Research On Automatic Parking System Based On Path Planning Algorithm

With the rise of cutting-edge technologies such as artificial intelligence and big data,autonomous driving technology has been developing rapidly,and some advanced driving assistance functions have been gradually produced on various types of vehicles.Among them,the automatic parking function is favored by many users because of its convenience.Automatic parking technology can not only effectively solve the environmental problems caused by narrow parking space and less time,but also solve the parking troubles for many drivers in need.Path planning is an important part of automatic parking,the quality of its algorithm directly determines the parking effect,and the adaptability of the planned route will also directly affect the success rate of automatic parking.In this paper,aiming at the existing research on automatic parking path planning,the parking path planning method that meets the vehicle’s kinematic constraints and obstacle avoidance constraints is developed and designed for the relatively high requirements on the initial posture and parking environment of vehicles.At the same time,based on the planned path,the tracking system is built and simulated.Firstly,the environmental awareness module of automatic parking is studied.the ultrasonic sensor is selected to sense the parking space and obstacle information,and the detection experiment is actually carried out to verify its effect and accuracy.At the same time,the kinematic model of the vehicle is established according to the Ackerman steering principle,and the equation of the low-speed motion is obtained through analysis and calculation.Secondly,the research on path planning is carried out in view of various complex scenarios and different parking modes with many limitations in parking environment.The whole design process of planning module is defined,and some basic principles and preconditions are preprocessed for later explanation.Two important sub-algorithms in the planning algorithm are described in detail: the simple connection algorithm and the collision detection algorithm between the path and the obstacle.Different planning schemes for parallel parking and vertical parking are given,including single-step parking and multi-step parking.On this basis,the exploration mechanism is designed to strengthen the redundancy of the planning algorithm.When the parking environment changes,it can be adjusted in time to ensure the success of the planning.Considering that some scenarios need the front-entry parking method,the forward parking route is proposed.In addition,in order to improve parking safety and adaptability to different parking environments,several optimization strategies are proposed.Finally,the path tracking control module is built and the simulation test of automatic parking is completed.By comparing several control strategies,the fuzzy controller is selected as the main tracking control method according to the characteristics of parking,and the membership function and rule base of fuzzy control are described in detail.In order to solve the problem of curvature discontinuity in the planned path,a corresponding tracking model is proposed.According to the parking route given by the planning module,MATLAB/Simulink software co-simulation was used to verify the effect of path planning and tracking control.The results show that the proposed algorithm has good parking accuracy and robustness,and especially it has strong adaptability to the complex parking environment.Therefore,the proposed planning algorithm has important application value.