Due to the fact that the current neural network-based traffic flow forecasting method embeds part of the manually designed features,the feature extracted by the network has single function,bad adaptability,poor robustness and inaccurate characterization of the local features.Therefore,this paper proposes a traffic flow forecasting method based on residual Long Short Term Memory(LSTM).The model uses the idea of ensemble learning to train the spatially distributed data into a residual LSTM network.In addition,dimension weighted units are introduced after each LSTM unit to present the interdependencies between different dimensions of modeling features.Experimental results show that the method can realize adaptive modeling analysis of short term traffic flow data.

In order to improve the tracking precision of the unmanned vehicle steering system to the target path,this paper proposes a new unmanned vehicle path tracking method.It introduces the lateral control scheme of unmanned vehicle based on tracking preview point,and gives the dynamic linearized data model of the system and its multi-step prediction equation.The estimation and prediction model of pseudo-gradient vectors is derived by using the method of least squares.Adopting the advantages of model-free adaptive predictive control,it can perform repeated online calculations through rolling optimization strategies,so as to obtain better dynamic performance.Based on the CarSim/Simulink co-simulation platform,the method is verified when the vehicle speed is 5 m/s and 20 m/s.Results show that compared with the MPC scheme based on vehicle dynamics,this proposed method has better tracking performance.

To address the re-identification problem of different individual vehicles with identical types,a new vehicle re-identification algorithm is proposed.According to the component detection algorithm,the window and the vehicle face region with large differences between different vehicles are obtained,and the vehicle features of the detected vehicle window and the vehicle face region are extracted and merged to generate new fusion features.The distance measurement between image features is calculated for classification and recognition.The test is carried out on the public dataset VRID-1 of Sun Yat-sen university and results show that the Rank1 matching rate of the algorithm reaches 66.67%,which is obviously better than the classical traditional feature representation algorithm,thus verifies the feasibility and validity of the algorithm.

Based on the traffic wave theory and VISSIM software simulation analysis,the critical condition of traffic flow that is appropriate for the intersection to set the Contraflow Left-turn Lanes(CLL) is studied.The critical calculation formula of setting left-turn lane is obtained by wave theory analysis and verified by example calculation and VISSIM simulation analysis.Based on the VISSIM software,the difference between the operational indicators before and after the intersection setting the left-turn lane is analyzed.Simulation results show that when the left-turn traffic flow is small that enters the CLL with unsaturated flow rate,it is not advisable to set the CLL.When the traffic condition which is appropriate to set the CLL,the setting of CLL can effectively improve the left-turn capacity and reduce the average delay,schedule time,length of queue,parking time,etc.The setting of CLL may not increase the number of the parking times,which is related to the left-turn traffic flow.

Aiming at the non-stationary characteristics of traffic trajectory big data,the preprocessing method of traffic trajectory big data is studied.According to the multi-resolution analysis characteristics of it,two-dimensional discrete wavelet,it is selected to denoise and compress the traffic trajectory big data.Building a preprocessing platform for traffic trajectory big data and combining with road traffic congestion state judgment criteria to analyze the real-time traffic situation in important sections.Analysis results show that the method can improve the speed of data processing and the analysis accuracy of congested sections.

Lidar has a great influence on the detection of ground obstacles due to the point swept to the ground.In order to accurately identify obstacles,it is necessary to accurately separate the obstacle point cloud and the ground point cloud.To address the shortcomings of traditional separation methods such as large computational complexity and poor scene versatility of the scene,a point cloud separation method based on three dimensional Mesh graph is proposed.The method connects the collected point cloud data into a Mesh map,finds the ground reference point according to different attributes of the ground point and other points on the Mesh map,spreads the ground attributes to other points and finally separates all the ground points.Experimental results show that the proposed method has a good ground point cloud separation effect while ensuring operational efficiency of the algorithm.

To address the multi-class Lightill-Whitham-Richards(LWR) traffic flow model on non-uniform roads,a low-dissipation central-upwind scheme is proposed.Based on the 4th-order Central Weighted Essentially Non-Oscillatory(CWENO) reconstruction and the low dissipation central-upwind numerical flux,the dissipative characteristics of the numerical format are optimized by constructing different forms of global smoothing factors and increasing the nonlinear weights corresponding to non-smooth templates.The Runge-Kutta method discretizes the semi-discrete numerical scheme in the time direction to maintain 4th-order precision.The numerical simulation of the lane number change and traffic signal control problem of multi-class LWR traffic flow model on non-uniform roads shows that the scheme has 4th-order solving accuracy and high resolution.

The car steering geometry model is an important part of the unmanned driving development process.However,in the face of complex roads,simple preview tracking can no longer meet the unmanned driving demand.Therefore,based on Stanley’s steering geometry preview model,an adaptive optimal preview time improvement algorithm is proposed.According to the vehicle’s current heading angle,longitudinal vehicle speed,front wheel rotation angle and the curvature of the primary planning path,etc.,it combines different previewing time to predict the running status of the vehicle within a certain period of time,and selects an appropriate preview time according to the designed optimization function.It adjusts the front-wheel rotation angle of the vehicle to reduce the horizontal deviation during the vehicle’s driving,so that the vehicle can smoothly pass through the test road section.Simulation test results show that the adaptive optimal preview model based on Stanley algorithm can realize unmanned driving on complex open roads.

Using Fast RCNN method for feature extraction takes a long time and the detection accuracy is low,this paper proposes an improved forward vehicle detection method that combines the forward vehicle detection model based on Faster RCNN with Convolution Neural Network(CNN) of three different sizes.The forward vehicle detection ability of different methods on three traffic scene databases is researched and compared.Experimental results show that compared with the deep CNN,this method improves the accuracy and robustness of vehicle detection,and has a generalization ability.

Adaptive Cruise Control(ACC) is one of the most core technologies in the field of smart driving.Researchers mostly use traditional hierarchical control methods or variable control algorithms to implement this technology.These algorithms can not respond effectively to unexpected follow-up road conditions.For this reason,this paper combines deep reinforcement learning with ACC,and proposes an ACC algorithm based on deterministic strategy gradient algorithm,so that the intelligent vehicle can complete adaptive cruise and continue to improve in the continuous self-learning process.The test results under the open source platform show that this algorithm can make the ratio of the acceleration of the smart driving vehicle within 1.8 m/s2 within 90% of the follow-up acceleration,which can reach the level of the cruise control of the human pilot.

The polynomial fitting strategy is used to generate the intersection trajectory,so as to reduce the complexity of the prior map acquisition and the road network data processing in the intelligent vehicle navigation.In this process,a coordinate rotation method based on latitude,longitude and heading angle is proposed.After fitting the intersection trajectory,the Dijkstra algorithm is used for global path planning.Road network data is collected and the experiment is performed with “Beijing Dragon” intelligent vehicle as experimental platform.The results show that the fitting strategy mentioned in this paper has the small deviation and realize the global path planning function,which can meet the safe driving requirements of the intelligent vehicle.

The traditional reinforcement learning method is limited by discrete state space and discrete action space,and can not be applied to ramp merging scene.Therefore,a reinforcement learning based autonomous driving ramp merging model is constructed.The reinforcement learning model is built by deep Q network.The ramp merging problem is incorporated into the category of reinforcement learning problem and solved.Experimental results show that the model can adopt different strategies for different environment vehicle speeds,thus improving the intelligent decision-making level of the autonomous driving in ramp merging scene.

Aiming at the shortcomings such as the dramatic change of background,the influence of illumination change and the indistinguishable background color in the intelligent driving process,an improved Kalman particle filter algorithm is proposed.A gray-scale projection algorithm is used to preprocess vehicle video and image sequence frames.Harris corners dertection is used to extract angles of the target and background in the image region,and Kalman-embedded particle filter is used to make a second prediction of particle filtering.The effectiveness and accuracy of dynamic tracking are ensured during smart driving.Experimental results show that compared with the traditional KPF algorithm,the Moving Object Tracking(MOT) ability of the algorithm in different scenarios is obviously enhanced,the tracking accuracy rate is 95.7% in a complex traffic driving environment,and has good real-time performance.

The Stanley method is less robust to disturbances and tends to produce larger deviations in path tracking.In order to solve this problem,an improved Stanley path tracking method is proposed.Road network data collected by the inertial navigation system of autonomous vehicles are used to build a high-precision road map.In order to get effective path information,A-Star algorithm has been chosen for path planning.The position of the look-ahead points are calculated by the logarithmic relationship between the vehicle speed and the fixation points,the look-ahead points are taken as the target points for the Stanley method.Experimental results show that the lateral deviation of path tracking of this method is reduced by 20% as compared with the Stanley method.

Only using CPU to process lidar point cloud data is unable to meet the real-time demand.NVIDIA Tegra X1 is chosen as a heterogeneous computing platform to accelerate the data processing algorithm for lidar.Combined the hardware architecture features and the characteristics of data processing algorithm for lidar,the load imbalance problem in the GPU optimization process is solved by coarse-grained parallelism.Meanwhile,zero copy and data localization methods are used to fine-tune the optimization.Experimental results show that compared with the current industrial computer used in the smart car,the performance of the improved lidar data processing algorithm is 5 to 6 times faster,which improves the real-time performance of the radar data processing for the smart car.

In the rail transit clusters dispatch system,the parameters of database connection pool is one-time set and can not be modified.Aiming at this shortage,a dynamic allocation strategy of multi-client database connection pool is designed.It uses a dynamic allocation algorithm to allocate the optimal number of connections according to the different access frequency of each client for current client,so as to achieve the purpose of improving the utilization of system resources.Experimental results show that the dynamic allocation strategy proposed in this paper can shorten the response time of connection pool and improve the efficiency of system.

Aiming at the problems in traffic sign detection process,such as that traffic sign shows dimensional changes,rotation distortion,projection distortion or the sign is partially occluded,in this paper,a traffic sign detection algorithm based on saliency map and Fourier Descriptor(FD) is proposed.Firstly,the Frequency Tuning(FT) method is used to get the saliency map.Secondly,binary operation is utilized on saliency map to achieve binary image and get the regions of traffic sign.By extracting outer contour of the regions,features of contour length and aspect ratio are utilized to filter interference information.Thirdly,the convex hull processing is applied for the eligible contours and FDs of convex hull are extracted and normalized.Finally,detection results are obtained according to the comparison with standard data.Experimental results demonstrate that the detection rate of the proposed algorithm is more than 95%.This algorithm meets the real-time performance requirement of traffic sign detection.

Focusing on the evaluation requirements of mobility support for the IEEE 802.11p protocol,this paper constructs a typical Internet of vehicles application scenario based on IEEE 802.11p and uses Poisson process to describe the vehicle arrival model in Internet of vehicles system.It uses Veins platform to build the simulation scene and sets the simulation parameters.From the end to end delay,packet loss rate and throughput,the influence of traffic density and vehicle running speed on the performance of IEEE 802.11p and the applicability of IEEE 802.11p in the Internet of vehicles environment are tested and analyzed.Test result shows that the vehicle running speed has a great influence on the Internet of vehicles performance of IEEE 802.11p,but the traffic density has little influence on the performance of it.At the same time,the IEEE 802.11p protocol is applicable in active road safety and it helps improve the cooperative transport efficiency.

According to the passenger movement pattern and the hot pick-up and drop-off areas extracted from taxi driving passenger data,this paper proposes a functions discovery method of taxi hot areas.Firstly,it uses taxi driving data clustering algorithm based on the temporal and spatial characteristics of traffic data to realize hot region division.Then,the passengers travel character discovery model of passengers in hot region based on Latent Dirichlet Allocation(LDA) is built to realize clustering hot taxi region with similar passenger travel mode.Finally,by summarizing the specific function of each area,it can find the relationship between area function and passenger movement patterns at different period of passenger flow.The experimental results show the method can effectively discovery the function characteristics of hot areas.

Aiming at the transmission instability problem when multiple users request a large number of data resources in intelligent vehicle network,this paper proposes a Set Covering Maximum Flow(SCMF) algorithm in mobile vehicle node environment.It uses set covering method to consider more request details and analyzes optimal stochastic demand model,which optimizes information resource sharing and data transmission process,guarantees vehicle information interaction between vehicle and wireless Access Point(AP) nodes with maximum flow efficiency transmission,and processes the customer’s big data resource request quickly and steadily.Experimental results show that the SCMF algorithm can obtain higher transmission efficiency with lower network overhead and transmission delay.It improves the utilization of network resources and comprehensive performance.

Aiming at the problems of low efficiency and single function of the existing illegal forensics analysis system for public security,this paper proposes an illegal forensics analysis system based on cloud platform and big data.It uses MongoDB to store original and mass vehicle trajectory data and MapReduce distributed computing to mine and screen trajectory data.The system scheduler on big data platform is used to schedule the resources of the system and monitor the running state of the system.Based on the extraction of discrete trajectory data for violation of the provisions of time travel,illegal provisions of the route,and fatigue driving for forensic analysis to find illegal vehicles.Experimental result shows that,when the vehicle trajectory data scale is large,the system still has high efficiency and stability.

In order to ensure the safety and efficiency of automatic driving,the vanishing points of road images are found based on Hough transform and voting method,thereby establishing the Dynamic Region of Interest(DROI).Then the illumination invariant lane detection algorithm is designed according to the features of white and yellow lanes to realize the detection of lane area under various complicated illumination conditions such as night and tunnel.On this basis,a polar angle constraint algorithm is designed to screen the candidate lanes to get the final effective lane.Experimental results show that the algorithm has a good detection effect which can reach an average accurate detection rate of 93.5% under various complicated illumination conditions.

Aiming at the problem that it is hard to review the accident scene or distinguish the authenticity of the known data when the information and data are insufficient in the traffic accident video,an algorithm based on the reverse method towards the earth celestial physics theory is introduced,which is aimed to estimate the geo-temporal position of the video through the shadow trajectories data to distinguish the authenticity.Utilizing the principle of the celestial bodies’ movement to generate the formula to describe the rule of the shadow,and it finds out the relationship among the length of the shadow,real solar time and average solar time by using the data fitting.It determines the local latitude.Towards the atmospheric refraction,the error between the real solar time and the average one,taking the advantage of solar azimuth and altitude,it optimizes the model,through which the accurate position can be estimated.Experimental result shows that the model can retrieve information and recovery scene.

Current related researches on parking methods are suitable for parallel parking scenario and vertical parking scenario.So,a general autonomous parking path planning scheme is proposed.The scheme takes the parking time as performance index,and formulates a time-optimal control problem of parking path planning with the vehicle kinematic constraint and path constraint.The pseudospectral method is used to solve the optimal control problem,and the obtained solution is guaranteed to be consistent with the path constraint by grid refining,so as to obtain the collision-free feasible parking path.Path planning in parallel parking scenario and vertical parking scenario are solved.Simulation experiment and real vehicle experiment verify the generality and feasibility of the scheme.

Aiming at the problem of anti-collision for intelligent vehicle control,a vehicle detection and ranging method based on monocular vision is presented.Multiscale Block-Local Binary Pattern(MB-LBP) and Adaboost are used to extract vehicle candidate area,the horizontal edge and gray features are used to eliminate false detection,the problem of the interference of road and green belts is solved effectively.The improved shadow location method is used to gain exact position of the vehicle,the accuracy of distance measurement is improved.A camera model based on position information is built to measure the vehicle distance ahead.Experimental results show that the average detection rate of preceding vehicles is 98.42% and the average error of vehicle ranging is 0.71 m.

The Two-dimensional Feature Fusion(2DFF) method based on two-dimensional feature matrix,i.e.,two-dimensional principal component analysis can the goal of feature fusion by decreasing the dimensions of the feature matrix,but it performs well got only when the difference in the dimensions of feature vectors is small.Some zeros after every single feature vector to get a two-dimensional feature matrix in the construction method of feature matrix of traditional 2DFF,which may change attributes of original feature vector at the condition that the difference in dimension of each feature vectors is huge and decreases the identification rate.Since the disadvantage above,a new construction method of feature matrix based on Singular Value Decomposition(SVD) is proposed.The new method groups all feature vectors end to end as a new one-dimensional feature vector which is decomposed into a two-dimensional feature matrix by keeping the phase of the signal unchanged based on the decomposition feature of SVD.Experimental result shows that the new method has a higher identification rate than traditional 2DFF feature construction method difference in the dimensions of feature vectors.

Taxi Global Position System(GPS) data contain macro information about the behavior of urban traffic and moving object behavior,from which valuable anomalous trajectory patterns can be mined.The location,geometry and travel time are taken as the spatial and temporal characteristics of the taxi trajectory respectively.According to the deviation of the feature,the trajectory anomalies are divided into temporal,space and spatio-temporal outliers.The trajectories of the same starting and ending points are extracted from the trajectory data,and are partitioned into segments.The similarity between trajectories is calculated and clustering based on distance and density is carried out.Frequent and the sparse trajectories are preliminary separated by the spatial characteristies.Based on kσ criterion,the separation threshold of temporal anomaly is determined to realize the classification of the temporal characteristic,and finally the trajectory outlier detection of the taxi is realized.The experimental results show that the method can extract personalized route as well as abnormal parking location and traffic section from abnormal trajectories,providing reference information for intelligent transportation as well as efficient logistics planning and execution.

In order to guarantee the message forwarding efficiency in Vehicular Ad Hoc Network(VANET),this paper presents a routing algorithm based on task allocation model.Network communities are divided according the social features of the node.Then,the consultation mechanism is used to allocate emergency safe and non-safe message tasks to communities and the community then signs tasks to the node.Incentive mechanisms are establised to improve the distribution efficiency of message tasks.Theoretical analysis and simulation results show that compared with other routing algorithms(such as Simbet,Prophet,Spray and Wait),the proposedalgorithm improves the messages delivery ratio,reduces the routing overhead,and meet the Quality of Service(QoS) requirement of diffrernt types of message tasks.

It is unavoidable that Vehicular Heterogeneous Network(VHN) need to exchange data by a gateway between each other.The data encapsulation method between heterogeneous networks is important for improving network protocol forwarding efficiency and determines the performance and reliability of the gateway.To improve the efficiency of gateway data forwarding,this paper proposes a new CAN/FlexRay network gateway data encapsulation method.Signals passing through the gateway in CAN and FlexRay networks are encapsulated into frames to make each signal frame contain more effective signals and improve the utilization of data.At the same time,it refines the BCBFD_LFS algorithm to confirm the optimum static time slot and decreases the time of operating message.Experimental result shows that the new method can improve the schedulability of messages,the network bandwidth utilization,and the data forwarding capability of gateway in VHN.

As the popularity of portable wireless broadband router MiFi equipment,the interference of urban rail transit train becomes more and more serious,therefore it is necessary to analyze coexistence anti-interference between Communication-based Train Control(CBTC) system and MiFi device.In order to improve the performance of system,this paper proposes an improved Distributed Coordination Function(DCF) scheme based on adaptive section,dividing MiFi node number into several sections and setting different contention window value.It firstly analyses the distribution model of inside carriage MiFi devices and interference model between MiFi and CBTC system.Secondly,it sets partition interval and contention window value according to the characteristics of interference.Finally,it uses the simulation tool to analyze throughput,transmission delay and transmission failure rate of CBTC system,in order to compare the different data frame length performance of traditional DCF scheme and modified DCF scheme.Analysis results show that the new DCF model optimization scheme can effectively improve the CBTC train-ground communication performance under MiFi interference environment,and it can achieve a certain degree of the coexistence.

In the Intelligent Road(iRoad) system,in order to meet the throughput requirement of packets with different priority,this paper presents an admission control protocol of dual-path routing based on priority.It analyzes the node available bandwidth based on priority and sends the paths which satisfies the requirement by data streaming to the final routing node.After receiving all the candidate paths,it selects the two best paths to meet the data streaming bandwidth requirement.When the data packet returns from final routing node,the local node uses its surrounding node available bandwidth prediction to do admission control to ensure preferential access of high-priority data streaming.Simulation results show that compared with the single-path-based admission control,the proposed algorithm has higher throughput,lower latency and jitter,which guarantees bandwidth demand for high-priority data flow.

To resolve privacy-preserving problem in Vehicular Ad-hoc Network(VANET),a message authentication scheme is presented by using secret sharing technology.This scheme can protect the identity privacy information of vehicle by the threshold proxy re-signature algorithm.It turns a signature generated by On-board Unit(OBU) into a signature from a trusted certificate authority on the same message,so it can effectively eliminate the risk of tracking vehicle according to signature.The new scheme distributes the re-signature key into Road-side Unib(RSU).Therefore it can reduce the probability of successful decoding re-signature key and prevent RSU from misusing the re-signature rights.Authentication center can trace back to the message of the real vehicle released,and solve the problem of illegal vehicle recall.Analysis result shows that the new scheme has high security and low storage overhead compared with similar schemes.

Considering the demand of low latency of intelligent vehicle access control system based on license plate recognition,a design for reducing the latency of system response performance from vehicle detection to lane release is proposed.This design mainly focuses on three aspects——pre-process in camera,image transmission and later-process in Industrial Personnel Computer(IPC).Using regional Auto Exposure(AE) in camera and asynchronous I/O image transmission to improve the performance of later license plate recognition,it realizes a kind of block transmission method combined with image region extraction while also optimizing image display in client-side of IPC,to improve the efficiency of the system’s each processing part.Experimental results show that from image acquisition to receiving,to processing and at last to show,this scheme can make system take about 136 ms while at the same time ensuring the accuracy of license plate recognition,which reaches the leading level in the field of vehicle access control system.

Aiming at the problem of long time delay and slow transmission speed of accident message publication method on highway,this paper proposes a Directional Single-hop Vehicular Broadcast(DSHVB) mechanism combining the application with Wireless Access scheme Vehicular Environments(WAVE).Each car node builds a car node list to record other cars driving in opposite direction,then accomplishes the directional accident message rebroadcast.It discusses the requirements to achieve reliable accident message forwarding and evaluates the performance of DSHVB using Veins.Experimental result shows that DSHVB has lower packet lost rate,higher quality of reliability and lower network load compared with flooding broadcast mechanism,and it meets the requirement of the fast broadcast of the accident message in highway scenario.

Task allocation is one of the key problems that need to be solved for design and implementation of multicore-based mixed-criticality system in automotive electronic system,and it needs to tradeoff among the schedulability,cost and resource efficiency.Aiming at solving this problem,this paper proposes a simulated annealing-based Criticality-aware Task Allocation(CTA) algorithm,which can realize the joint optimization of cost and resource efficiency by conforming to the constraint of system’s schedulability.The comparison experiment based on real-life automotive applications and simulated dataset verifies the effectiveness of the proposed CTA.

Cyber Physical System(CPS),as a new topic in research of computer science and technology,is the fusion of computation,communication and control. The Auto Adaptive Cruise Control (ACC) system is thought as a typical CPS,which has wide application prospect. This paper builds the mathematical model for the vertical travel of the car,and gives the architecture and the system state machine model of the ACC based on the theory of CPS. It designs the model of vehicle ahead,own vehicle and the distance between the two vehicles based on Ptolemy,and builds the hierarchical model of the system. In the submodel of the system,it constructs the system behavior model which is combined with model based on time and state model with the modal model. Simulation results show that the method can meet the requirements and ensures the security of ACC system.

The current train energy-saving optimization only considers the coordination between individual interval or train to lead poor applicability and guide optimization results. Aiming at the problem,this paper uses the energy-saving optimization of a single line interval as the research object to establish the interval train energy-saving optimization model,and resolves the model by genetic-annealing algorithm. It sets the time step-length to bring about the interval operation mode transition from the time-saving to energy-saving,constructs a knapsack problem to come true the energysaving optimization of line via distributing line reserve time and considering the time value of passengers,and solves it by using an improved greedy algorithm. A simple line is as an example to illustrate the maneuverability of the energy-saving optimization method. Experimental results show that the optimization method is effective,the energy consumption saves 41. 54% compared with time-saving mode,and it has good energy saving effect.

On the basis of the NaSch (NS) model,a one-dimensional Cellular Automaton (CA) traffic flow model is proposed,in which the relationship of vehicles moving distance and the vehicle speed is emphasized,the safety distance is considered,and the reduction process and location update rules are redefined. The model confirms current vehicle speed by vehicles moving distance and previous vehicle speed. By computer simulation,the relationships among speed,density and traffic volume are given to show the influence of moving distance on the traffic flow. The presence of metastable state,phase separation and hysteresis are revealed,which have been observed in real traffic. Simulation result shows that the model is reasonable and effective,and when the vehicle is on the road,its availability of road resources is high and traffic flow is big.

When environment of ballast is very terrible,the working states of track circuits are difficult to differentiate in spatial domain. In view of this,the time domain finite difference method is used to obtain the time domain solutions of track circuits in this paper. The partial differential equations group is dispersed based on the theory of partial differential equation numerical solution,and a differential formation of track circuits is got. Then time domain responses of track circuits for adjusting state are analyzed according to the boundary conditions of voltage and current at the beginning and terminal. Through an example,voltage changes of the receiving end for the adjusting state are simulated corresponding to different initial electrical parameters. Results show that the solutions of time domain conform to the transmission characteristic of track circuits,therefore this method can provide the theoretical basis for transient analysis on track circuits.

Aiming at the problem of the now high localization degree of China’s urban rail transport Automatic Fare Collection(AFC) system,a rail transport automatic fare collection oriented bill acceptor is designed. The whole structure is of high-reliability so that it can promote the security and reduce jammed banknote. In hardware aspect,it is of modular design which helps product maintenance and update. High performance with ARM processors of every independent module,high-speed brushless DC motor,an array of sensors ensure the processing speed of the system. And software aspect,it uses the PCA dimensionality reduction algorithms and Euclidean distance to build the template. Test result shows that the bill acceptor can discriminate the bill quickly and reliably,and reach the engineering requirements on receiving rate and bill jammed rate. In addition,the receiving speed can be improved by more than 1 s,which solves the problem of big services density,and sensitive failure rate.

In vehicular network,broadcasting messages are not often able to arrive and receive properly due to the vulnerability of wireless channels and high mobility of vehicles. To solve this problem,this paper proposes an index coding based message broadcasting scheme in purpose of improving the message transmission efficiency. Index coding is a variant of source coding scheme that exploits the side information at different receivers,and this paper focuses on implementing index coding technology in the message broadcasting of vehicular networks. It proposes a distributed feedback based side information collection mechanism and an improved graph coloring algorithm to find the maximum clique,and the indexing coding is done. Simulation experimental results show that the scheme can reduce the number of transmissions,thus save wireless channel bandwidths and improve broadcasting efficiency.

Mechanical fault of bogie seriously affects the security and comfort of the high speed train. Vibration signal of bogie and car body change with the fault occurrence, therefore, this paper proposes a fault diagnosis method of high speed train bogie based on Ensemble Empirical Mode Decomposition(EEMD). There are four typical working conditions in simulation experiment, such as air spring fault, yaw damper fault, lateral damper fault and normal condition. Vibration signal becomes several intrinsic mode functions after ensemble empirical mode decomposition. Energy moment feature is extracted to reflect the time distribution rule of energy. The 2nd to 6th energy moment are chosen to constitute 5-dimension eigenvector. In speed of 200 km/h, the Support Vector Machine(SVM) gets recognition rate. Simulation experimental result shows that the correct recognition rate of this method can achieve more than 95%.

The trains in the railway section communicate with each other by track circuit or base station indirectly, which has too many nodes and low reliability, also brings safety risk. In order to realize direct communication of train-to-train and the need of railway collision avoidance system, the train-to-train direct communication scheme is proposed, the ultrashort wave band, the maximum communication distance between trains and the structure of transceiver are selected for railway application requirements, also the feasibility of this technology is described. The link of the direct train-to-train communication is analyzed, the budget model of path loss of station and section are given, the calculated results under different conditions of the receiver power also with simulation are illustrated. Analysis result shows that the proposed technology can be realized. The fading in communication link is discussed with the character of train running environment, the result shows that the multipath propagation and Doppler effect lead to serious signal fading.

Train speed detection and positioning are key technologies for improving the safety and efficiency of train operation. According to the domestic and foreign research in this field, a train detection and positioning system is designed based on embedded processor and multi-sensor information fusion. An axle speed sensor, a Doppler radar speed sensor, an acceleration sensor and query bails are employed to collect train status information. Federal Kalman filtering and the multi-sensor information fusion method are used to process these information in an embedded system. Problem of errors caused by train wheel diameter wearing, idling, sliding and other factors in tradition system is solved. Simulation results in Matlab show that the system can effectively improve the precision of train speed detection and positioning.

On analysis of the braking force distribution strategy in existing Electric Multiple Unit(EMU) electro-pneumatic braking control system, an optimization strategy of electro-pneumatic braking cooperative control is proposed, which aims at solving the imbalance problem of the motor car and trailer car’s electro-pneumatic braking. It uses electro-pneumatic braking priority control principle in EMU, and distributes air braking force of motor car and trailer car in inverse ratio on the basic of load. By modeling the EMU electro-pneumatic cooperative braking control and braking force distribution optimized algorithm while taking one motor car and one trailer car in CRH2 EMU as the basic unit, the Matlab/Simulink software is used to simulate different braking condition. The results indicate that the electro-pneumatic cooperative braking control optimization strategy which based on the load inverse-proportion for braking force distribution has the significant effect on improving brake efficiency, reducing wheel tread wear of the motor car and the trailer car.

In order to overcome the verification and test difficulties caused by the absence of sensors when developing safety computer systems, a model-based simulation system for safety computer system is presented. Modeling on real sensors to control real hardware, by the method of combination of hardware and software, the simulation system provides actual physical input and output signals for safety computer system. At the same time, by modifying the model parameters, it can provide no damage fault injection for safety computer systems. The test results show that the simulation system can effectively verify the security, reliability and availability of safety computer systems and can speed up the development process of safety computer systems.

For a safety critical computer, large-scale data like database which has to be transferred in an instant time cannot be voted directly. This paper proposes a database update algorithm for safety critical computer based on status vote, which is to vote the database status instead of database itself. This algorithm can solve the problem of voting too much data in a short time, and compare versions of database of different modules in real time. A Markov model is built to calculate the safety and reliability of this algorithm. The results show that this algorithm meets the update requirement of safety critical computer.

This paper proposes a visual inspection and classification algorithm of rail surface defect. An overall structure of the visual inspection and classification system for rail surface defects is designed. It extracts the rail surface subarea based on horizontal projection method, combines the inspection result based on logic or operation, and classifies the defect based on BP neural network. Experimental results show that the algorithm is able to detect and identify the two rail surface defect, scarring and corrugated abrasion accurately, the accuracy of classification is 99% and 95%.

Session Initiation Protocol(SIP), as a core session control protocol of VoIP, its base functions are a bit simple. It can not meet the needs of force disconnect which is a core business of wireless train dispatching system. To the question above, this paper provides the architecture of the wireless railway dispatching system under the environment of WiMAX, extends the SIP message methods and transactions, designs the signaling process of new SIP message method, and provids the extension scheme on the client side and server side. Experimental results show that this scheme can meet the establishing time requirement of force disconnect on railway, and can realize force disconnect successfully.

For the designing complexity of CTCS-2 system, this paper proposes the method combining Unified Modeling Language(UML) and Symbolic Model Checking(SMC) for modeling and formal verification. It analyzes the mode conversion scene of CTCS-2 on-board equipment. The mode conversion scene of CTCS-2 on-board equipment is modeled by using the UML, and UML class diagrams and UML state diagrams are gotten as well, through formulating some exchanging rules to extend and abstract UML model and exchanging it to the NuSMV model. The property of to be verified system and system Symbolic Model Verifier(SMV) model are inputted to symbolic model verifier to check. The verified results are true, and it shows that mode conversion scene of CTCS-2 on-board equipment has activity, accessibility and security.