Research On Planning And Control Algorithm For Autonomous Vehicles

In today’s intelligent era,artificial intelligence technology ushered in a boom of research.Autonomous driving,as one of AI technologies,has attracted wide attention.Autonomous driving technology plays a key role in reducing traffic accidents,reducing energy consumption and improving travel efficiency.Autonomous driving technology is the cross integration of many functional technologies,including perception technology,planning technology,control technology and so on.The perception technology is responsible for sensing and modeling the surrounding environment using sensors;The planning technology is responsible for making decisions on the driving route and motion trajectory of the vehicle according to the environmental model;The control technology generates the executable control signal of the vehicle according to the planning results,so that the vehicle can drive according to the planning decision.Planning and control technologies directly affects the quality of vehicle driving behavior,manifests the degree of intelligence of automatic driving vehicles,and is very important in the automatic driving system.In this paper,the planning and control technologies in automatic driving are studied,and simulation experiments are designed to verify the effectiveness of the proposed algorithm.The specific contents include the following:(1)The sequential programming algorithm is used to solve the vehicle planning problem.The algorithm is logically divided into global planning and local planning.Global planning is route planning,which is responsible for searching the route from the current road to the road where the target location is located and obtaining the road sequence.After the global planning is completed,the local planning is responsible for solving the vehicle trajectory planning in the local scene,generating the feasible smooth trajectory of the vehicle,and finally the vehicle is controlled by the control algorithm.The algorithm state machine is designed for different scenarios that local planning needs to deal with,and the planning algorithm most suitable for the current scenario is designed and used in different scenarios.(2)Aiming at the defects of low efficiency and kinematic infeasibility of traditional planning algorithm,a corresponding improved algorithm is designed.For the RRT algorithm in unconstrained scenario,the probability distribution is generated according to the initial solution to guide the sampling.Select the sampled vehicle control signal for forward trajectory deduction to ensure the kinematic feasibility of the search results.For the state lattice algorithm in the constrained scene,the spiral curve is used as the model to generate a smooth curve that meets the boundary condition constraints,and then the optimal trajectory in the local scene is obtained through the graph search algorithm.(3)The vehicle controlling problem is decoupled.The vehicle transverse control is based on pure pursuit algorithm,and the longitudinal control is based on traditional PID algorithm,which reduces the complexity of the algorithm and is easy to implement.A pure pursuit algorithm for adaptive selection of forward-looking distance is designed.The indexes such as vehicle speed and tracking effect will affect the selection of forward-looking distance to ensure the control effect of vehicles in different states.In this paper,the designed planning and control algorithm are realized in the simulation environment,and the simulation experiment is carried out.The experimental results show that the algorithm designed in this paper can effectively complete the planning and control task of autonomous vehicles.