A Research On Path Planning And Tracking Control For Intelligent Vehicle’s Autonomous Parking Systems

The popularity of automobiles has provided humans with a convenient mode of transportation and way of life.However,as the number of cars continues to increase,parking spaces in cities have become very compact due to architectural constraints.For drivers who lack parking experience,maneuvering a car into a parking space in a narrow area can be quite challenging.Autonomous parking systems,relying on technologies such as environmental perception,path planning,and vehicle control,can effectively solve the problem of difficult parking.Currently,autonomous parking technology still has certain limitations,making it difficult to achieve its function in some vulnerable scenarios.This paper aims to conduct research on the technology of path planning and tracking control of intelligent vehicle’s autonomous parking systems for the disadvantaged scenario of parking in narrow spaces.The main research contents are as follows:1)Design a path planning algorithm based on the Inverted-Expanded Hybrid A~★algorithm for autonomous parking.This paper researches the parking path planning problem based on the grid-based cost map and the inflation collision detection algorithm.For narrow parking spaces,improvements are made to the path node expansion logic of the conventional Hybrid A~★path planning algorithm.By exchanging the start and end points of the path,the path nodes can be expanded from the interior of the narrow parking space in the opposite direction towards the open traffic lane,reducing the number of invalid path nodes and shortening the expansion time of the path nodes.Path planning simulations are conducted in various types of narrow-space parking spots.The results show that using the Inverted-Expanded Hybrid A~★path planning algorithm designed in this paper improves the real-time performance of parking path planning in narrow-space parking spots.2)Design a path tracking control algorithm based on model predictive control(MPC)for autonomous parking.This paper designs a linearized and discretized path tracking error control system based on the kinematic vehicle model.In the objective function,three variables,namely the error of vehicle state quantities,the error of control input quantities,and the error increment of control input quantities,are constrained to ensure that the vehicle can quickly and smoothly track the reference path.A kinematic vehicle model-based MPC path tracking controller is designed to realize longitudinal and lateral motion control for parking path tracking.The calculation logic of the steering angle reference value of the vehicle front wheels is designed based on the Pure Pursuit algorithm,enabling the MPC path tracking controller to adapt to continuously changing path curvature.Autonomous parking"planning-control"closed-loop simulation verification is conducted in narrow parking space scenarios.The results show that using the MPC autonomous parking path tracking control algorithm designed in this paper,the Car Sim vehicle model can successfully track the reference path and safely park in a narrow parking space.3)Conduct real-world validation of path planning and tracking control for autonomous parking.The Inverted-Expanded Hybrid A~★path planning algorithm and MPC path tracking control algorithm based on the kinematic vehicle model designed in this paper were deployed to the onboard computer.Real-world validation was conducted in a narrow parking space scenario,including vertical parking,parallel parking,and diagonal parking in both field and road environments.The results show that the path planning and tracking control algorithms designed in this paper can achieve safe entry of the test vehicle into a narrow parking space.