Local Path Planning And Tracking Of Autonomous Driving Cars

With the development of artificial intelligence technology and advanced sensing control technology,autonomous driving cars have achieved great development.From the technical point of view,the autonomous driving cars can be divided into three parts: perception and location,decision-making and planning,execution and control.Among them,decision-making and planning as the bridge of the autonomous system are getting more and more attention from researchers and being one of the key factors in the degree of security and safety margin.This paper improved and designed a method based on model predictive control algorithm for path planning and tracking control in the reference of previous studies.This paper first clarifies the research value and social significance of the development of autonomous driving cars,as well as the current development status of typical autonomous driving cars at home and abroad.After that,it briefly introduces the software and hardware system architecture of autonomous driving cars and analyzes several common planning and tracking algorithms and simulation software platforms for autonomous driving cars.Then,based on the kinematics and lagrange analysis mechanics method,the kinematics model and simplified dynamic model of the autonomous driving cars are established.Based on the model predictive control algorithm,the vehicle’s lateral tracking algorithm is designed.Based on the dual-PID closed-loop algorithm,the vehicle’s longitudinal control algorithm is designed.Based on the dangerous potential field method combined with the model predictive control algorithm,the vehicle’s local obstacle avoidance and the local path planning algorithm are designed.The algorithm simulation platform of the autonomous driving cars is built by Carsim,Prescan,and Matlab/Simulink,and the systematic simulation test from the environmental model re-emergence to the vehicle dynamics simulation is completed.In the premise of ensuring safety,through the positive development method,the iteration of autonomous driving cars is sped up.Based on the sightseeing car platform,the software and hardware platform of autonomous driving cars is built.The control parameters obtained by simulation optimization is taken as real vehicle control values.The algorithm is deployed and verified on the sightseeing car platform,and the results of the real vehicle performance are fed back to the simulation platform for parameter re-optimization.