This paper introduces the basic principle of the vehicle-mounted doubly-fed machine, and analyzes the energy flow direction of the machine in different operating states. In order to reduce the number of distributed drive electric vehicle inverters, a new distribution pattern is proposed that the stators share the same inverter, which simplifies the structure of drive system effectively. The mathematical model of the doubly-fed machine is established, then by analyzing the influence factors of the active power and the reactive power of the rotor side and the stator side, the control strategy, which minimizes the total capacity of the inverter, is put forward. Matlab/Simulink is applied to establish the driving system model, and the validity of the proposed control strategy is verified, so as to reduce the production cost and the occupation space.

The control objective for Extended Range Electric Vehicle (EREV) is to control battery State of Charge (SOC) to reach the lower bound as the vehicle arriving at charging station. First, the minimum fuel consumption and SOC varying along target curve were added in cost-function, then, Dynamic Programming (DP) algorithm was adopted to solve the optimization problem. To achieve real-time control performance, a neural network framework was applied to train the optimal results and generate control model. Simulation results show that the generated model has better learning performance that the difference error is within 3%. Since the control model is only fit for special driving cycle and driving distance, an adaptive energy management controller was proposed. It consists driving pattern recognition module, control model library and control model selection module, etc. Hardware-in-Loop test was conducted to validate the controller performance. The results show that the proposed controller can control battery SOC just dropped to lower bound at the end of driving cycle, and it improves fuel economy for 9.2% compared to CD/CS algorithm under HWFET driving cycle.

Engine as the power source of refrigeration system during the night/midway parking causes large energy consumption and much exhaust emission. To solve this problem, a refrigeration system supplied by solar auxiliary power of refrigerator car was proposed. The dynamic mathematical model was established and matching methods of the system were proposed. The effectiveness of the model was verified by experiments and the cooling performance of the solar auxiliary power was analyzed. The results show that the power generation of the system is able to supply the refrigerator for continuous operating for 4.45 h in a day. Temperature in the compartment needs about 580 s to drop from ambient temperature to the specified value at three typical driving speeds and then remains stable. Compared with the refrigeration effect of traditional refrigerator car, the system has remarkable energy-saving effect and increases the delivery range of the refrigerator car.

In order to provide reference and technical guidance for the practical application of adhesive technology in steel-aluminum body, the influence of hygrothermal environment on the performance of adhesive joints under different stress states is studied. The Single Lap Joints (SLJs), 45°Scarf Joints (SJs) and Butt Joints (BJs) are selected as the research objects. After the adhesion and curing of specimens, experiments of 20, 40 and 60 hygrothermal cycles are completed according to hygrothermal cycle spectrum. The results show that with the increase in hygrothermal cycles, the failure loads of three kinds of joint all decrease, and the SJs have the biggest decrease amplitude, namely the failure load decreases significantly under the combine stress of shear and normal. At the same time, the failure modes of the joints have also changed after the hygrothermal cycles. The normal stress and shear stress of the adhesive layers are calculated and the curves are fitted. The stress criterion is selected as appropriate expression of joint failure criterion after different hygrothermal cycles. Based on this, the failure criterion formula and 3D surface with different aging cycles of the joints are established.

This paper employs homogeneous coordinate matrix and homogeneous transfer matrix to establish inverse position solution model of the three 6-DOF motion platforms. Then the obstacles around the test bench and two lower motion platforms and vertical actuators of motion platforms are simplified into hexahedral model using the bounding box method. According to spatial kinematics and spatial convex polyhedron theory, the collision detection model between vertical actuators and lower motion platforms with obstacles around the test bench was constructed. The shortest distance of two objects is calculated to determine whether they are in a pose of collision. Combined with example and Matlab simulation, the correctness of the collision detection model is verified by simulating the obstacles, lower motion platforms and vertical actuators. This mode can provide valid data for the study of collision avoidance among the obstacles, the vertical actuator and the lower platform.

The operation of pedestrians and vehicles at signalized intersections differs between common right-turning and channelized right-turning. Compared with common right-turning, in channelized right-turning, the vehicle turning radius is larger, and velocity becomes faster. For pedestrians, there is no such aggregation characteristic during the red light. To get a better understanding of the conflict between pedestrians and channelized right turn vehicles at signalized intersections, odds-ratios were used to analyze the conflict risk between pedestrians in crossing the road and vehicles. Results show that the occurrences of pedestrian-vehicle conflict and pedestrians' yielding behavior to vehicles are influenced by the scale of crowd, the velocity of pedestrians, the trajectory chosen by pedestrians to cross the intersections and the direction whether the pedestrians are exiting or entering the intersection. On this basis, we made a penetrating analysis of different trajectory factors and obtained their influences on conflicts and pedestrians' yielding behaviors.

Bike sharing is a new dockless public bicycle system based on the mobile Internet technology. The solution of recycling problem of faulty bikes can reduce the bike sharing company's operating costs. A recycling process model of the faulty bikes is proposed in this work. Firstly, a K-means algorithm is used to cluster faulty bikes to different service points. Then a model of faulty bikes recycling is present in order to minimize the total recycling costs. The optimization software Cplex is used to solve the model. Finally, a case study based on a certain area in Beijing is carried out to validate the proposed model. The results show that the value of loading capacity greatly affects the results of recycling, and that the model and method proposed can be applied effectively to the recycling of faulty bikes.

Based on the arrival characteristics of traffic flow between adjacent intersection, the mechanism of arterial coordination was analyzed first considering six states of traffic flow arriving at downstream intersections. Then, the traffic delays were divided into three types, named motorcade head suffocated delay, motorcade trail suffocated delay and zero delay. Third, according to the aggregation and dispersion of the queue between associate intersections under different conditions, the length of the queue was calculated, which was then introduced into the calculation of the delay. Finally, an arterial coordination control optimization model was established. Five intersections along the Coastal Road in Qingdao were selected as an example to verify the proposed model. The quantitative relationship between offset and delay in correlated intersections was obtained, and the minimum delay as the target of offset was determined. The results indicate that the offset model based on queue characteristics is feasible and effective in arterial traffic coordination control.

To solve the distribution route optimization problem for not-fully loaded vehicles of Hazardous Materials (HM), a piece-wise linear approximation method was used to evaluate the potential transport risk dynamically with consideration of the impact of HM quantity. According to different optimization criteria of the transportation enterprises and the uncertainties of route attributes, a multi-criteria optimization model for distribution routes of HM with capacity constraints was formulated based on the credibility theory and expected value method. An improved simulated annealing algorithm was designed to solve this model. The fast non-dominated sorting method and the dynamic crowding distance calculation method were used to increase the solving efficiency and to improve the distribution uniformity of Pareto optimal solutions in solution-space. The variable neighborhood search strategies were designed based on the coding structure of the solution to enhance the local and global search ability of this algorithm. Different examples were used to illustrate the rationality of the model and the effectiveness of the algorithm. The research results can provide decision support for the distribution route selection of HM transport enterprises under multiple uncertain conditions.

Rutting is closely related to the pavement temperature field. As a result, using the concept of rutting equivalent temperature to predict rutting distress is a common method. Different from previous studies, this study puts forward a calculation method of the rutting equivalent temperature for asphalt pavement at different depth, which is based upon the pre-established prediction models of asphalt temperature field and rutting distress. Meanwhile, the rutting equivalent temperature prediction model considering regional differences and asphalt material properties is initially established and tested. Compared with the existing rutting equivalent temperature prediction model, the proposed model reflects the influences of different depth, different regions and different types of asphalt materials. It is demonstrated that this model has better applicability.

Eight groups of centrifuge tests of pit foundation supporting with scattered piles and soil nailing were carried out by Tongji University. Based on the results of eight group tests, the analysis model of the final form of soil arching effect in pit supporting structure with scattered row piles and soil nailing was put forward. Not only the dynamic development process of soil arching effect between piles was explored, but also the effects of the pile spacing and soil nailing parameters on the soil arching pattern was analyzed in combining with three-dimensional finite element numerical simulation. The results show that, for pit supporting structure with scattered row piles and soil nailing, when the ratio of pile spacing to pile diameter is greater than 8 or the ratio of soil nail length to the excavation depth is greater than 2/3, the soil between piles can not form soil arching effect. In addition, when the ratio of pile spacing to pile diameter is in the range of 4~6 and the ratio of soil nail length to excavation depth is less than 2/3, the soil arching effect is obvious. Furthermore, for pit supporting structure with scattered row piles and soil nailing, the final form of soil arching effect is a curved surface, which along the top surface appears as parabolic and along the depth direction has a certain vertical depth.

This paper presents a study on seismic response, which influences the Bridge Crossing Active Fault (BCAF). One bridge crossing strike-slip active fault is taken as an example. The seismic ground motions at the site of the bridge are generated following a hybrid simulation methodology. Multi-support excitation displacement input models and time history analysis method are used to calculate seismic response of the structures. The fault crossing position, crossing angle, arrangement of the pier are taken as the key parameters, and their influences on the seismic response are discussed. The results show that the fault crossing position, crossing angle, arrangement of the pier are important parameters influencing the seismic responses of the bridge, which shoul be taken into consideration in the design. With the change of the fault position from the mid-span to side-span of bridge, the most disadvantageous seismic response shows a trend of increase, being unfavorable to stress of the structure. When the fault perpendicularly crosses the bridge the seismic response caused by fault dislocation is minimal, and the optimal rationality and economy for piers internal force are presented. Generally, the demand of maximum transversal bending moment of BCAF does not necessarily occur at the highest pier, but on higher bent on the side of the fault with two bents. The structural response characteristics of the BCAF and the near fault bridge have significant influence, therefore should be paid particular attention in seismic design. The results in this research can be used as a reference to the seismic design and bridge location.

The composite girder bridge with multi-cell corrugated steel web has been widely used in recent years, but there are some deficiencies in the corresponding specification and fine calculation methods. At present, the design mostly adopts the computational analytical method combining the spatial bar system model, plane beam grillage model and solid model. However, the spatial bar system model is short of the refinement analysis on the space effect, such as the shear lag effect, effective distribution width problem, and eccentric load factor problem etc. Due to the similarity of the plane beam grillage method in the equivalence principle, it can not accurately reflect the shearing stress distribution and local stress of the top and bottom plates of the box type composite beam. The solid model is very difficult to combine with the overall calculation. Moreover, the spatial grid model can achieve the refinement analysis, with the integrity of the analysis and the comprehensiveness of the stress checking calculation, and can make up the deficiency of the analytical method currently. Through the example verification of the solid model and spatial grid model, it can be seen that the calculation results for the stress and the displacement of two models are almost consistent, indicating the applicability and precision of the grid model.The real bridge analysis shows that the space grid model of multiple chamber box girder with corrugated steel webs can be calculated the web shear transverse distribution clearly, and the each web shear distribution rules of cable-stayed bridge at the key sections of side and mid span, so it can provide some suggestions for this kind of engineering design.

A combustion simulation model is built with AVL-Fire, and the dynamic distribution characteristics of the mixture concentration field and temperature field are analyzed by verification of the simulation results on test bench. The start point of combustion, premixed rate, diffusion rate and combustion duration are taken as the evaluation parameters. The effect of the control parameters on Heat Release Rate (HRR) at different stage is analyzed. An Artificial Neural Network (ANN) model is established for quantitative prediction of the evaluation parameters. The structure and of the ANN algorithm is confirmed. The effects of the injection parameters and intake parameters on the evaluation parameters are analyzed and predicted. The results show that, when the structure 5-18-4 ANN with trainlm algorithm are adopted, the robustness, responsiveness and convergence precision are good, indicating the ANN model is a powerful tool to predict HRR.

A two-stage fuel injection warm up test was carried out on a GDI engine, and the influence of the fuel ratio of the two injections on the combustion and emission was studied in the warm-up process. The results show that with the coolant temperature rising from 30 ℃ to 80 ℃, the combustion and emission conditions are improved gradually that in different fuel ratio, the peak cylinder pressure is increased by 5%~10% , and HC emission is reduced by 32%~37%.With the fuel ratio changing from 9∶1 to 3∶1, during the coolant temperature raising from 30 °C to 80 °C, the peak cylinder pressure is increased by 16.68%, and the phase of peak cylinder pressure was 4° CA earlier, the peak heat release rate is increased by 2 J/°CA, HC emission is reduced by 15% on average, NO_x emission is increased gradually, the total concentration of particles is decreased gradually, particle size distribution is gradually shifted from the bimodal distribution of nuclear mode and accumulation mode to unimodal distribution of nuclear mode.

A rollover prewarning method for articulate loader is proposed. Firstly, the steering angle, boom lift angle, bucket rotation angle and boom cylinder pressure are measured. Second, with these information and using multi-body dynamics algorithm, the stability diagram is calculated, which is an enclosed plot in polar coordinates to show the rollover angles in different directions. Third, according to the slope angle and the relative orientation of the loader toward slop measured by gyroscope, the state point of loader stability is determined in the stability diagram. Finally, early warning is achieved by judging if the state point is close to the boundary of the stability diagram. The accuracy of stability diagram and the feasibility of early warning method are verified by the model experiments.

Comparing with compressive primary crushers such as jaw crusher or gyratory crusher, mineral sizer breaks the material mainly by shear or tensile force. Besides, mineral sizer has unique characteristics that it can crush larger mineral particles and shorten the crushing process. Because of the complexity of the crushing process of the mineral sizers, it is difficult to calculate the breaking force acting on the breaking tooth and its distribution theoretically. Therefore, a DEM (discrete element method)-FEM (finite element method) coupling method was applied in this article for the tooth strength analysis in order to improve both the authenticity and the accuracy of the calculation. In this method, the distributed loads, which are the output from the DEM simulation software of EDEM, are applied to the FEM model constructed in the software of ANSYS. It is essential to calibrate the bonding parameters in DEM simulation so as to properly describe the breaking behavior. In this article, uniaxial compression experiment and Brazil disk test were taken as the calibration methods for the bonding parameters according to the compressive and tensile strengths of different bonded materials considering the size effect. After the calibration of bonding parameters, crushing process was simulated in software of EDEM and the breaking force distribution on the tooth at any second was obtained and analyzed. Finally, the force distribution at the moment when the largest breaking force occurred was exported and applied on the FEM model in the software of ANSYS for the strength analysis. The results of the FEM analysis show that the breaking force load mainly distributes on the back of the breaking tooth and the stress concentration occurs at the front part of it.

The tensegrity mechanism is a new type of rigid-flexible spatial model, with advantages of light weight, easy-folded, extended flexibility and high precision. In order to improve the compliance of precise manipulators during normal operation, a model for tensegrity parallel mechanism in six DOF was proposed, which was built in quadrilateral prism type. The proposed model was mainly composed of four rigid drive branched chains, two rope drive links and two branched springs. The kinematics equation of the mechanism was established by closed vector method, and the relation between the angular velocity, the angular acceleration and the Euler angle during platform moving was deduced. According to the screw theory, the static equilibrium equation and the stiffness matrix of the mechanism were derived to solve out the numerical values. After that, the pseudo-rigid-body dynamics pattern of the mechanism was set up by the principle of virtual work. Finally, the velocity, acceleration and dynamic equations of the mechanism were numerically calculated. Based on the calculated values, simulations were conducted in MATLAB to verify the above mathematical solutions. This research provides theoretical references for the further studies on dynamics characteristics and control strategies of this mechanism.

In order to improve the wear resistance of grinding roller, imitating the form of biological surface morphology in nature, we designed and machined the bionic stripe structure with different depths and different circumferential numbers on the surface of the roller. The wear test of the grinding roller was carried out according to the design scheme of orthogonal test. The results show that the bionic stripe structure can effectively improve the wear resistance of the roller. Both the stripe depth and the number of stripe circumferences influence the wear amount. Compared with the standard roller, the maximum wear resistance of the bionic grinding roller increases by 72.1%. Analysis of the wear resistance mechanism of bionic grinding roller suggests that, the stripe structure makes the grinding roller contact with the quartz sand more fully and evenly, and the slippery grinding probability of quartz sand reduces on the roller surface at the same time, thus the overall and local stress conditions of the grinding roller are optimized, which is an important reason to improve the wear resistance of the roller.

Based on flow characteristics around the gill of alopias, a bionic thermal surface with a rib turbulator was designed in a simplified gas turbine transition piece. The thermal performance of the heat transfer surface with a rib turbulator in a double-chamber model was investigated. The CFD simulation was carried out using software FLUENT to obtain the temperature contour and the velocity vector distribution on the thermal surface. The impinging cooling efficiency and the Nusselt number were computed also. The regression equation about the area weighted average temperature and the minimum pressure on the thermal surface was obtained by using a comprehensive experimental design and regression analysis method to study the effect of height and width of the turbulence generator. The optimal height and width of the ribbed turbulence generator were evaluated by the particle swarm optimization algorithm. The simulation results indicate that the bionic thermal surface can improve the convective heat transfer and the cooling efficiency of the double-chamber is improved by 340%.

Thermomechanical simulation test is an effective method to study the new 9Cr heat-resistant steel. The experiment is carried out under the conditions of temperature 900~1050 ℃ and strain rate of 0.001~10 s ^-1 to determine the thermal deformation parameters. The constitutive equation of the new 9Cr heat-resistant steel based on the peak stress was established and the hot processing map of materials was drawn from the dynamic material model and Murthy instability criteria. The results show that the temperature and strain rate of deformation have little influence on plasticity of new 9Cr steel, but play a decisive influence on the dynamic recrystallization. The microstructure evolution of new 9Cr steel under different conditions was observed by optical microscope, and then the dynamic recrystallization and the variation were studied.

The 22MnB5 ultra high strength steel quenching joint in state welding is taken as the research object, intensive processing experiments are carried out in spinning machinery in the process of welding. The mechanical properties of the welded joints, microstructure, microhardness, corrosion resistance are tested and analyzed. The results show that the welding joint after mechanical spinning, a plastic deformation layer is formed on the joint surface. The grains in the head spinning are flattened and elongated and the grain distribution is more uniform. The tensile strength of welded joint is about 940 MPa, which is about 100 MPa higher than that of untreated welded joints. The microhardness is slightly increased, but the corrosion resistance is changed little.

This paper proposes an adaptive Backstepping control method with H_∞ tracking performance for mobile robots, which considers the uncertainty of parameter and disturbance in the dynamics equation of mobile robot. Firstly, the trajectory tracking controller with asymptotic stability is designed based on the kinematic model with the linear velocity and angular velocity as the virtual control input. Secondly, in order to deal with the parameter uncertainty in the dynamics system, an adaptive method is used to estimate the parameter online. The method ensures the boundedness of the estimated parameters by introducing the discontinuous mapping. Finally, the dynamics controller is designed by Backstepping method, meanwhile, the same time, the H_∞ performance of the proposed controller against disturbance is proved in the Lyapunov framework. The simulation results show that the controller can control the mobile robot tracking the reference trajectory even if the system parameters are unknown. For the bounded external disturbance, the controller can control the output deviation of the system within the bounded range. It is proved that the proposed method can suppress the influence of the system parameters and the disturbance on the control deviation in the presence of unknown parameters and disturbance in the system.

It is an important application of robot in the field of nuclear power station to realize welding for a nuclear reaction pool in emergency and patrol daily. A cable underwater robot system with a full range of motion which can attach on the wall by propellers was put forward in consideration of bump and pipeline at the bottom or side wall of the reaction pool. The changes of the center position of gravity, the center position of buoyancy and the moment of inertia were analyzed when the movements of the body and an operating device cooperate with each other, and a dynamic model of the robot was established. The position and attitude changes with time were compared before and after adjusting the position of an operating device under no thrust input by Matlab/Simulink. The rotating efficiency of a robot was analyzed by regulating the position of gravity, buoyancy center. The results demonstrate that the attitude stability of the robot without thrust input is increased and the rotation efficiency is improved by adjusting the position of an operating device. The dynamic equation for this robot lays a good foundation for controller design.

This paper studies the planning problem of investigation, strike and evaluation task for known targets with multi-heterogeneous fixed wing Unmanned Aerial Vehicles (UAVs). Firstly, the task execution time-consuming constraints, collaborative attack constraints are added to the cooperative multiple task assignment problem (CMTAP). Then, based on the distributed planning architecture, the integrated solution of the two sub-problems is completed by improving the gene coding method. In order to optimize the uniformity of the UAVs' planning trajectories, the total time index of the task completion is added to the cost function. The feasibility of this method is verified by the six-degree-of-freedom model and the Vector-based path following algorithm. The simulation results show the rapidity of this method.

In order to realize micro Electrical Discharge Mchining (EDM) surface roughness evaluation, a fast Gauss evaluation algorithm of surface roughness is proposed based on the results of scanning electron microscopy (SEM). First, the Gauss evaluation algorithm is described in detail, and the data to be evaluated are preprocessed. Then, according to the SEM data of the micro electrical discharge machining, a fast evaluation algorithm is designed based on a Gauss Evaluation benchmark. Finally, with the input of the results obtained from micro electrical discharge machining and SEM on actual samples, the simulation experiments of surface roughness evaluation are carried out using the proposed fast Gauss evaluation method, convolution algorithm and FFT algorithm respectively. Experimental results show that, compared with the other algorithms, the proposed fast Gauss evaluation method significantly improves the efficiency of evaluation. Estimated time for 512×512 sampling points can be shortened to 2 s from 6 s and 4s of the other two conventional methods respectively.

In order to extract and match the texture features of dorsal hand vein, this paper presents a dorsal hand vein recognition system based on optimized texture features. First, we design a image acquisition device to collect vein images. Then, after image preprocessing and three-layer haar wavelet decomposition, we use different scales and directions of gabor kernel function to extract texture features of low-frequency sub-band images. Finally, PCA is used to reduce the dimensionality of features and the nearest neighbor classifier based on Euclidean distance is used to match the features. In this paper, a dorsal hand vein database of Jilin University is established. The experimental results show that the proposed recognition system can effectively improve the identification speed of features and the recognition rate can reach 98.5%. Good efforts on recognition accuracy and efficiency are achieved by the system.

In order to promote the conditional dependence expression ability of K-dependence Bayesian Classifier, in this paper we relax the conditional independence between predictive features according to the basic idea of Markov blanket for feature extraction, and carry out structure learning based on greedy search strategy. We propose a macro model from the training set, and a micro model from each test sample. The decision is made by the ensemble of both models. We select data sets from UCI machine learning repository for cross validation. The experimental results demonstrate the rationality and effectiveness of the proposed algorithm in terms of 0-1 loss, bias and variance.

With the rapid development of cloud computing, more and more people shift their workload to the cloud data centers, the energy consumption of the cloud data centers is non-negligible. In order to reduce energy consumption and achieve green cloud service, an energy saving strategy in the cloud data centers based on synchronous sleeping of the reservation virtual machines is proposed. All the Virtual Machines (VMs) are divided into two groups: the main group and the reservation group. The main group is always activated. According to the traffic load of the cloud data centers, the reservation group may be asleep or activated. Considering the cloud data centers served for the user requests with multiple tasks, a batch arrival queuing model with part servers synchronization multiple vacation is established. By using Gauss Seidel method, the steady-state distribution of the system model is derived, and the performance measures, such as the energy saving rate of system are evaluated. Numerical experiments with analysis and simulation are provided to study the system performance of the energy saving strategy in the cloud data centers based on synchronous sleeping of the reservation VMs and verify the effectiveness of the proposed strategy.

In order to improve the ability of moth optimization algorithm in solving practical problems, a novel hybrid moth optimization algorithm based on single storage is proposed. The algorithm uses particle swarm optimization to obtain the initial optimal location of the moths, so that it has a better and richer population. The optimal solution is also obtained by the simplex method. The proposed algorithm is compared with other four kinds of algorithms by tests on ten functions. Experimental results show that the convergence speed and solution quality of the proposed algorithm are better than other algorithms, it has better solving and optimizing performance, and can be used as an effective optimization tool.

For the judgment of the lesions of large-scale planted crops, this paper uses machine vision technology to propose an agricultural disease judgment algorithm based on surface multiple feature decision fusion. First the images of multiple agricultural leaves are collected that different surface feature extraction algorithms are applied to each leaf. Second, the feature weights are set and different features are combined to determine the likelihood of lesions in the leaf. Finally, the method of majority voting decision is used to judge the lesions in the region of each leaf. Compare with the artificial judgment method, this algorithm reduces the workload and can provide effective early warning of crop diseases.

In multisource multirelay cooperative networks, to avoid the interference among the relays that assist the same source and maximize the relay utility, we propose the joint power control and source selection distributed algorithm, which is based on two-level game. By using the alternate iterations between the non-cooperative game of power control and the evolutionary game of source selection, the proposed distributed algorithm can effectively suppress the interference and choose the optimal source. Furthermore, we prove that the two-level game model has the unique Nash equilibrium. Simulation results show that the proposed two-level game distributed algorithm can converge to the Nash equilibrium.

Considering the low robustness of existing saliency detection methods in detecting complex nature images, a new saliency detection algorithm is proposed by combining the absorbing markov chain and manifold ranking. First, the entropy of the gray image is calculated to obtain the number of superpixels. Secondly, a two-stage saliency detection is used. In the first stage, the image edge superpixels are pre-processed, and then the first saliency map based on random walk on Absorbing Markov chain was computed by using background prior. In the second stage, the saliency scores computed by the first stage are used as foreground query seeds for manifold ranking in order to further optimize the saliency detection results. Experiments on publicly ASD and ECSSD databases demonstrate that, compared with the existing saliency detection methods, the proposed method can accurately highlight the saliency target, and effectively suppress the background area, while improve the performance evaluation of precision, recall and F-measure values.

In order to avoid the problems of manually selecting features, difficulties in feature extraction and low visibility in traditional underwater image enhancing methods, an underwater image enhancing method based on convolution neural network was proposed. First, an underwater image dataset was established by simulation of underwater image formatting process by using clear images and underwater degraded model. Then, a convolution neural network model for underwater image enhancement was established, which can directly create a mapping relationship between the clear image and the underwater image. Finally, underwater images were restored by the features extracted by convolution neural network of underwater image. The experimental results show that the proposed method has less noise and improves contrast than tradition methods, and the algorithm provides a new idea for the underwater image enhancing research.

The speckle noise causes the target feature unobvious in the SAR image. The traditional CFAR algorithm only considers the strong scattering feature of object, and it is easily affected by much false alarm target. This paper proposes a novel conjugate CFAR algorithm which is more precise for object detection. The conjugate CFAR algorithm extracts the strong scattering feature as the traditional CFAR algorithm does. Then, the shadow feature detection is implemented by modifying the integral interval of the traditional CFAR algorithm. Therefore, the better results can be achieved by deleting the false alarm target according to the spatial relationship between the strong scattering feature and the shadow feature. The experimental results show that the false alarm rate of the algorithm is much lower than that of the traditional CFAR algorithm at the same detection rate.

Capacitive touch panels (CTPs) have been more popular owing to their excellent interactive performance. However, electronic magnetic interference, display noise and process variation negatively influence the accuracy of predictions of touch positions. A smooth tracking algorithm for capacitance touch panels based on Kalman filter was proposed. The original touch capacitance data and touch noise characteristic were detected using a micro-controller unit and a sensor integrated circuit. Additionally, the noise of touch trajectory was suppressed by Kalman algorithm timely through process model and measurement model of touch trajectory smooth system. Experimental results indicate that the proposed touch trajectory smooth algorithm has better performance compared to the traditional filtering algorithms, and the algorithm can be used to CTP systems.

Inherent non-determinism in parallel program brings huge difficulty and challenge to programmers to test and debug. The major source of non-determinism is the undetermined order of shared memory accesses. Deterministic replay provides the possibility of reproducing parallel program execution, and deterministic replay method based on page protection mechanism can reduce the recording overhead effectively. However, the exception of page protection is the main performance cost. This paper deeply analyzes runtime overhead incurred by release synchronization points in record-runs, and proposes a deterministic replay method based on adaptive release strategy. Performance evaluation shows that the record slowdown is much smaller than prior user-level replay system.

The image information of spaceborne synthetic aperture radar is rich and complex, and no information can be used to accurately express all the targets. The images of the same kind are very different, and the images of different kinds have great similarity. The traditional method can not solve the above problems, which leads to the low cutting precision. To solve this problem, a new small feature segmentation method for spaceborne SAR image is proposed. The gray-scale feature of the spaceborne SAR image is extracted, and the texture features are extracted by gray-scale symbiotic matrix. The feature contour model of spaceborne synthetic aperture radar image can be identified by spectral embedded clustering method, and the contour feature is extracted. The SVM is used to segment the small feature of spaceborne SAR image, and the spatial relationship of the over segmented parts is studied by using different small feature information. In the global range, the similar over segmentation parts are processed by iterative growth, and the small feature segmentation is carried out on the spaceborne synthetic aperture radar image. Experimental results show that the proposed method has high accuracy.

It is difficult to search the global optimal solution when traditional region adjacency graph is used to describe the data structure. To solve this problem, an image segmentation algorithm based on SLIC and fast nearest neighbor region merging is proposed. This method introduces the nearest neighbor graph to optimize the global search based on the region adjacency graph. First, an image is divided into small regions by SLIC superpixel algorithm, in which the relation between regions is described by the adjacency table structure of the Region Adjacency Graph (RAG) and the Nearest Neighbor Graph (NNG). Then, the value of the dissimilarity function is calculated between each region to be merged with all of its adjacent regions. Finally, the region with the least similarity is merged. The experimental results show that the proposed algorithm can combine the most similar regions, and reduce the complexity of the merging calculation compared with the traditional region merging algorithm, which greatly improves the accuracy of the regional merging.

In order to solve the problem that the performance of Direction of Arrival (DOA) tracking of dynamic target algorithm deteriorates under single snapshot, a DOA tracking method using dynamic compressed sensing is proposed in this article. Firstly, by extracting priori information about sparsity from last signal vector, we estimate the probability of the elements being non-zero and build a sparsity probability model of dynamic DOA in the present signal vector. Secondly, we get the locations of nonzero elements of present signal vector and achieve real-time DOA tracking of dynamic targets by applying this sparsity probability model to the reconstruction of sparse signals and minimizing a weighted norm. This method can obtain real-time DOA tracking of dynamic targets in a single snapshot condition and has better tracking performance than particle filter under the same conditions. Finally, the effectiveness of proposed algorithm is verified by numerical simulation results.