An automatic parking method based on deep reinforcement learning

Abstract

The invention provides an automatic parking method based on deep reinforcement learning. The basic idea of ​​the invention is to use the powerful perception ability of deep learning to perceive the current state of the car, that is, to perceive the relative positional relationship between the car and the parking space, and then place the car Input the state of the state into the deep deterministic policy gradient model, use the decision-making ability of reinforcement learning to get the action that needs to be executed in this state, input the action that needs to be executed into the environment model to get the next state and reward function value, and then go through a reward mechanism Evaluate the quality of the action just performed, and guide the car towards the direction of the parking point, and repeat the previous operation in the next state until the model converges, that is, the car accurately parks in the parking space. Through the above design, the present invention solves the problem that the current automatic parking system cannot achieve real-time interaction with the environment, and cannot realize parking at any place and any angle.

Description

technical field [0001] The invention belongs to the technical field of automobiles, in particular to an automatic parking method based on deep reinforcement learning. Background technique [0002] With the continuous development of the automobile industry and the continuous increase of car ownership, the problem of difficult parking in cities has become more and more serious, and automatic parking technology has begun to develop. Automatic parking can not only free your hands and facilitate people's lives, but also reduce the parking pressure in the city and reduce the safety hazards in the parking process, which plays a very important role in urban construction. [0003] Now the main research of automatic parking system is to plan the parking path according to the parking space detected by the sensor, and then guide the car to follow the planned parking route according to the path tracking module. However, the parking process is a dynamic process, which requires continuous...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the above-mentioned deficiencies in the prior art, the present invention provides an automatic parking method based on deep reinforcement learning, which solves the problem that the current automatic parking system cannot achieve real-time interaction with the environment, and cannot realize parking at any place and at any time. Angle start parking problem

Images