Research On Highway Trajectory Planning And Control Strategy Of Unmanned Vehicle

The development of technology and the people’s requirements for living standards have promoted the development of unmanned vehicles.The development of unmanned driving technology can effectively improve traffic congestion and reduce traffic accidents.At present,unmanned driving technology is mainly researched on structured roads.Therefore,in this article,the main research is to study the driving trajectory planning and control strategy of unmanned vehicles in the highway environment.The research content of this article mainly includes the following points:(1)In this article,the driving environment of the vehicle is the expressway lane,so the highway is simplified first.Driving in the expressway lanes mainly includes three driving states:cruise,following and changing lanes.Based on the three driving states,a safe distance model for following driving and a safe distance for changing lanes are established respectively.In terms of driving decision-making,the safety distance is used as the basis for driving state switching.(2)There are curve conditions on expressways.Trajectory planning in the world coordinate system cannot accurately obtain the information of the vehicle relative to the road in the curve conditions,and the planning method is more complicated.In order to solve this problem,this paper introduces the Frenet coordinate system,and In the Frenet coordinate system,a fifth-degree polynomial is used for trajectory planning.The planning module will plan trajectory clusters for different driving states at each planning moment.The trajectory is filtered by the trajectory evaluation function,and the trajectories that violate the constraint conditions are eliminated.Then,by predicting the trajectory of the obstacle vehicle,the collision detection between the planned trajectory and the predicted trajectory of the obstacle vehicle is performed,and finally the optimal trajectory without collision is transmitted to the control module.(3)Taking into account the speed changes during the driving of the vehicle,this paper uses PID to control the longitudinal position and speed of the vehicle.First,build the vehicle longitudinal driving dynamics and longitudinal braking dynamics model,and use Carsim and Simulink to complete the calibration of the vehicle motor accelerator and brake,and finally complete the construction of the longitudinal control module.(4)In this article,the driving environment is highway,and its main feature is fast driving.Based on this feature,the controller should have high real-time performance in addition to ensuring control accuracy.Therefore,LQR is used to achieve lateral control in this article.Firstly,a two-degree-of-freedom dynamic model of the vehicle is established,and the linear differential equation of the lateral error in the Frenet coordinate system is derived.Taking into account the curvature of the road,feedforward control is introduced in LQR to eliminate the steady-state error caused by the road curvature.In addition,a prediction module is added before the control module to improve the real-time performance and control accuracy of the control algorithm.(5)Through Carsim and Simulink to build a simulation platform,the planning module and the control module are combined,and the simulations are carried out under three working conditions of car following,straight lane change and curve lane change.The results show that the planning method designed in this paper can still accurately plan the optimal trajectory in a curve environment.The horizontal and vertical control algorithm combined with LQR and PID has high control accuracy and can ensure the stability of the vehicle during driving.