Research On Decision Planning And Motion Control Of Intelligent Vehicle In Urban Traffic Scenes

With the advancement and development of science and technology,intelligent vehicle have become one of the most promising research directions in the current automotive industry.Intelligent driving technology can effectively reduce the frequency of traffic accidents and avoid road congestion,road congestion,etc.caused by different driver styles.Road traffic safety issues such as traffic accidents.The urban structured road environment is unique,and smart cars face many complex and uncertain factors in this environment.Behavior decision planning and motion control are the key technologies of intelligent driving vehicles.Improving the decision planning and motion control technology of intelligent vehicles in urban traffic scenes has profound research significance and application prospects.The thesis is supported by National Key R&D Program of China Intelligent Vehicle Control and Performance Improvement Technology(Grant No.2018YFB0106203).The research takes intelligent driving cars as the research object,mainly studies the decision-making planning and motion control problems in urban traffic scenarios,and designs a simulation test to verify and analyze the algorithm proposed in this paper.The main research work of this paper is as follows:Research on the decision-making and planning algorithms of smart cars in urban driving scenarios.Firstly,analyzing the characteristics of urban road environment.Based on the hierarchical state machine model,the decision-making and planning problems of intelligent vehicles in the urban traffic environment are divided into sub-state problems in normal road driving and intersection driving scenarios.In the normal driving road environment,focus on the decision-making and planning of lane-change driving of intelligent vehicle,analyzing the vehicle’s lane-changing intention through time and space constraints,and judge the feasibility of vehicle lane-changing based on the concept of minimum safe lane-changing distance when the vehicle has the intention of changing lanes and meets the feasibility,the trajectory cluster planning in the lane changing process are carried out based on the fifth degree polynomial,designed the cost function under the given constraints and the optimal solution is performed to obtain the optimal lane changing time solution to get the best lane change trajectory.For intersection driving scenes,consider the traffic laws and regulations to build the velocity plan algorithm when passing the intersection.For the problem of the lack of reference trajectories for vehicles at the intersection,path planning at the intersection based on multi-segment line splicing.Aiming at the trajectory following problem of the vehicle.The vehicle motion control problem is devided into two parts:lateral motion control and longitudinal motion control.For lateral motion control issues,build a vehicle’s lateral trajectory following controller based on the principle of model predictive control,model the lateral dynamics model of the vehicle,and use it as a predictive model to design the target under the constraints of control and output function,convert it into a quadratic programming problem to solve and obtain the control input,so as to realize the tracking of the desired trajectory.For the longitudinal speed control of the vehicle,the reverse longitudinal dynamics model of the vehicle is analyzed,and the upper-level controller is built based on PID theory.The desired acceleration obtained from it is solved by the reverse longitudinal dynamics model to solve the control value,thereby realizing the follow-up control of the desired vehicle velocity.Finally,a joint simulation platform is built through PreScan and Simulink,and the algorithm is implemented based on Stateflow and Simulink.Typical operating conditions are designed for different driving scenarios to verify the algorithm proposed in this paper,and the rationality and safety of the decision-making planning algorithm and the effectiveness of the control algorithm are verified.The test results show that the research results of this subject have a certain degree of safety,availability,practicality and reliability.