Incorporating the 2 traditional vehicle stability control methods, the one is to follow the ideal yaw rate, the other is to limit the mass center sideslip angle, a new optimal control method was proposed. Following the linear 2-DoF vehicle handling performance model as control target, a control system model was built based on the dynamic relationship between the vehicle yaw moment and the vehicle state deviations. Using the Linear quadratic regulator(LQR) method to decide the vehicle yaw moment required to stablize the vehicle state. The control performances of the LQR method were verified by simulations and tests on the hardware-and-driver-in-the-loop simulation test rig.

A 14DOF vehicle dynamics model was built in ADAMS/Car based on the conceptual suspension module. The dynamic simulation results using the model were in good agreement with those of road test. The effect of the properties of the front suspension of the understeer rate was analyzed using the model, and it was found that the C property of the front suspension shows larger effect on the vehicle understeer. The effects of the K&C properties of the front suspension on the vehicle handling stability were analyzed, introducing a train of thought to dissolve the whole vehicle performance to the suspension K&C properties.

The basic structure of the electronic hydraulic brake system(EHB) was briefly introduced and a hydraulic system model was built for the brake circuit when the EHB works normally. A control strategy based on a single variable—the vehicle yaw rate—control was proposed to control the vehicle stability. The simulations for the vehicle stability control with EHB were performed under typical conditions. The results showed that the proposed vehicle stability control algorithm of EHB can ensure the vehicle stability effectively under the high vehicle speed and low road adhesion factor conditions.

Pulley Deflection losses of Continuously Variable Transmission is the one of the most important losses. Reduce the Pulley Deflection losses can improve the transmission efficiency. For this the relation between the loss and the transmutation is analyzed. The CVT-transmission model is founded on ANSYS/LSDYNA. And simulation is carried out under different stiffness and different transmission ratio. In the end the influencing factors of the energy losses of CVT is analyzed. The conclusion show that this way is a feasible way to find the fittest stiffness which can improve the efficiency of the M-belt CVT.

The effects of biodiesel blending fuels with different biodiesel fractions on the exhaust emission characteristics of the diesel engine were investigated on a heavyduty automotive electronically controlled highpressure commonrail diesel engine. Five diesel/biodiesel blends were tested according to ESC steadystate 13mode test cycle without modifications of engine construction and parameter adjustment. The test results showed that the brake specific emissions(BSEs) of CO, PM and DS decrease with the increase of biodiesel fraction in the blending fuel, while the BSEs of HC and SOF decrease a little, but the BSE of NOx increases gradually. Compared with BD(pure diesel) the 13mode weighted average BSEs of CO, PM and DS of fuel B30(30% biodiesel) decrease by 46.0%,27.5% and 44.2% respectively, while the BSE of NOx increases by 10.9%. The weighted BSEs of CO,PM and DS of biodiesel decrease more obviously under high load conditions.

A new membrane device for gas separation was developed and tested on an engine. The tests were done in the pressure operation mode and vacuum operation mode respectively to investigate the flow characteristics of both streams of oxygenenriched air(OEA) and nitrogenenriched air(NEA), as well as the performance of gas separation process. The results showed that in the pressure operation mode, the composition concentrations of both streams are in inverse proportion to mass flow rate, which is in favor of the preparation and regulation of NEA stream. In the vacuum operation mode, the concentration of OEA stream is in direct proportion to mass flow rate, which is in favor of the preparation and regulation of OEA stream. It was found that the air supply pressure, the suction vacuum and the pressure difference between both sides of the membrane are the decisive factors for gas separation, and the temperature is also significant. Therefore, it is necessary to incorporate all factors, coordinate the flow rate and gas concentration to meet the requirements of engine intake process.

A 6 DoF tractorsemitrailer train model was established and combined with a driver model incorporating the path preview, the lowand highfrequency compensatory gains, the time delay, and the vehicle state prediction function. The design parameters of the train coupled with different drivers were analyzed. The GaussNewton method was used to define the ranges of the train parameters to adapt the drivers with different skills. It was concluded that the safety performance of the train can be enhanced through optimization of the train parameters. A driver with excellent driving skill can easily adapt a train with large size, soft suspension and high degree of oversteering. The adaptation of the train to the driver reduces the driver steering effort, vehicle roll angle and yaw rate remarkbly.

The selection principle of flywheels in the inertial system of the automobile chassis dynamometer was analyzed to enhance its test accuracy. The gradation principle of flywheels was proposed based on the double multiplication method, and an inertial system with 9 steps of flywheel groups was designed, the simulative mass ranging 500~10 000 kg was divided into 512 levels with accuracy of ±10 kg. The accurate matching between the simulative flywheel mass and the tested automobile mass was realized. The optimal design of the flywheel was performed. The braking method of the flywheel was studied to remove the additional inertia moment from the inoperative flywheel in the test process, ensuring the accuracy of test results. An experimental prototype of the inertial system was developed. The feasibility and accuracy of the developed precise and fractionized inertial system were validated by comparative tests.

Aiming at the tractor semitrailer with liquid hazmat tank, a mathematic model of vehicle rollover is built to reflect its movement response characteristics. Based on the real vehicle parameter, a dynamic simulation model of multibody vehicle is built in virtual environment of ADAMS. Through the dynamic simulation under two typical work conditionlanechanging and swerving, the rollover law is analyzed systematically under the condition of different vehicle speed and different load, the threshold values of the rollover of tractor semitrailer are gotten. The research conclusion is meaningful on the theoretical guide for the early warning of the risk during the running of the tractor semitrailer with liquid hazmat tank.

A new vehicle wheelbase difference detection method was proposed based on the stereovision theory to meet the wheelbase difference detection requirements stipulated by the national standard. A mathematical model was built for the wheelbase difference detection. An identification scheme of the coordinates of the wheel image center was defined on the basis of image processing technology. An experimental system was established to detect the wheelbase difference, the intrinsic and extrinsic parameters  of the dualcamera were calibrated and the systematic errors of the detection system were analyzed. A theoretical model was set up for the 3dimensional reconstruction of the wheel coordinates, the calculation process of the wheel hub center coordinates was discussed. The results of application of the established experimental system to the real vehicle showed that the proposed theoretical model is correct and the experimental system is characterized by high detection precision.

A vehicle dynamics model incorporating mainly the steering system for self-control and the vehicle body model for output of state variables was built using the mechanical dynamics simulation software ADAMS/CAR. A control system for lane-keeping was designed in the environment of MATLAB/Simulink based on the BP neural network and PID control algorithm. The communication between the vehicle dynamics model and the Matlab module was established by input/output interface. The co-simulation of the lane-keeping control was realized by the established Matlab control module and the mechanical system dynamics model. The simulation results showed that the designed lane-keeping control system can achieve the reliable self-control of the lanekeeping,guarantee the vehicle to move along the lane mask smoothly and steadily at dangerous conditions.

The vehicle accumulationrelease mode under the signal control was differentiated through analysis of the start-wave and stop-wave of the vehicle behind the stop line, concluding that the absolutely one-stop queue condition does not exist under steady input. The trajectories of vehicles driving in steady mode were classified and their travel times were investigated respectively. The case calculation showed that the road travel time(or road resistance) is not simply an increasing function of the traffic flow, it is also affected by the signal control. Even if the traffic flow is low, the travel time of the vehicle may be more than that at high flow rate.

In order to provide a theoretical basis for drawing up a reasonable vehicle minimum speed limit(MSL) on freeway standard to ensure its traffic safety and energy saving, the method to set the MSL was studied based on the vehicle speed scattering constraint and considering the vehicle minimum fuel consumption as the target. The concept of the vehicle speed scattering was proposed and its calculation method was given. The model of relationship between the speed scattering and the traffic accident rate was built using the regression analysis method(RAM), and the threshold value of the speed scattering was defined. The constraint condition for the MSL was given from the need of traffic safety,i.e.the MSL should be higher than 62% of the maximum speed limit. Taking the minimum fuel consumption as the target function of setting MSL, the model of relationship between the fuel consumption and the vehicle speed was built using the RAM, the economic speeds under different grades were given. The MSL should be the higher value from the 62% of the maximum speed limit and the economic speed, which on the one hand ensures the vehicle speed distribution as uniform as possible, reducing the probability of road traffic accident, on the other hand is in favor of decreasing the vehicle fuel consumption, providing a support for energy saving of freeway operation from the management side.

A method was proposed based on the improved near-field acoustic holography(NAH) technique to identify the sound source of the in-cabin sound field, to predict the sound field, and to analyze the acoustic contribution under the condition that the sound source of the sound field is unknown. Because the in-cabin noise field is not a free sound field, the prediction method of sound field with reflecting sound was analyzed. The sound source identification, sound field reconstruction and acoustic contribution of the reflecting sound field were studied with case calculations and the satisfactory results were obtained. The results showed that applying the NAH technique to acoustic contribution analysis is obviously advantageous than traditional methods.

The dynamic properties of the fly ash soil improved by randomly distributed polypropylene fiber were studied by the dynamic triaxial tests. The test results demonstrated that the dynamic shear strength of the improved fly ash soil is higher than that of the raw fly ash soil, both of them decline with the increase of the load cycle number and increase along with the surrounding pressure. When the surrounding pressure increases, the strength enhancing rate of the improved soil is higher than that of the raw soil. Under higher surrounding pressure, the improving fiber shows larger effect on the strength of the specimen. The effect of dynamic strain εd on the improved soil is lower than that on the raw soil, indicating more stable property of the improved soil. The dynamic modulus Ed declines with the increase of εd. At the same εd, the Ed of the improved soil appears about 20% higher than that of the raw soil. The damping ratio of the improved soil is a little higher than that of the raw soil.

The affecting factors of the 15 ℃ splitting strength of AC-13 modified asphalt mixture, their significantly order and their correlation with the Marshall performance index were investigated by means of the orthogonal test design and the statistical analysis method. The results showed that the fine aggregate gradation affects significantly the 15 ℃ splitting strength, the significance order of the factors is:2.36 mm passing rate>4.75 mm passing rate>0.075 mm passing rate>OAC>9.5 mm passing rate. The 2.36 passing rate shows highly significant effect on the splitting strength, while the OAC and the 9.5 mm passing rate have no significant effect. The splitting strength is obviously linearly correlated with the Marshall performance index with an absolute correlation coefficient of more than 68%. Along with the gradual increase of VV, the splitting strength of the asphalt mixture decays nearly by 30%.

A viscoelastic 3-D finite element model was built to calculate the horizontal stresses in the surface and bottom layers of typical road asphalt pavement structure, and the effects of loading mode and temperature on their viscoelastic responses were analyzed. The results showed that the surface layer and the bottom layer of the road pavement experience the repeated horizontal tensile and compressive stresses under cyclic loading and small residual tensile and compressive stresses build up in the surface layer and bottom layer respectively. This is probably one of the factors causing top-down or bottom-up fatigue cracking in the asphalt pavement. The residual stresses, the compressive stress amplitude in the surface layer and the tensile stress complitude in the bottom layer of the asphalt pavement of the road all decline with rise of temperature.

The characteristics of the interface between the soil and the geogrids are lack of study. To make a contribution to this respect, applying the uniaxial and biaxial geogrids and the soil weathered from the expansive marlite in Xinxiang section of the middle route of South to North Water Transfer Project(SNWT), a giant laminatedring type shear test apparatus with soil filling of dimensions of 600×600×600 mm was adopted to do a series of pullout tests to investigate the ralations of the average friction versus the relative displacement and the interface strength characteristics under different water contents and dry densities of the expansive soil. An ideal elastoplastic model was used to simulate the interface behavior, the interface strength curves were expressed as the linear functions, and the parameters were analyzed based on the average friction and the average relative displacement of the interface between soil and geogrids. Furthermore, the feedback analysis of the interface parameters was discussed when the progressive yield occured at the interface.

The shrinkage deformations of specimens of concrete added with different amount of silica fume with onesided drying at various depths from drying surface were investigated by fixation of several groups of displacement sensors. The relative humidity changes of concrete at various depths were also studied using embedded humidity sensors. The results showed that the shrinkage and the humidity reduction in the internal layer of concrete are markedly lower than those in the nearsurface layer. adding silica fume to concrete reduces significantly the shrinkage difference from drying surface to internal layer of concrete and the humidity gradient. Under the onesided drying condition, there are good linear relationships between the shrinkage and the humidity at various depths in concrete. However, the shrinkage in the surface layer of concrete is lower than that in the internal layer at the same relative humidity.

A new calculation method was proposed to predict the ultimate load capacity of the reinforced concrete slabs under the compressive membrane effect based on its loading mechanism and the geometrical conditions at ultimate state with the lateral restraint load as intermediate variable. The method is according to the theory of the additional load capacity, considers the load capacity of the compressive membrane effect as the sum of the puve bending load capacity and the additional load capacity by the threehinge arch effect, uses the computeraided programable algorithm to easily get the required results. A series of specimens of the shearwall restrained reinforced concrete slabs were designed with the horizontal support with constant lateral restraint stiffness to restrain the dateral deflection of the specimen. The ultimate load capacities of the specimens obtained by the proposed method were on an average 38.3% higher than those calculated by the yielding line method. Compared with the other calculation methods, the proposed method is easy to use and the test data proved its calculation precision suitable for engineering application.

In this paper, boron alloy steel EN 10 292 TL4225 is selected for the study. The laser spot welding process, one kinds of advanced jointing technology, is used. We get the perfect welded joints of ultra-high strength stamping parts gained by hot stamping process through using the optimum welding parameter. The welding performance of the welded joints, such as the metallography phase is tested, the tension strength is tested, and the microstructure of the fracture visage is analyzed by scanning electron microscope. The test results showed that the welded joints have good mechanical performance by using laser spot welding process, and the mechanical performance meet the actual need of product well.

The thermo-structural and thermo-electrical coupling analyses during the resistance seam welding(RSW) process in the laser beamresistance seam welding(LB-RSW) were performed using the software ANSYS to simulate the effects of RSW current, welding speed and RSW wheel space on the RSW temperature field in mild steel. The results showed that the electrode pressure on the RSW wheel can control the clearance between welded plates effectively, and increase of RSW current, decrease of welding speed and increase of wheel space all lead to increase of the surface temperature of the welded mild steel plates, the temperature gradient, and the temperature change rate. The simulated results are in a good agreement with measering ones by thermal image formation technique, providing a basis for prediction of optimal matching between the laser beam and electric resistance heat sources and the study of mechanism of LB-RSW.

The pincers of manipulator with step-walking beam in the automatic production line of the fracture splitting connecting rod was briefly introduced. A series of dynamics simulations were performed for the pincers jaw using the mechanical system simulation software ADAMS. Based on the simulation results, the pincers jaw structure was optimized, the optimal clearance between the connecting rod and the locating jig was defined, eusuring the pincers meet the basic requirements during the simulation of clamping and transmitting. On the basis of optimization results, the physical test of the pincers jaw for the straight connecting rod was performed and its results showed that it works with stability, reliability and good clamping effect while the jig clearance and the chamfer of the big/small end bores are more than 0.15 mm, 0.5 mm×45°,respectively.

According to the working principle of the piezoelectric pump(PEP), a PEP with an active valve actuated by a rectangular piezoelecltic vibrator(PEV) was developed based on the initial twoendclamping active value design, and its structure and working principle were described. The theoretical flow rate of the developed PEP was calculated and its affecting factors were discussed. The test result indicated that the optimal phaseshifts between the PEVs in the valve chamber and the pump chamber are 70 ° and 250 °. The output characteristics of PEP when the PEV in the valve chamber was exeited by drive signals with different voltages and frequencies were measured. The twoway pumping capacity of PEP with active valve was analyzed comparatively. The measuring results showed that the square wave as control signal for the active valve with the rectangular PEV results in flow rate higher than that by sinusoidal wave. The maximal flow rate is 140 mL/min at frequency of 70 Hz. Under the high exciting frequencies, the PEP with active valve using the rectangular bimorph keeps better flow rate capability than that using the circular PEV, the flow rate of which decreases significantly with rise of frequency.

A dual-pipeline brake system controlled by electronic-hydraulic servo was developed for the heavy-duty vehicle. The steering brake and the emergent brake of the vehicle were realized by a dualpipeline brake system in which an electronichydraulic servo controlled dual brake valve is used to control the dual-chamber brake cylinder. A nonlinear mathematical model was built for the brake system through analyzing the effects of the composing modules and linking pipelines of the brake system on the brake performance. The dual-pipeline steering brake performance of the system was simulated using the software MATLAB/Simulink and the result proved the rationality of dual brake scheme. The test-bed tests showed the performance of the developed brake hydraulic system meets the requirements of design.

In order to meet the requirements of the lunar rover for energy saving and light weight, taking the energy consumption of driving motors as the objective function, the parameters of rocker-bogie suspension were optimized for rover passing obstacle condition. A parametric model was built for the rocker-bogie suspension, the conditions for rover-passing obstacle were analyzed, and a mathematical model was established for the multi-objective optimization of the suspension parameters. The maximum-minimum method was adopted to unify the multi-objective function to a single-objective function. The optimization analysis was performed by the sequential quadratic programing algorithm to obtain the optimal values of the suspension parameters. The results showed that the energy consumptions of motors during rover passing obstacle reduce significantly.

The numerical simulations were performed for the injection process of the common injector needle and two kinds of injector needles with zigzag and wave shape non-smooth surfaces using the non-linear contact finite element method. The histories of the stress, the velocity, the kinetic energy, and the internal energy in the human soft tissue were obtained, demonstrating intuitively the effects of the non-smooth surface during the dynamic process of injection. Comparison among these three kinds of injector needles showed that the bionic non-smooth injector needles are characterized by less drag and more effective pain allviation than common needle. The wave shape non-smooth surface is better than zigzag shape surface, being in agreement with experiment result.

Part or all of the mesh can be eliminated by using the compact support, the orthogonality and other properties of the wavelet function and the meshless algorithm to enhance the computing capability and reduce the error. A meshless algorithm was proposed based on the wavelet basis function. The field function of the meshless algorithm was generated and the corresponding control equation was derived using the Daubechies wavelet as the basis function. Calculation showed that the solution of the proposed algorithm is in excellent agreement with the theoretical solution. Compared with the finite element method, the proposed method is characterized by reduced error and improved precision under the roughly same computing time. The calculation results proved that the wavelet based meshless algorithm effective.

A cubic spline BernoulliEuler beam element was proposed to accurately analyze the large displace ment, large rotation and small strain problems based on the updated Lagrangian formula and the moving coordinate method. The deformation field was built using the cubic spline function, its curvature DoF was eliminated by static condensation method, the geometric nonlinear tangent stiffness matrix of the beam element was derived with consideration of the secondorder effects. The nonlinearity of the axial length change caused by bending deformation was considered, the additional stiffness caused by the bowing effects was derived, and the tangent stiffness matrix of the beam element was amended. The moving coordinate system of the hingedhinged beam element was built on the updated configuration after deformation, and the large displacement total equilibrium equations of the cubic spline beam element were derived. The correctness and efficiency of the proposed element were shown by several typical numerical examples.

The early warning of airport runtime flight delay is a multi-class classification problem. There are two issues when solving this problem using the normal Support Vector Machine (SVM). The first issue is that the prior knowledge is not adequately utilized, and the second issue is intensive time and space consumption for data training. A new algorithm, which is called as center-constrained Minimum Enclosing Ball (MEB) based weighted margin multi-class algorithm is proposed. First, the proposed algorithm uses the prior knowledge to build a new methodology which is based on a new relative affinity function. Then this new methodology is used to calculate the weights of the sample data and add them to the SVM. After adding these features, the SVM is converted to a center-constrained MEB and can be trained easily. Experiments show that the proposed algorithm not only gives more reasonable classification results comparing to normal SVM, but also obviously speeds up the data training processing.

The generation of initial population with heuristic information and the process of coding chromosome was proposed. It presents strategies of adjusting the fitness function dynamically. In the section of genetic operators, the selection of crosspoints was given, as well as the method of amending the crossover operator, mutation operator. In the population evolution, it refers to the iterative backward control algorithm, which is used to solve the problem of the emergency resource deployment by simulation. By comparison with Dijkstra algorithm, the algorithm proposed in the paper can efficiently solve the emergency resource scheduling and provide decisionmakers with strong decision support.

An improved DV-Hop algorithm (RHDV-Hop algorithm) in Wireless Sensor Networks (WSN) is proposed. In this algorithm, the Received Signal Strength Indicator (RSSI) ranging technique is employed to replace the distance measurement of one hop away from the anchor node in DV-Hop algorithm. Also a 2-D Hyperbolic algorithm is used to replace the trilateration algorithm in DV-Hop. Contrasting simulation experiments on the proposed RHDV-Hop algorithm, the DV-Hop algorithm and other improved DV-Hop algorithms reported in the literature are carried out in NS2. Results show that the proposed RHDVHop algorithm outperforms the DV-Hop algorithm and other improved DV-Hop algorithms in localization accuracy.

The current flame detection algorithm can not meet the requirements of high detection rate and low falsealarm rate in engineering application. A multifeature fusion flame detection algorithm based on Choquet integral is proposed. In this algorithm, first the background modeling is carried out by optimizing the Running Average. Then the edge fractal dimension feature, the corner centroid trajectory feature and image correlation feature in the candidate region are detected. Finally, the detected results are fuzzificated and the Choquet integral is calculated to determine the existence of flame. Experiments show that the proposed detection algorithm is efficient and fast with good application prospects.

Aiming at H1N1 gene and starting from the point of RNAi, siRNA prediction is conducted by means of multicharacters analyses, such as sequence and structure. First 46 viral sequence's PA fragments in year 2009 are analyzed. Then the siRNA target genes with strong ability of interference are selected on the basis of the secondary structure of siRNA target sequence. Our research reveals that the outbreaking H1N1 virus in 2009 is characterized by steady heredity, high sequence conservativeness, uniform siRNA target gene sequence, and high conservative structure. This method can be employed to choose the possible siRNA and obtain less but more valuable siRNA target genes by further sifting. It provides a new idea for the design of complex sequence siRNA, and it is of instructional significance for optimal design of siRNA. The research is helpful to the study of H1N1 treatment by RNAi.

A method for extraction of qualitative concepts and construction of concept hierarchy is proposed. First, the cloud transformation is used to extract qualitative concept from database. Then, the former algorithm for pan concept tree generation is modified. When processing the cloud synthesization, the new algorithm takes the influence of the magnitude coefficient into consideration, which makes the output concept hierarchy more close to reality. As a case study, this method is used to extract qualitative concept and construct the concept hierarchy from car price data. Results demonstract that this method can extract qualitative concepts exactly. The concept center extracted by the proposed method is in good agreeement with that obtained by conventional clustering algorithm. The proposed method can also describe the concepts with uncertainty, thus the ontology based on these concepts could give better specification of the conceptualization.

Nonlinear control methods for friction compensation are proposed to improve the precision of position control of ball and plate system. Friction forces are estimated by a nonlinear state observer constructed in an augmented space. A stabilizing position controller with friction compensation is designed using the methods of input/output precise feedback linearization and optimal theory. A position tracking controller, which eliminates the influence of friction, is designed with the theory of variable structure control. Simulation shows that the proposed nonlinear control methods for friction compensation are feasible. Stabilizing control experiments prove that the proposed methods can significantly reduce the position errors in steady state. Tracking control experiments prove that the proposed methods can significantly improve the tracking accuracy.

The quadratic stability conditions for a class of discrete TS fuzzy systems are studied. In these systems uncertain terms exist in the state matrix, input matrix and disturbance input matrix. New stability conditions for fuzzy static output feedback control are proposed. The sufficient Linear Matrix Inequality (LMI) conditions are also proposed, which ensure the existence of fuzzy static output feedback controllers based on the TS fuzzy systems. The sufficient LMI conditions are not only simple, but also consider the interactions among the fuzzy sub systems. Numerical simulation using Iterative Linear Matrix Inequality (ILMI) technique demonstrates the correctness and the effectiveness of the design methodology.

The dynamic uncertainty of space robot caused by the change of base mass or load mass is analyzed. According to the characteristic that the operation load of space robot varies significantly, an adaptive terminal sliding mode controller is designed, and the upper bound of the dynamic uncertainty is learnt using a RBF neural network. The designed controller ensures that the system states stay on the sliding surface all the time. Further, the control errors converge to zero in finite time. Simulation with different loads is carried out on a space robot with two DOF. Results show that the proposed method is robust to dynamic uncertainty.

A head pursuit tracking method is proposed to reduce the speed of mobile robots during the course of tracking moving targets. The tracking law positions the mobile robot ahead of the target and to move along the target trajectory. Thus the robot can capture the target in front of the target trajectory, and the speed of the robot is required to be lower than that of the target. The tracking mathematical model of the mobile robot and the target is derived, and a nonlinear variable structure tracking algorithm is presented based on Lyapunov stability theory. Meanwhile an adaptive algorithm is proposed to estimate the uncertain acceleration of the target. Simulation results verify the tracking model and the effectiveness of the tracking algorithm.

A controller design method for strapdown inertial navigation system in initial alignment was proposed.  State space mathematics model of SINS in initial alignment disturbance compensate was constructed by adopting flight pose deviation，velocity increment，and position increment. Based on Backstepping function control algorithm, virtue feedback control variable was imported and design nonliner controller to draught all states to attractive region of equivalent point collection. According to H2 optimal control theory, optimize choose method for weight matrix was introduced, and the linear optimal control law was obtained finally. Digital simulation and experiment results show that, the controller designed by this method can restrain disturb deviation for SINS in initial alignment quickly, and the proposed control method offer an useful reference to SINS used in engineering.

To overcome the lower data fusion precision, which is caused by the restriction of noise measurement to dissimilarsensor, a novel data fusion method based on weighted least square is proposed. The multisensor measurement model is analyzed. Then the weighted fusion formula is derived for parameter fusion estimation from multiple dissimilarsensors, which is based on the principle of least square; and the scaling coefficients of weighted fusion of various sensors and estimated variances of erroneous vectors are obtained. Coparing with the error of singlesensor and mean estimation, the measurement accuracy is higher for dissimilarsensor data fusion based on weighted least square. Numerical case result is in consistent with theory deduction that further validates the proposed methodology.

A new design method is proposed and used to analyze the characteristics of microstrip discontinuities of a new type of annular branch hybrid. First, the evenand oddmode theory is employed to analyze the ideally simplified circuit model and to obtain the preliminary circuit dimensions. Then, based on these dimensions, the Method of Moments (MoM) to the electric field integral equation together with the full spacedomain Green's function is used to find the currents in the microstrip circuits and the junction scattering parameters. All the physical effects are taken into consideration in this proposed approach. Numerical and experimental results of the new type of annular branch hybrid demonstrated the accuracy and efficiency of the proposed method.

A quasirectangle waveguide structure is designed using organicinorganic hybrid material to meet the needs of the fabrication of electrooptic switches. The optical field distribution on the crosssection of the quasirectangle waveguide structure is simulated and compared with that of the rectangle waveguide structure. Then two linear waveguides, the quasirectangle waveguide structure and rectangle waveguide structure, are fabricated. The output nearfield profiles of the two linear waveguides are captured respectively on a testing system. Comparison shows that the testing results are in good agreement with the simulation results.

A new optimization algorithm for threedimensional (3D) sonar beamforming in the frequencydomain is proposed. The relationship between the beam intensity and azimuth angle is derived based on the resolution of echo displacement vector, and then the algorithm is optimized. Simulation of the algorithm is conducted in Matlab and two datapaths for the proposed algorithm are demonstrated. The memory and the computational requirements for the proposed datapaths are also analyzed. Finally, under the Xilinx ISE environment, simulations for the two datapaths are carried out on FieldProgrammable Gate Arrays (FPGAs) respectively, and the resource utilizations are listed. It is demonstrated that the proposed algorithm possesses the advantage to reduce the memory requirements without sacrificing the system performance.

According to multipath flow optimization mathematical model, a MPLS system model was described. An adaptive flow allocation for LSP based on gradient algorithm was presented. This proposed algorithm can reduce network congestion caused by traditional routing algorithms and optimize network resource utilization. Simulation results indicate that the average packet loss rate is significantly lower than that of the traditional routing algorithms in MPLS networks and this algorithm possesses little running time and high stability.

A five dimensional parameter joint estimation method for near field sources was proposed based on cross array. The method adopts centralsymmetric cross array, each array consists of orthogonal dipolepairs with the same axis direction. The dipolepairs measure the xaxis and yaxis electricfield components of the nearfield signals. Six matrixes pencils can be constructed by receiving polarized array data. The range, elevation, azimuth and polarization parameters of the signal sources can be jointly estimated by generalized eigenvalue of the matrix pencil decomposition. The proposed method does not require any spectral peak search, does not construct highdimensional cumulant matrix, and has simple parameter pairing. It can be applied to additive Gaussian noise environment. The performance of the method is verified by computer simulation.

This paper presents a fast approach for estimation of Direction of Arrival (DOA) of coherent signals on a Uniform Circular Array (UCA). The mode excitation method is applied to transform the output signals of UCA in element space into a virtual Uniform Linear Array (ULA) in mode space. Then a special matrix is constructed using the characteristic of the array, and the DOA of coherent signals can be estimated according to ESPRIT method successfully. Spatial smoothing techniques and spectral peak searching are not needed in this method, so the computation is reduced. The proposed method is validated by computer simulation.

The characteristics of JPEG2000 compression stream are studied and an image distortion model is proposed. By transferring the error rate of JPEG2000 compression stream to the error rate of bit channel coding data, and analyzing the different impact of channel error on JPEG2000 compression stream, the relationship between the image distortion and the error rate of bit channel coding is established. Simulation results show that the distortion value calculated by our proposed algorithm matches well with the actual value, which could be a good reference for estimating image distortion on the receiver.

In coded VerticalBell Laboratory Layered SpaceTime (VBLAST) systems, an enhanced Minimum Mean Square Error (EMMSE) based detection algorithm can significantly mitigate error propagation and improve the system performance. However, under quasistatic fading channel condition, its complexity is much higher than that of conventional VBLAST detection algorithm as it requires matrix inversions for almost every symbol in a data block to be detected. To overcome this shortcoming, a simplified detection algorithm is proposed. This algorithm does not need matrix inversions for all symbols to be detected and can also mitigate the error propagation. The complexity of the proposed algorithm is significantly lower compared with that of EMMSE algorithm and is not obvious higher compared with conventional detection approach. Furthermore, simulation results show that the performance of the proposed algorithm is almost the same as that of EMMSE algorithm.

In order to investigate and develop advanced flight simulator with high fidelity and provide the simulator with actual and reliable navigation data, the data structure requirements of navigation database for the system of flight simulator prototype are presented. These requirements including the airport, runway, very high frequency omnidirectional radio range, instrument landing system, distance measuring equipment system, nondirectional beacon and marker beacon. The complete and authentic navigation data of domestic civil aerodrome is obtained by integrating the navigation database of FlightGear and FS2004 with civil aerodrome navigation chart. The transfer function of navigation data structure is built. Based on Microsoft Visual Studio.NET, the structure transfer of navigation data is implemented; the practicable data file of navigation database for the system of flight simulator prototype is generated; complete and authentic navigation database with reasonable data structure for the system of advanced flight simulator prototype is constructed. The running results of the simulator program indicate that the proposed method for exploiting navigation database is effective and feasible.

A yeast, which can perform biotransformation from xylose to xylitol, was isolated from corncob and was identified as trichosporon coremiiforme by morphological identification and 26S rDNA D1/D2 domain sequence analysis. Besides, the optimal fermentation conditions were determined through single factor test and orthogonal experiment. The optimal conditions are as follows, carbon source concentration: 40g/l, nitrogen source concentration: 6g/l, initial pH: 7.0 and inoculum size: 6%.

An egg is excited by a light mechanical impact on different locations of the eggshell, then the cracks are deteted by dynamical frequency response using flexible piezoelectric film sensors. Analyse show that the dominant resonance frequency can be observed and the frequency value is lower in the intact eggs. While for cracked eggs, there are several peak frequencies, the magnitudes of these frequencies are approximate, and the frequency value is higher. In normalization average of frequency domain, 10 or 20 magnitudes of interval frequency, 10 magnitudes in turn and frequencies by magnitude in turn are extracted as eigenvalues. Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) are used to investigate whether the dynamical frequency response is able to distinguish intact egg from cracked egg. Results show that the frequency response could distinguish the intact egg from the cracked egg using LDA and PCA based on magnitudes in turn and frequencies by magnitude in turn, and higher distinguish rate can be obtained based on frequencies by magnitude in turn. Also higher distinguish rate can be obtained using LDA or based on 10 eigenvalues.

The sensitivity index of the river valley runoff process as dependent variable versus the atmospheric circulation factors as independent variable was defined based on the theory of sensitivity analysis. The coupling mechanism of the atmospheric circulation factors to the runoff process was identified by the BP algorithm to get the mapping relationship between the dependent and independent variables. Based on the historical data of the Nenjiang river valley since the 1950s, the sensitivity index of the valley runoff to the atmospheric circulation index was calculated. It was found that the key factors of atmospheric curculation affecting the Nenjiang river valley runoff process are the area and intensity of circumpolar cyclone, circumfluence at middle latitudes, and Tibetan plateau circumfluence.

Paper designs a kind of ocean seismic data acquisition and recording system based on CPCI bus and twostaged distribution architecture. The system includes underwater data sampling and communication subsystem and on shipboard recording subsystem. The main characteristic is single hydrophone, 3.125m channel distance and 24 bit ADC. Its dynamic range achieves 115 dB, THD -106 dB. It supports 1920 channels per cable，1ms sampling interval，connecting 4 streamers at one time. The system has been tested on the sea，it could collect seismic data，plot a profile graph，achieved all technical target. This is the first near ten thousand channels ocean seismic survey equipment which China owns the intellectual property.