Research On Trajectory Planning And Control For Unmanned Vehicles On Structured Roads

The development of unmanned driving technology can not only reduce the rate of traffic accidents,but also improve the operating efficiency of vehicles in the road environment,and local path planning and motion control technology are the keys to realizing unmanned driving.Based on the Frenet coordinate system,this paper carries out the following research and analysis on the trajectory planning and motion control of unmanned vehicles on structured roads:(1)Research status of driverless cars at home and abroad.The development of unmanned driving technology at home and abroad and the research status of path planning technology are expounded,and the commonly used trajectory planning algorithms are summarized.(2)Development of a convex approximation and three-element obstacle avoidance model in the Frenet coordinate system.Aiming at the path planning problem in the structured environment,firstly,the vehicle is guided by the reference line in the structured road.The reference line is a series of waypoints,and the information of the waypoints is collected in advance by the inertial navigation equipment.Based on these waypoints,the local environment is modeled by the grid method of road adaptation in the Frenet coordinate system,and the principle of convex approximate obstacle avoidance is proposed for the perceived static obstacles.Secondly,a three-element obstacle avoidance model is proposed for the moving obstacle.Through this model,the connection between the moving obstacle and the ego vehicle can be established,and the measure of the threat degree of the obstacle to the ego vehicle can be obtained.Then use Matlab/Simulink and Carsim to simulate the actual collected waypoint route to verify the effectiveness of the algorithm and prepare for the subsequent real vehicle test.(3)Design of cost function and trajectory tracking controller.The kinematics and dynamics model of the target vehicle is constructed.In order to meet the requirements of vehicle driving safety,comfort,efficiency,and continuity between the current decision cycle and the previous cycle,a cost function based on MPC(Model Predictive Control)is designed;By continuously debugging the parameter values of various cost functions,the waypoint with the lowest cost function is obtained as the route to be tracked at the next moment.Then a trajectory tracking controller based on time delay dynamic prediction is designed,which can eliminate the vehicle position deviation caused by the delay and achieve the effect of accurately tracking the desired path.Finally,a planning algorithm for the leftturn scene at the intersection under multi-vehicle interference is developed.(4)The construction of the experimental platform.In order to verify the application effect of the algorithm on the real vehicle,a multi-sensor unmanned vehicle hardware platform was built,a hybrid system architecture software platform based on Lab VIEW and Simulink was developed,and a Lab VIEW based GNSS acquisition system was designed on the platform.And the horizontal and vertical control system of the vehicle chassis,and then carry out the real vehicle test of path tracking and obstacle avoidance.In this paper,the trajectory planning and tracking control technology under structured road are studied,and the convex approximation and three-element obstacle avoidance model under the Frenet coordinate system are proposed,and the vehicle is modeled.At the same time,the cost function and the horizontal and vertical control algorithm are designed,and finally the real vehicle experiment is carried out.The experimental results show that the algorithm developed in this paper can successfully avoid obstacles,and the algorithm can meet the requirements of ride comfort,safety and real-time performance when the vehicle is at a low speed.