict pseudoknot structure.

The stripe noise removal method based on block optimal linear predictor is developed to reduce strip noise effects in moderate resolution imaging spectrumradiometer(MODIS) imagery. Through analyzing the correlation between detector images, the original image is divided to several detector images which include stripe noise images and nostripe noise images. By using the block optimal linear predictor, the new detector images can be obtained with nostripe noise images, thus the stripe noise can be removed from the original image. Experimental results demonstrate the proposed method can achieve better results than present methods, which effectively remove stripe noise and preserve most information of original image. The method is proved to be perfect for stripe noise removal in MODIS imagery.

A multi-agentbased eventdriven intelligent scheduling system structure model was presented. In the model, Petri nets act as eventhandler to describe the dynamic status of agent. Coordinator dynamically establish a multiagent coordination network according to the events occuring in the runtime. Using Petri net, agent status can be analyzed and forecast, and then multiagent scheduling queues can be determined and optimized, ensuring multiagent system to carry out tasks effectively.

To improve the ability of structural damage detection and its noise restraining ability, damage detection was transferred into optimization problem by the concept of residual force vector. And the multiobjective model was proposed for structural damage detection. Because of its higher dimensions, the improved grey particle swarm algorithm was presented to improve the model precision. The experiment on structural damage detection of cantilever beam shows that the proposed method is efficient on damage detection.

For the purpose of improving the safety and reliability of mechanical components, the theory of Reliabilitybased robust optimization design (RROD) and multiobjective decisionmaking approach are employed in this paper. The multiobjective optimal modes are established for RROD of structural system.The Pareto sets for constrained conditions are solved by fuzzybased multiobjective optimization. And the component size is decided by Pareto set and practical cases. Experimental results show that the information of reliabilitybased robust optimization design can be obtained effectively and quickly by this method through the experiment of RROD for drag rod with normal distribution parameter and arbitrary distribution parameter.

Considering the social foraging swarm task under the uniform viscous environment, and the concept of internal average kinetic energy is introduced. Adjacent matrix method is presented in this paper which is due to the distance of robots. A decentralized controller is designed based on local information interaction, then analysis of system stability is performed, Simulation experiments have been done to verify the designed controller. The results show that the proposed controller enables the swarm robots system to reach forage task more efficiency.

This work presents an adaptive synchronization between two different order fractional chaotic systems with unknown parameters. Based on the nonlinear feedback, the proposed adaptive scheme not only can estimate the unknown parameters successfully in certain condition, but also realize the synchronization of chaotic fractionalorder system. Taken the fractional liu system and Duffing system as an example, numerical simulation results verify the effectiveness of the proposed scheme.

On the foundation of analysis of expertise character of assembly sequence planning process, a method of obtain assembly sequence expertise knowledge of assembly components is proposed based on rough set. Finally, the algorithm is realized by VC++ and the method is proved for its validity by assembly model of engine.

An image feature fusion method was proposed based on scope similarity scores minimization. Aiming at the uncertainty of image feature fusion, it converted the problem of fusing image features into that of minimizing similarity scores of restricted samples on line. From different feature space, the similarity scores between a query image and images in database were calculated, and then some of them were selected to construct a candidate image pool, the size of which was relative to the number of return images. For the images in candidate pool, AG algorithm was employed to select return images, the purpose of which is to minimize the total similarity scores of return images. By this way, the adaptive weights of different image features were gained, and similar images were returned. The experimental results show that the proposed method is competitive.

Some classic methods for DOA estimation about wideband signals need constructing focusing matrices and preestimating directionsofarrival(DOA) of wideband signals. So the focusing matrices and estimative directionsofarrival of signals will directly influence the performance of the classic methods. Based on subband data matrix, a set of new matrices was reconstruct, and the DOA estimation was carried out using MUSIC method. The simulation results show that the performance of the proposed method is close and even superior to the RSS method.

A multiplecell coordinated multiple point (CoMP) scheme based on limited feedback was proposed, then specific operations at the transmitter and the receiver were described. A feedback scheme for general unitary matrices was presented where the QR decomposition precoding was used at the transmitter. Moreover, an enhanced feedback scheme with minor quantization error and less feedback overhead was presented because of channel quality information sharing with the cooperatingcells, which improved the system performance dramatically at the expense of a few feedback bits, since the multiple subchannels were combined coherently at the receiver. The equivalent channel matrix was not a triangle matrix due to the quantization error. Hence, a modified successive interference cancellation (SIC) receiver was presented. Simulation results demonstrate that the proposed scheme outperforms traditional ones on the system throughput.

The virtual source method was proposed to improve the reconstruction accuracy of acoustic pressure fields. Sound field of the sound source surface is reconstructed by acoustic holography to determine the initial location of the sound source. The source is virtualized and the ideal location is searched using nonlinear damped least squares to determine the reconstruction result of the sound field. This method will strip the influence factors of the reconstruction accuracy and avoid the errors from observation data and other errors origin. The simulation validation and experimental results show that the reconstruction error is only 8.6%, which are better than existing methods significantly. The reconstruction accuracy of the sound field is improved effectively.

Facing the new requirements of chip development for arc welding inverter power source, a new dualDSPbased control program was designed, a rapid twomachine communications was achieved through the DSP chipchip SPI communication module, which communication rate is up to 10 Mbit/s. The CRC 16 methods of communications test was used,the results indicate that error detecting rate is as high as 100% when data bits less than 16.

In this paper, the body weight and average daily food intake of mice which were reared for 30 d or 90 d were determined in order to optimizate different compound of the bioactive peptides derived from egg white protein which had no effect on body weight changing condition and food intake. The results showed that: no matter the egg white hydrolysates which only hydrolyzed by Alaclase or the high F ratio oligopeptide under through both Alaclase and flavorenzyme, they had no significantly effect on the body weight of mice reared for 30 d and 90 d compared to the blank group (p>0.05). Egg white hydrolysates and the positive control (Tineng soybean peptide)could reduce the average mouse food intake in different degrees, but they had no significantly effect compared to the blank group (p>0.05) in 30 d. Except for 4.5 mg/(g·bw·d), different concentration of high F ratio oligopeptide also had no significantly effect compared to the blank group (p>0.05) at 30 d. During considering the effect of different bioactive peptide from egg white protein on food intake of mice which were reared for 90d, we found that except for the EWPE4, bioactive peptide from egg white protein could significantly decrease the average daily feed intake (p<0.05 or p<0.01).

This research focused on the steady state protection factors of the antioxidant peptide derived from egg white protein which was hydrolyzed both by Alaclase and flavorenzyme at the pilot scale. The DPPH scavenging activity of antioxidant peptide from egg white was selected as the index to investigate the effects of temperature, pH, food ingredients and metal ions on the stability of antioxidant activity. The results show that antioxidant activity of peptide derived from egg white protein was very strong and the IC50 of DPPH free radical reached to 1.07 mg/mL. The scavenging activity got to 94.34% when its concentration was 5 mg/mL. The antioxidant peptide from egg white treated for 1 h to 3 h has a strong heatresistant ability at 25~40 ℃. And it can keep high activity at pH 2～10. Sucrose had no significant effect on it (P>0.05).Glucose and NaCl changed the activity significantly (P<0.05), and NaCl concentration relatived to DPPH antioxidant activity positively. K+, Ca2+, Mg2+, Fe2+, Zn2+ ions all improved the antioxidant activity of it in which Fe2+, K+, Zn2+ could enhance the activity significantly (P<0.01 or P<0.05 ), while Ca2+, Mg2+ had no significant effect on it(P>0.05).

Mice drank pure water, tap water and mineral water from Changbai Moutain. The functional properties of the mineral water were investigated via the studies on the mice's growth, calmdown activity, antifatigue behavior, memory ability, exploration behavior and immunity. The results have shewn that the mineral water has greatly improved the mice's antifatigue ability, calmdown ability and immunity, but do little on their memory and exploration behavior.

In order to improve the activity, stability and reusability of watersoluble βglucosidase in the reaction of alkylpolyglycoside synthesis, chitosan and sodium alginate, a new composite vector (alginatechitosanalginate, referred to as ACA),was used  to immobilize βglucosidase, and the mathematical optimization model of immobilized enzyme's technological parameters and mathematical model of enzyme activity (glucose conversion rate) were studied in detail. The results show that chitosan, glutaraldehyde, trigger time in CaCl2 and sodium alginate are influence factors of glucose conversion rate successively; optimization parameters of immobilized enzyme are chitosan of 0.2 g, sodium alginate of 0.6 g, trigger time in CaCl2 of 30 min, glutaraldehyde of 0.25 g/L, and glucose conversion rate is 85.8%; chitosan and wall materials solution show a positive interaction among trigger time in CaCl2, glutaraldehyde and sodium alginate; wall materials solution show a negative interaction between trigger time in CaCl2 and glutaraldehyde; optimized glucose conversion rate and main factor code value are regression mathematical model of variables secondary universal rotary combination design and regression mathematical model of variables and test effect.

Adding the milk, egg, apple juice, mushrooms and kinds of accessories additives to the meat mixed with pork and chicken, the fruit and vegetable compound fermented sausage was developed by the fermentation of Lactobacillus and Bifidobacteria. Optimizing the craft and the formula and analyzing the properties of fermentation by the orthogonal experiment, range analysis and sensory evaluation, the optimal ratio of bacterium (1∶1),the optimal quantity of bacterium(1g/kg),the optimal time of fermentation(26 h), the optimal temperature for fermentation(35 ℃) were obtained. The optimal main formula were apple juice (6%), the ratio of pork to chicken (2∶1), mushrooms (10%), isomaltooligosaccharide (3%) (The percentage was compared with the pork and chicken). Increasing fermentation temperature, the reduce rate of the pH, water and nitrite increases, the amino acids and the peroxide value increase obviously; the amount of bacteria has significant effect on the amino acid; the fermentation time and temperature has significant effect on the product value,and the fermentation time has obvious effect on the nitrite reduction; The result of the analysis of physical and chemical properties of the optimal sample and the optimization of process parameters basically consistent.

A new method was developed for realtime monitoring the Nisin titer (NT) and the concentration of cell during Lactococcus lactis subsp. Fermentation by near infrared spectroscopy (NIR) combined with radial basis function neural network (RBFNN). Three different 5 L fermentors were applied to implement 15 batches of the Lactococcus lactis subsp. Fermentation. Samples were collected interval 1 h during these fermentations. The Nisin titer and the cell concentration were determined by reference methods. At the same time the NIR spectra of the samples were recorded using UVVisNIR spectrophotometer. RBFNN was applied to model the relationship between NIR spectra and Nisin titer and the cell concentration. The RBFNN models were optimized by selecting efficacious preprocessing methods, wavelength, the number of hidden nodes and the spread constant. The optimum models for determination of the NT and the cell concentration were developed. The coefficient of NIR predictive values and reference values of calibration set were 0.8649 and 0.9914, the root mean square error of prediction set (RMSEP) were 2865.05 and 0.1414,respectively. These results demonstrate that fit and the predictive capability of these models were satisfied. This method can be used for monitoring the key parameters during fermentation processes.