The cooling circuits of Electric-vehicles (EVs) are designed according to the temperature characteristics of all parts. The heating principle of the brushless permanent magnet synchronous electric motor is analyzed. The theoretical heat transfer equation from generation to dissipation is derived based on the equivalent heat capacity of the parts. The minimum power consumption is taken as the evaluation index to search for the optimum cooling temperature under certain environment temperature. Based on this the optimum cooling temperature control strategy is built. Compared with pump fixed-displacement control, the vehicle efficiency with the proposed optimum cooling temperature control strategy is improved from 60.5% to 66.1%.

In order to reduce the development period and cost of hybrid electric vehicle, and improve the efficiency of development, a more readable Matlab/Simulink based control strategy architecture for hybrid electric vehicle is proposed. The architecture is suitable for different simulation environments, different configurations and different vehicle types. The ideals of standardization, modularization and stratification are applied to build the architecture. It is verified that compared with the existing architecture, the new proposed architecture is more readable, extendable, universal, portable and inheritable.

Using software ADINA, accurate finite element mesh models of the superposition throttle slice and fluid field are established based on the Fluid-Structure Interaction (FSI) method. The velocity property, the indicator diagram, internal pressure field and velocity field of the shock absorber are obtained using the post-processing module of ADINA. The nonlinear dynamic characteristics are analyzed. Results show that, under high-speed impact of the shock absorber oil, the pressure field inside the valves changes significantly. When reaching the valve-opening pressure, the fluid velocity suddenly jumps, making the pressure change rapidly in the damping channel. The finite element mesh models of the superposition throttle slice and fluid field significantly influence the accuracy of the simulation results. The FSI model is close to the physical reality because the consideration of the turbulent flow of oil, the contact slip and large deflection of the throttle slices. The simulation results are in good agreement with the experimental results.

On the basis of Thermoelastic Instability (TEI) theory, a theoretical model is proposed to investigate the local hot zones and influence factors of wet clutches during engaging process. The effects of the surface roughness and the mixed lubrication on the temperature and pressure of the friction pair are considered. Under mixed lubrication, the ratios of the lubrication area to the contact area and the fluid shear velocity induced heat flux to the frictional heat flux are calculated. A plasmon resonance analyzer is used to investigate the variation of the surface roughness in different stages. The results show that strong inhomogenous temperature field, hot zones and cold zones appear when the clutch engaging speed exceeds the threshold. The material properties, such as thermal conductivity and elasticity, and structural properties, such as the thickness of the matting disk, significantly influence the stability of the system.

A Finite Element Model (FEM) of a homemade car is established, and according to the side impact and Noise-Vibration-Harshness (NVH) requirement, Multidisciplinary Design Optimization (MDO) is carried out for the Body-in-White (BIW) of this car. First, on the basis of the simulation results of side impact and NVH, the thickness and materials of several parts of BIW are selected as the design variables; the mass and stiffness of BIW are taken the objectives; and the constraints include the velocity and intrusion of side impact. Then, a Radial basis Function (RBF) approximate model for stiffness and a stepwise-Response Surface Model (RSM) for side impact are established. Finally, multi-objective optimization is performed using Adaptive Simulated Annealing (ASA) algorithm. Results show that the mass of BIW can be reduced by 15.7 kg while keeping its performance of crashworthiness and NVH.

In order to identify different types of driver individualities, a driver behavior signal acquisition system was developed using the dSPACE real-time simulation platform. The signals of driving behaviors of 30 drivers were collected under the test condition of double lane change. An identification model of driver individualities was proposed using Gaussian mixture model. Three kinds of typical standard drivers were chosen to optimize the model parameters. Identification experiments were carried out with testing drivers using the optimized model, and the Design of Experiment (DOE) method was applied to analyze the identification method. Experiment results show that the proposed identification method based on Gaussian mixture model can effectively identify the drivers' individualities.

Computational Fluid Dynamics (CFD) simulation with overset grid is carried out to investigate the transient aerodynamic characteristics of the flow field around a simplified model of straight driving vehicle. Wind tunnel test is conducted to verify the CFD simulation results. Based on this, CFD simulation of the flow field around a turning vehicle is carried out. The aerodynamic drag coefficient and velocity vectors are obtained. The results are analyzed in contrast with the results of straight driving vehicle. The turning vehicle experiences a transient side force and a yawing moment, both of which sharply increase as the turning radius decreases. Both the side force and the yawing moment change with the vehicle speed, which would affect the vehicle stability. This study provides reference for further study of the transient aerodynamic characteristics of turning vehicle.

In the current status of the design practice, scientists/engineers do not know the specific process of emotional awareness and only conduct research using empirical validation. To improve such status, the significance of the methods, which use a model to summarize the process of psychological cognition of product modeling, is analyzed. The projection principle of the autobody styling cognition-intention is proposed. A projection evaluation model of the autobody surface modeling cognition-intention is established. A case study is carried out to demonstrate the application of the model in the design practice.

The existing driver model less considers the driver characteristic factors. To overcome this shortcoming, the expressway weaving area is chosen as the research object to build a gap acceptance model. First, a large number of vehicle trajectory data is generated using the video processing software VEVID. Second, a continuous driver type integration model is proposed using fuzzy clustering theory and K-S test method. Third, the mutual cooperation mechanism is introduced into the lane change process of the expressway weaving area, and the influence of driver characteristics on the driving behavior is considered to build the gap acceptance model. Finally, the microscopic traffic simulation system MTSS is taken as the simulation platform to build a merging simulation model to validate the gap acceptance model. Results show that the error between the simulation value and real measured value is less than 10%, which indicates that the proposed gap acceptance model can be used to describe the impact of driver type on driving behavior.

For un-dedicated left-turn phase intersections, the pedestrians' crossing characteristics were analyzed under the influence of left-turn vehicles at the crosswalk. The spatiotemporal information of the pedestrians crossing the road is obtained by video image processing technology; then their crossing trajectory and speed are analyzed. The typical pedestrians' behaviors are classified. Under different interference degree of left-turn vehicle and green light time, the pedestrians' characteristics are analyzed. The pedestrians' delay at the crosswalk due the influence of left-turn vehicle is statistically counted. The results may provide reference for signal timing and left-turn vehicle phase setting at the intersections.

In order to further improve the online analysis efficiency of traffic operating reliability of road network, a concept of reliability matrix of road network is proposed, which is based on the adjacent matrix method of graph theory. The aims to propose this reliability matrix are to reduce the information storage space and accelerate the running speed of the algorithm. On this basis, a new online analysis algorithm of traffic operating reliability is designed based on factor theorem. The algorithm is validated and comparatively analyzed using simulation data. The results show that the proposed algorithm can ensure the analysis effect of traffic operating reliability; meanwhile, it significantly reduces the running time by about 92.7% in average.

To fully utilize the time-space resources of an intersection with reversible approach lanes, an optimization model was developed for pretimed signals. The objective of this model is to minimize the total delay for all vehicles. The decision variables are the function of each reversible approach lane and the effective green time per signal phase. This model was validated by a case using the traffic simulation software VISSIM.The results indicate that there is an optimal combination of the two variables by which the average delay per vehicle for each movement, the average queue length and the maximum queue length on each approach can be minimized. To further consider the multi-period signal control, the procedure of the multi-period pretimed signal timing design for an intersection with reversible approach lanes was proposed. The research achievement can reasonably allocate the function of each reversible approach lane and the signal timing scheme during any time period for a signalized intersection, effectively improving the operation efficiency of traffic flow and providing a theoretical basis and method for the control of reversible approach lanes.

In order to evaluate the impact of multi-modal guidance information on the traffic assignment in the integrated transportation network, a trip-chain-based probit stochastic user equilibrium model was proposed to study the variations of travelers' choice of combined trip mode, trip-chain cost and trip-chain structure under the multi-modal guidance. A MSA algorithm with Monte-Carlo method embedded was presented to solve the model. In the model, travelers were divided into two groups by accepting multi-modal guidance information or not, and both groups make choices according to their own expected cost of trip chains. In the course of solution, hyper-network theory was introduced, and traffic flow assignment was carried out on the hyper-network. The results show that multi-modal guidance information helps to increase the mode split rate of public transport by 5.63%, and the network benefit reach 196.254. It can be confirmed that multi-modal guidance information contributes to encourage travelers to decrease the use of car, and reduce trip-chain cost and simplify trip-chain structure.

A multi-mode choice model is proposed to effectively investigate the impact and feasibility of charging taxis with link-based toll in the pricing zone in designing congestion pricing scheme. The model is developed as a bi-level programming problem. The lower-level is formulated as a variational inequality, with the logit-based mode split, route choice, elastic demand and vacant taxi distributions. The upper-level is for social welfare maximization. An efficient algorithm is implemented to solve this specific problem. The numerical experiment indicates that whether to charge the taxis relies on the operation cost ratio of private car to taxi. The threshold of the ratio is 5.6 (which can hardly be reached in real world) in the numerical experiment based on realistic data settings. If the ratio is lower than the threshold, charging taxis always outweighs exempting them from the toll in terms of the maximum social welfare.

An online vehicle classification method based on decision tree model is proposed, in which multi-functional magnetic sensors are used to collect field magnetic data. First, eight speed-independent time-domain waveform features are extracted as the inputs of the decision tree model. Then the decision tree model is trained based on the Classification and Regression Tree (CART) algorithm with the Minimum Number of Slit-samples (MNS). Finally the trained decision tree model is pruned with a Minimum Error Pruning (MEP) role to obtain an optimal pruning tree, which is more robust to new samples. For the field samples collected from a road in Beijing with two types of vehicles, the average classification accuracies of the forward and reverse tests are 88.9% and 94.4% respectively. The proposed classification method is compared with existing methods. The results show that the proposed method enables online vehicle type classification with the advantages of high classification accuracy, sample robustness and less algorithm execution time.

Braking deceleration accuracy and velocity accuracy are two important indexes to measure the brake system performance. However, in real braking, the deceleration cannot be accurately measured because that it is easily affected by the line conditions and brake shoe coefficient. Conventional closed-loop control method cannot deal with the phenomena of input delay and nonlinearity in the system. According to the operation principle of the electro-pneumatic brake system, a nonlinear brake model with input delay is built. To deal with the nonlinearity, input delay and bounded disturbances in the system, a robust brake controller based on backstepping technique and Krasovskii functional operators is designed. By Lyapunov-Krasovskii method, the close-loop system is proved to be Input-State-Stable (ISS) with respect to bounded disturbances. To validate correctness of the robust controller, the brake model is simulated on Matlab. Results demonstrate that the proposed model not only improves the brake system performance, but also is robust to disturbances and input delay in the brake system.

Experiments are conducted to investigate the damage feature, quality degradation and dynamic elastic modulus decay of concrete under multi-salt soaking, freezing-thawing and dry-wet cycles. Furthermore, the corrosion mechanism of concrete is revealed by electron microscope scanning, analysis of EDX spectrum and chemical composition analysis of the concrete microstructure. Results show that under the common action of multi-salt soaking, freezing-thawing and dry-wet cycles, quality degradation and dynamic elastic modulus decay tend to rise with the increase in the test cycles. The corrosion damage of concrete is mainly caused by the expansion effect of salt crystallization. Combined with the salt chemical erosion, the damage becomes faster and more serious. The total corrosion damage is the result of accumulation of damages caused by a variety of harmful factors.

Using Maple software as the platform, a method is proposed to calculate the internal force of flexible pier by solving the differential equation set. In this method, first, each bridge pier is considered as a whole, and the differential relationship of each calculation unit of the pier is established according to the coordinating deformation of each node and the force equilibrium condition. Then, all the bridge piers are considered as a whole, and the differential relationships among the piers are established according to the lateral displacement produced by shrinkage, the creep of concrete and the temperature influence on each node. Finally, by simultaneously solving the differential equation set, the internal force functional analytical formula of each pier was obtained. Compared to the currently used integrated stiffness method, the proposed calculation method has the advantages of simple and convenient, and the result is more precise and reliable.

In order to analyze the combustion performance of a spark ignited free-piston gasoline engine for linear alternator, according to the piston motion characteristics, a coupled multidimensional computational fluid dynamics model is used to simulate the combustion process. The effect of free-piston motion on the combustion process is analyzed in comparison with a conventional crankshaft-driven gasoline engine. The results show that, compared with the conventional engine, the peak in-cylinder pressure and the mean temperature of the free-piston linear alternator are lower, and the period of the peak in-cylinder pressure is shorter, which benefits the reduction of emission of NO exhaust. The heat release rate of the free-piston linear alternator has the characteristics of slow heat release at prophase, gentle heat release at metaphase and long combustion at anaphase. Less gas mixture burns before the piston reaches the top dead centre. The loss work in compression stroke is also less than that of the conventional engine. However, the indicated efficiency of the free-piston linear alternator is lower because of its serious after-combustion period.

In the search of parallel hybrid hydraulic excavator power matching, a power prediction model is proposed based on Least Square Support Vector Machine (LS-SVM) algorithm for constant power variable duplex pump. Based on the analysis of the principle of variable capacity pump, the load spectra obtained from actual excavator operations are used as training samples, the pump outlet pressure, the pilot secondary pressure and the variable displacement pressure are taken as the input, and the pump displacement is taken as the output to build the model. Such the model has better adaptability to the working environment. The LS-SVM parameters are optimized with grid-search and cross-validation methods. Experiment results show that the proposed model has good prediction accuracy and generalization capacity.

Based on the structure of kind of V type engine, a design method for the timing chain system was established. The overall layout, the chain, sprocket, guide rail, tensioner of the timing chain system were designed and the length of the chain system was calculated. A dynamics analysis model of the system was proposed. The trajectory and the chain tension force, the contact force between the chain and the transmission parts, and the tensioner's tension were analyzed. The results demonstrate that the proposed design method is reliable, and the dynamic analysis results can meet the requirement of the design. This study provides a reference for timing chain system design.

In this paper, a novel approach for analysis of the kinematics of lower limb joint based on non-intrusive virtual sensors is presented. With the signals obtained from a wearable sensor system, the method, which combines a physical-sensor-difference-based algorithm and virtual-sensor-difference-based algorithm, is used to calculate the rotational angles of the virtual sensors supposed to be fixed on the knee joint. Then the data is used to analyze the knee joint kinematics in anatomical coordinate system. To test the approach, a wearable sensor system composed of accelerometers and magnetometers is developed and evaluated on the lower limbs of five subjects. The results show that, without intrusive sensors, the proposed approach is suitable for detecting and analyzing the parameters of the kinematics of lower limb joint with wearable sensor system. After the measuring accuracy is improved, the approach can be applied for gait analysis.

Theoretical calculations are carried out based on a vertical line source 3D abnormal potential of finite difference to verify the effectiveness of potential gradient in the detection of water injection direction. Then, a low-noise, high-resolution oil borehole-surface potential gradient detection system is designed and developed based on virtual instrument technology. The system applies a differential input. The short circuit noise amplitude is lower than -130 dB. The A/D acquisition, which adopts the expansion slot design ideas, realizes the synchronization of multi-channel data acquisition. The upper layer software consists of Matlab and Labview programming techniques, thus the extraction of weak signals and real-time imaging of detection results can be achieved. Experiments show that the potential gradient detection system has higher resolution and good stability compared to the potential detection system. By processing and analyzing the experimental data obtained from the injection wells in Daqing oilfield, it can locate the cutting-edge of water drive. Furthermore, the injection well data prove the accuracy and effectiveness of the system in detecting the water injection direction.

The stability of brightness of the O^- beam generated from the douplasmatron ion source is influenced by the gas pressure, the magnetic field strength and the arc current. An ion optical system for measuring the brightness of a duoplasmatron of the SRHIMP II was presented. The brightness of the O^- beams as a function of the gas pressure, the magnetic field strength and the arc current was experimentally studied. The results show that longtime, stable operation of the ion source can be realized at the pressure range from 110 mTorr to 170 mTorr and an arc current above 50 mA. The maximum brightness measurements give 52.4 A/(cm²·sr) for O^- with 0.25 T magnetic field strength, a low pressure of 140mTorr ere and an arc current of 200 mA. Controlling these parameters reasonably can increase the brightness of O^-, improve the lateral resolution and the sensitivity of the secondary ion mass spectrometry.

Experiments were performed to investigate the sinkage of wire mesh wheel under light load (in the range of 20 N to 60 N), which mimic the low gravity environment on lunar surface. The wheel load and wheel speed are taken as the experimental factors, the wheel sinkage is taken as the experimental index, and orthogonal experiment design method is adopted. The experiments were carried out on a single-wheel test bed for the wire wheels under three simulant regolith conditions. The actual wheel sinkages were collected, and the apparent sinkages were measured using a rut morphology laser scanning device. Using regression analysis, three linear regression equations were worked out with confidence degree of 0.99 and square correlation coefficients all above 0.9. The relative error between the actual wheel sinkage and the calculated sinkage by the regression equations ranged from 0.34% to 8.24%. The variation trend of the laser scanned apparent sinkage was in consistent with the actual sinkage. The linear relationship between the sinkage and the wheel load was obvious. This study verified that the actual sinkage can be predicted from the apparent sinkage.

Based on biomimetic principles, the ridge structure of shell surface of chlamys farreri, which has excellent wear-resistant property, was employed to design the wear surface of subsoiler tines. Four subsoiler tine samples were designed and manufactured with 65Mn and T10. Abrasive wear experiments were conducted using JMM type abrasive wear tester. The results show that the wear losses of three bionic rib type samples are less than those of the flat samples obviously. With 65Mn, the wear losses of the bionic samples are reduced by 17% to 30% compared with flat samples. With T10, the wear losses of bionic samples are reduced by 13% to 24% compared with flat samples. The sliding velocity and distribution space of ribs have significant impacts on the wear-resistance of the samples. The wear-resistance of T10 samples is superior to that of 65Mn. The results show that the bionic ridge structure has remarkable effect in improving wear-resistance of sobsoiler tine.

The effects of laser power and welding current on the appearance of weld, the microstructures and mechanical properties of welded joint were studied by CLSM, SEM, microhardness tester and universal tensile tester. The results show that laser power and welding current greatly promote the weld penetration. Laser power has no obvious influence on the weld width except that the weld is "bread twist" like when the laser power is relatively low. The multi-layer hybrid weld is goblet-shaped and there are a small amount of pores within the weld metal. The formation of the pores has nothing to do with the gasification of element Mn in the base metal and filler metal. Obvious necking emerges in the base metal zone where fracture happens during the tensile test. Furthermore, the fracture appearance has a typical characteristic of plastic deformation.

Axial compressive failure test of composite cylindrical shell was carried out. The failure load and the load-strain curves were collected. The results show that the damage form of the cylindrical shell under axial compressive load is buckling failure. Software ANSYS was used to simulate the buckling process of cylindrical shell. The simulation results agree with the test results, which verify the finite model and the computation method. Based on the model, modified design of ply stacking-sequence for the cylindrical shell was carried out, and the influence of fiber ply angle on buckling load of the shell with opening was analyzed. A composite cover was installed to reinforce the composite cylindrical shell, and the buckling load of the cylindrical shell with cover under different ply methods was calculated. Results show that optimization of ply stacking-sequence can enhance the buckling load of the shell. The buckling load of the cylindrical shell with opening decreased significantly; while installation of the cover can reinforce the cylindrical shell effectively. A method to improve the axial buckling strength of composite cylindrical shell with cover was proposed, which can provide a reference to stability design of composite shells.

The coupling relationship between the manipulator and the hydraulic servo system is constructed by a driven Jacobian matrix. Then, the entire system dynamic model is established. Considering the unmodeled dynamic and external disturbance, the Adaptive Second Order Sliding Mode Control (ASOSMC), which can guarantee the tracking performance in finite time and strong robustness, is obtained by defining virtual control law and selecting suitable Lyapunov function. Simulation results show the effectiveness of the proposed control scheme, which can reduce the chattering phenomenon in traditional sliding mode which is harmful to hardware.

In order to further improve the isolation degree of disturbance of traditional controller, a new active disturbance-rejection controller based on frequency adaptation of disturbance is proposed. It is verified in the linear inverted pendulum system. The results show that the controller can avoid phase lag when extended state observer observes the disturbance with the high-order form. It can also at least reduce the disturbance error by 7.85 dB compared with the state feedback controller when the perturbation frequency is between 0 Hz and 2 Hz. The proposed controller has strong robustness, allowing the range of system parameters to change within 15%.

Supercharged boiler has the characteristics of strong nonlinearity, big time delay, parameter time-variation and uncertainty, which make it difficult to obtain an ideal control performance using traditional control methods. Since Laguerre filter networks can approximate system linearity with a lower model order without knowing the system delay, while Fuzzy model can describe system nonlinearity and uncertainty effectively, a Nonlinear Model Predictive Controller (NMPC) based on Laguerre-Fuzzy Hammerstein model is proposed and applied in the controller design of supercharged boiler combustion system. Simulation results show that the proposed controller can adjust the main-steam pressure to set-point fast under varying load. Compared to the predictive controller with one input variable, the proposed controller comprehensively considers the influence of the two nonlinear variables on the system, and avoids the influence of the varying time delay and uncertainty, thus validating its effectiveness.

In order to suppress the injection of three-phase Electromagnetic Interference (EMI) of switching power supply into the power grid, a design method of three-phase EMI filter is proposed. By modeling and analyzing the three-phase common-mode inductance, the ideal equivalent circuits of common-mode and differential-mode filters are built. The effect of high frequency on ideal equivalent circuits is analyzed. The insertion loss expression based on parameter matrix is determined. Parameter matrixes of common-mode and differential-mode filters are derived under ideal and high frequency conditions. According to the characteristics of load impedance in the power grid side and source impedance in the switching power supply side, the relationship between insertion loss and frequency is analyzed. Because that high conductivity of soft magnetic ferrite has gyromagnetic frequency limit and nonlinear hysteresis loop, their inhibitory effect on common-mode inductance and their influence on common-mode insertion loss are analyzed. The filter's parameters are determined based on the related suppress standards and requirements. Experiment results show that using three-phase EMI filter the noise peak of the switching power supply can be suppressed by about 21dB.

In model based diagnosis of discrete event system, an incremental method is proposed to determine the diagnosability and improve the efficiency of determining diagnosability. By reversely spreading the faulty labels on states, the pre-diagnoser is built. Then based on the diagnoser and the virtual online observation windows, the diagnosability is determined incrementally, and it is decided whether to prune the current state and return the result of diagnosability. An incremental algorithm is proposed and its correctness is proved. The experiment not omnly verifies the efficiency of the incremental algorithm but also tests the relationship between the size of the observation windows and the efficiency of determining diagnosability.

A new algorithm is proposed, which combines arc consistency algorithm and max-restricted path consistency algorithm. The new algorithm makes full use of the low computation cost of arc consistency and the strong ability to remove invalid values of the max-restricted path consistency, thus reducing the visited nodes when coping with unsatisfied problems. As a result, the efficiency can be improved. Experiment results show that the proposed max-restricted path consistency algorithm based on the half range of domain outperforms the traditional maintaining arc consistency algorithm by large margin when solving the unsatisfied problems.

The techniques of statistics, frequent item mining, pattern matching, heuristic learning and active learning are employed to learn the concepts (including instances), taxonomic relations, semantic relations and the concept properties from the documents based on preprocessing tool Gate and general ontology WordNet. The concept property learning was first proposed in this paper. Experiment results show that the proposed ontology learning method can improve the effect of word semantic disambiguation, enrich phrase concept learning and semantic relationship learning, increase the accuracy of automatic ontology construction and reduce the cost of ontology learning.

Lazy slicing is an effective state-space reduction method, but it can not directly determine whether a model satisfies the expected linear property. A lazy slicing based method is proposed to solve this problem. First, it constructs a negation Büchi automaton for given linear temporal logic formula and a product automaton of the system model. Then it searches for an acceptable trace on this product automaton lazily. Thus, the problem of verifying a linear time property is converted into the problem of invariance validation, which makes it possible to search the acceptable state by reachability analysis. Experiment results show that this lazy slicing based algorithm not only makes lazy slicing be able to verify linear time property, but also remarkably improves the scalability of LTL model checking.

To capture the impact of concurrency on query performance we present an experiment-driven modeling approach and propose an interaction-aware scheduling strategy to achieve better performance by reducing the negative interactions. Experiment results demonstrate that, compared with other methods, the proposed scheduling strategy can effectively improve the database performance.

Dynamic Voltage Scaling (DVS) is an effective low-power technology. In this paper, a low-power scheduling algorithm is proposed for mixed task in real-time system. In this algorithm the mixed task set with both periodic tasks and aperiodic tasks is considered using the DVS technology. The algorithm is composed of two phases: in the first phase the static speed is computed off-line; in the second phase the running speed of the task is adjusted based on the slack time. Simulation results show that the proposed algorithm can reduce the average energy consumption by 27.35% over the existing mix task DVS algorithms.

The service deployment with multi-service level agreement is a basic problem to be solved for the service based software system. To overcome the shortcomings of the existing algorithms, a multi-objective model for this problem is built and a Simulation-based Multi-Objective Genetic Optimization Algorithm (SMOGA) is proposed. According to the characteristics of this problem, the genetic operations are redefined, including the introduction of individual domination strength into the environment selection, and the redefinition of crossover strategy and individual mutation strategy incorporated with local search. The performance of the proposed SMOGA was analyzed and compared with that of recently proposed algorithms E3-MOGA and NSGA-II for this problem on different scale test cases. Experiment results show that the proposed SMOGA can solve this problem more effectively.

Existing algorithms for hot route detection are difficult to solve the complex coupled problem of hot routes, or they need the support of road network topologies. In order to overcome these disadvantages, we present a hot route detection algorithm based on grid clustering. In this algorithm the trajectory is converted to grid sequence, and the density reachability of the neighbor grids is determined based on their common traffic, and then the grids are abstracted to a graph model. So the grid clustering algorithm, GridGrowth, can be presented based on the graph theory, i.e. the hot route detection algorithm. Experimental results show that the proposed algorithm can effectively detect the hot routes and can accurately solve the complex coupled problem of the hot routes.

The classic grey theory does not adequately take into account the distribution of data set, and lacks effective methods to analyze and mine large sample in multi-granularity. Considering the universality of normal distribution, a normality grey number is proposed. Moreover, the corresponding definition and calculation method of the incidence degree between the normality grey numbers are constructed. On this basis, the grey incidence analysis method in multi-granularity is put forward to realize the automatic clustering in the specified granularity without any experience knowledge. Experiments fully demonstrate that the proposed method is effective in knowledge acquisition for large data.

Considering the difficulty of phase synchronization for Continuous Phase Modulated (CPM) signals, a general Soft Carrier Phase Synchronization (SCPS) algorithm for different signals is proposed. Using Rimoldi decomposition of CPM signals, the SCPS algorithm is derived based on maximum likelihood criterion and the Expectation-Maximization (EM) algorithm. The synchronizer and demodulator are iterated jointly. The soft demodulation information provided by CPM soft demodulator is used to aid the carrier phase estimation. Simulation results indicate that the Root Mean Square Estimation Error (RMSEE) of the SCPS algorithm is very close to that of the Modified Cramer-Rao Bound (MCRB) for general CPM signals, and the bit error rate could achieve approximate ideal performance.

A Load and Interference-Aware Transmission Time (LIATT) routing metric is proposed. It inherits the advantage of describing intra-flow and inter-flow interference uniformly. The buffer queue length of node is used to capture traffic load, and the average load of adjacent areas is used to capture interference, thus, achieving the load-balancing function and interference-aware function, which can avoid routing packets into heavy load and intense interference area. Simulation with NS-2 software is carried out to evaluate the performance of the proposed LIATT routing metric. Results show that the LIATT routing metric can effectively achieve load balancing and improve the whole network performance.

First, spatio-temporal interest points containing enough human motion information are detected, and a set of spatio-temporal interest points are selected based on the information of connectivity of human silhouettes. Then, the GMM clustering is performed on the points in the training set and the spatial-temporal words are generated. Finally, these spatial-temporal words are trained to obtain the GMM of each behavior for human behavior recognition. This algorithm can be applied both to single behavior recognition and to interactive behavior recognition. Experiment results on activity database show that this approach has a satisfactory identification rate of human activities.

When the visual perception effect of the reconstructed stereoscopic image is good, the Peak Signal to Noise Ratio (PSNR) value may be low. To solve this problem a method is proposed to improve the PSNR in order that it is consistent with the perception of human eyes. This method is based on weighted filter, membership function and sliding window. Experimental results show that the proposed method has good effect, which can promote the universality of the commonly used PSNR evaluation method in the quality evaluation of reconstructed stereoscopic image.

Considering three dimensional target tracking systems with unknown systematic error, in order to obtain the joint estimation of target state and sensor systematic error, an augmented target tracking algorithm based on Square-root Cubature Kalman Filter (SRCKF) is proposed. The performance of proposed estimation method is analyzed with a numerical example taking account the root mean square error and average computational cost. Simulation results show that the effectiveness of the proposed algorithm is higher than that of the algorithm based on extended Kalman filter in the aspects of estimation accuracy and filtering stability.

The BCS-SPL algorithm, which combines the Wiener filter with PL compressive sensing signal recover algorithm, can overcome the blocking artifacts but its convergence speed is slow. To overcome the slow convergence speed of the BCS-SPL algorithm, especially on small measure rate, the BCS-SPL algorithm is combined with smooth L0 norm compressive sensing signal recovery algorithm, and a BCS-SSL0PL algorithm is proposed, which can speed up the image compressive sensing. The performance of the BCS-SSL0PL algorithm was compared with the BCS-SPL algorithm in terms of recover image quality, recover time and iteration times. Results show that the proposed BCS-SSL0PL algorithm surpasses the BCS-SPL algorithm that it can achieve the same recover quality with less time; it needs nearly the same time for different measure rate conditions; furthermore, it can improve the blocking artifacts on small measure rate conditions.

Spurious Free Dynamic Range (SFDR) measurement of Analog to Digital Converter (ADC) is restricted to the specified ADC parameters and matching test methods. In this paper, first, a theoretical analysis of approximate upper bound performance of ideal ADC with different quantization bits is carried out based on spectral analysis of quantization noise. Then, the relationship between the SFDR performance and the SNR of input signal under Additive White Gaussian Noise (AWGN) is illustrated. The effect of sampling on SFDR is also considered that the characteristic of discrete spectrum of ADC output and its influence on SFDR with different sampling rate is investigated. The mathematical deduction results are verified with FFT analysis test method.

To overcome the complexity in prototype filter design induced by the increase in the number of channels and the requirement of narrow transition band, a new design method of channel filter banks with narrow transition zone is proposed. Using Frequency Response Masking (FRM) techniques, the prototype half-band filter is interpolated to obtain a narrow transition band. First, the prototype masking filter is shifted in spectrum to generate masking filters for upper and lower branches. Then, the narrow transition band filter is obtained by synthesizing the upper and lower branches. Finally, the filter banks with narrow transition band are attained by filter modulation, poly-phase filtering and combination of adjacent bands. A 8-channel design examples with Matlab demonstrate effectiveness of the proposed method.