Research On Trajectory Planning Of Low-speed Intelligent Vehicles In Complex Scenarios

Autonomous driving is a technology that various auto manufacturers and Internet companies are competing to develop,and it is one of the trends in the development of automotive technology in the future.The application of autonomous driving technology plays a prominent role in solving traffic congestion,improving traffic efficiency and reducing traffic accidents.At the same time,under the trend of rising labor costs,autonomous driving can also provide new solutions for transportation problems in logistics,mining,ports and other related industries.As key technologies in autonomous driving,trajectory planning and trajectory tracking have a significant impact on the driving safety and rationality of autonomous vehicles.Therefore,based on the distributed electric drive intelligent vehicle test platform,this thesis studies the trajectory planning and trajectory tracking control technology.Firstly,two trajectory planning methods are designed.According to the different road types,the A* algorithm is improved so that the heuristic function can be adjusted according to different road types.Combined with the quintic polynomial path planning method,a hybrid A* trajectory planning method is designed.Considering the uncertainty of dynamic obstacles in complex scenes,a reachable set trajectory planning method is proposed.According to the kinematic constraints of the vehicle,the set of motion states that the vehicle can reach at different times are obtained respectively.Then,considering the distribution of obstacles in the environment,the motion state set is trimmed.Finally,a reasonable driving corridor is selected from the trimmed reachable set,and the final planned trajectory is obtained by numerical optimization.Secondly,according to the driving conditions of the vehicle and the structural characteristics of the intelligent vehicle test platform,a trajectory tracking control method is proposed.The trajectory tracking control is decoupled into lateral path tracking control and longitudinal velocity tracking control.According to the driving characteristics of the vehicle,the feedforward-feedback path tracking strategy is determined.Based on pure tracking control theory,a path tracking feedforward controller is established.Taking the lateral deviation and directional deviation at the center of mass of the vehicle as control variables,a path following feedback controller is designed based on the sliding mode variable structure control theory.Based on the longitudinal kinematics model of the vehicle,the deviation between the actual vehicle speed and the desired vehicle speed is taken as the control variable,and the longitudinal speed controller is involved by using the sliding mode control theory.Then,a joint simulation environment is built in the computer to simulate and verify the trajectory planning method.The vehicle dynamics model is provided by Car Sim,the simulation scene is built in Pre Scan,and the trajectory planning program is run in the ROS environment.The three establish communication through MATLAB and realize joint simulation.The two trajectory algorithms are tested in different simulation scenarios,and the effectiveness of the planning algorithm in complex dynamic scenarios and the rationality of the planned trajectory are verified.Finally,based on the distributed electric drive intelligent vehicle test platform,the real vehicle test of the trajectory tracking control method and trajectory planning method is carried out.Using the test platform,the tracking effect of the trajectory tracking control algorithm in the single-line shift and the "8" circular road was tested at different speeds,respectively.The test results show that the tracking control algorithm can well control the vehicle to run stably along the large curvature path,which verifies the effectiveness and real-time performance of the algorithm.Combining the trajectory planning method and the trajectory tracking control method,a real vehicle test is carried out in a simple static scene,which verifies the feasibility of applying the two methods to a real vehicle.