Research On Overtaking Path Planning For Autonomous Vehicles

Autonomous vehicle is an important participant in intelligent transportation systems.Among other things,path planning and path tracking technology have become two of the core technologies for realizing autonomous driving.Through the research on path planning and path tracking algorithm,designing safe,reliable and stable obstacle avoidance path for intelligent vehicles,realizing accurate obstacle avoidance of vehicles in different environments and reaching the target position smoothly have become the focus of current researches.In order to improve the optimization performance and path stability of the path planning algorithm,on the basis of classical artificial potential field(APF)algorithm,the thought of Bug algorithm and the theory of model predictive control(MPC),an integrated hybrid path planning algorithm and a model-predictive-control-based path tracking controller are established.Finally,the proposed hybrid algorithm is verified by a real experimental vehicle.The main work is as follows:Firstly,the models of safety distance and kinematics of autonomous vehicle are constructed.This paper analyzes the overtaking and lane changing process of autonomous vehicle,obtains the collision conditions between autonomous vehicle and obstacle vehicle in front,and establishes the minimum safety distance model and vehicle kinematics model,which provide a basis for the path planning and path tracking of autonomous vehicle overtaking process.Secondly,the path planning of overtaking of autonomous vehicle is constructed.In view of the advantages and disadvantages of traditional APF algorithm,the influence of physical characteristics of autonomous vehicles on path planning has been investigated.Furthermore,this section introduces vehicle kinematics parameters into the potential field function to improve the safety of the path planned by the APF algorithm,and designs a static obstacle avoidance scenario to simulate and verify the improved algorithm with vehicles featuring different masses and different driving speeds.At the same time,with a view of solving the problem that the APF-based path planning for obstacle avoidance of autonomous vehicles is always naturally easy to strapped by the local extreme value in complex dynamic driving environment,to enhance the optimization ability of the improved APF-based path planning algorithm,the thought of Bug algorithm has been employed.Taking the zero resultant force point as the bridge,and combining the improved APF algorithm with Bug algorithm,a hybrid path planning algorithm is proposed and verified by a variety of static obstacle avoidance scenarios.Then,with the aim of validation for the effectiveness and stability of the proposed integrated hybrid path planning algorithm under the scenario of dynamic overtaking process,a dynamic overtaking scenario with a stable-driving obstacle in front is designed,and the simulation results are compared with other existing literature.To verify the robustness of the proposed algorithm,a dynamic overtaking scenario with an unstable-driving obstacle in front is constructed.Then,on the basis of MPC theory,a smart tracking controller was established to achieve an autonomous path tracking behavior along with the planned optimal path.With the goal of ensuring the vehicle travels along the planned optimal overtaking path,the path tracking controller is designed by using the MPC theory to track the overtaking path,and the Car Sim / Simulink joint simulation platform is built to verify the tracking performance of the path tracking controller.Finally,an autonomous vehicle test platform is established to verify the practical availability of the hybrid path planning algorithm proposed in this paper.Taking the smart vehicle controlled by a microprocessor of STM32 as the research object,the path planning algorithm and path tracking controller are transformed into C code and embedded into the vehicle control unit.The distance between the autonomous vehicle and the obstacle vehicle in front is measured by the on-board ultrasonic sensor,and the optimal overtaking path of the autonomous vehicle is obtained and tracked by the vehicle controller.The simulation and experimental results show that,compared with the traditional APF and other existing path planning algorithms based on APF,the novel integrated hybrid intelligent path planning algorithm built by this paper performs an effective improvement on success rate,stability and safety during overtaking process.