Research On The Path Planning In Specific Scenarios Of Intelligent Vehicle

With the continuous progress of technology and the sustained development of national economy,automobile is gradually becoming a substitute tool that can not be ignored and replaced in people's daily life.In view of this,entrusted by the Ministry of Industry and Information Technology and relevant departments,the Automotive Engineering Society published the Road Map for Energy Saving and New Energy Vehicle Technology,which was prepared by more than 500 industry experts.It is self-evident that Intelligent Network Unified Vehicle is one of the seven major areas involved in the circuit diagram and has been put on the agenda as a major breakthrough.Among them,the research and development of ADAS system is the most important part of the Intelligent Network Link,shouldering the strategic significance of "Made in China 2025".Based on the research of intelligent vehicle path planning in specific scenarios,this paper aims to extend to the hottest AVP system in ADAS research and development,that is,the automatic parking system for passengers.The research work includes intelligent vehicle research status,dynamics and kinematics modeling,global path planning and related optimization under specific scenarios,local path planning(lane-changing obstacle avoidance)algorithm and simulation analysis.Firstly,according to the background and significance of the development of intelligent vehicles at home and abroad,the current research and development hotspots of ADAS system are introduced.The key technologies involved,such as environment perception,information fusion,path planning and intelligent control,are briefly introduced.The research of path planning under specific scenarios,such as the application of AVP system in semi-closed scenarios such as fixed parks and underground parking garages,is emphasized.Based on the theoretical knowledge of vehicle dynamics and kinematics,a two-degree-of-freedom model using the tire model is proposed,and the turning control model is analyzed,which lays the foundation for subsequent local path planning,i.e.simulation control of lane change and obstacle avoidance,and chooses the front wheel angle to constrain,so as to ensure the trajectory safety and feasibility.By comparing and analyzing all kinds of obstacle avoidance planning algorithms,the A-star algorithm is selected for global planning and simulated in MATLAB.Thenthe potential field method is used to model the obstacles of the turning nodes and the road boundary,so that the path of A-star is close to the center line,which solves the shortcoming of A-star's path being too close to the obstacle boundary.Finally,B-spline curve is introduced to smooth the turning peaks.Optimize rationally so as to achieve the best global planning path.Aiming at the unexpected obstacle avoidance environment in global planning and combining with the driver's perception and judgment of the obstacle avoidance environment in lane-changing,a new X-Sin lane-changing model algorithm is proposed to solve the problems of excessive curvature of the sinusoidal model's initial and final nodes and sudden change of the rate of the constant velocity migration model.A mathematical model is established.The parameters of the model,such as lateral acceleration time and lateral acceleration,are combined with the dynamic The number is analyzed and determined to satisfy the comfort and safety reliability of lane change,while ensuring the robustness and scientificity of lane change trajectory model.Based on lane-changing trajectory,simulation verification and analysis of lane-changing obstacle avoidance for intelligent vehicles are completed.The test environment is built with Carsim software,and the track tracking controller is designed according to MPC control algorithm in Simulink to realize the joint simulation of vehicle tracking control.At the same time,the stability of MPC controller is proved by comparing the experimental data of different vehicle speeds,referring to the indexes of front wheel rotation angle,lateral error and lateral acceleration,and the X-Sin lane change is verified.The effectiveness and robustness of the obstacle avoidance model are discussed.