Research On Algorithm For Obstacle Avoidance And Motion Control Of Intelligent Vehicles On Structured Roads

In recent years,Vehicles have gradually towards the "software-defined vehicle" direction.Intelligent driving as the core of intelligent vehicle software has become a hot research topic.How to plan the obstacle avoidance path in real-time and control the fast,stable and accurate tracking of vehicles in complex structured road scenarios;it is the research difficulty of intelligent driving planning and control algorithms.In this paper,the following study is carried out on the obstacle avoidance and motion control algorithm of intelligent vehicles on structured roads :(1)Aiming at the obstacle avoidance problem of intelligent vehicles in structured road scenes,an obstacle avoidance path planning algorithm based on finite state machine driving decision strategy and fusion of different curve fitting methods is designed.In cruise mode,the lane lines sampled by splicing the third-order Bezier curves are used to generate cruise candidate paths.In lane change mode,quintic polynomial curve is used to generate lane change candidate paths.A variety of cost functions are designed and weighted to select the best candidate path for obstacle avoidance,which meets the requirements of intelligent driving vehicle obstacle avoidance path for safety,smoothness,and lane keeping performance.By Carsim/Simulink co-simulation,it is verified that the obstacle avoidance algorithm designed in this paper has good obstacle avoidance and lane-keeping performance in structured road multi-obstacle scenarios.(2)A longitudinal velocity tracking controller with a hierarchical structure is designed.The upper controller uses PID control,and the lower controller is a look-up table with logical switching,the throttle opening and brake master cylinder pressure are obtained by interpolating the throttle and brake calibration tables,and the switch logic between throttle and brake is designed.By Carsim/Simulink co-simulation,it is verified that the longitudinal controller designed in this paper has good speed tracking accuracy under acceleration,uniform and deceleration conditions,and the throttle and brake switching is smooth.(3)Aiming at the problem of poor control accuracy and real-time performance of intelligent vehicle tracking curvature changing road at high speed,a vehicle dynamics model considering path tracking is built,and a variable step MPC lateral path tracking controller is designed based on this model.Considering the cumulative error of the model,a short-step discretization prediction model based on a zero-order holder with higher accuracy is adopted in the early stage of the prediction interval,and a long-step discretization prediction model based on a first-order holder is adopted in the post-stage.Considering the vehicle path tracking performance and control smoothness,the objective function and constraints of MPC are designed.By Carsim/Simulink cosimulation,it is verified that the variable step size MPC has better tracking accuracy,real-time and robustness than the traditional MPC lateral controller on double lane change and highway.Finally,the real vehicle experiment is carried out based on the wire-control chassis experimental platform.The experimental results show that the lateral controller designed in this paper has high tracking control accuracy under the path of curvature change.