Design of auto obstacle avoidance system based on machine learning under the background of intelligent transportation

Abstract

With the process of urbanization and the increase in car ownership, traffic problems are becoming increasingly prominent. In order to improve traffic mobility and improve traffic safety, a machine learning based autonomous obstacle avoidance system was studied and designed in the context of intelligent transportation. Design an obstacle avoidance hardware system consisting of a tracking sensor module, an intelligent patrol module, an obstacle avoidance sensor module, and a motor module. Through the coordination and cooperation of multiple modules, the adaptive ability of the obstacle avoidance system is improved. On the basis of hardware design, a road coordinate system is established, and the lane‐changing path is planned with the longitudinal, lateral distance and speed of the ego vehicle and the preceding vehicle as input, and the vehicle steering and lane‐changing control is completed using the front wheel angle of the ego vehicle as the control quantity. The model predictive control method is used for obstacle avoidance trajectory planning. Based on the obstacle avoidance path planning results, the reinforcement learning method is used to design the vehicle's autonomous obstacle avoidance early warning to improve the efficiency of obstacle avoidance. The experimental results show that the designed system can maintain the lateral stability of the vehicle under continuous steering conditions, and the fit between the path tracking and the reference path is better, that is, the vehicle obstacle avoidance control effect is better; the convergence speed is faster. The vehicle autonomous obstacle avoidance warning time is short, which can ensure the safety of the vehicle to the greatest extent. This research achievement will provide important support for the development and practical application of intelligent transportation systems, and promote innovation and progress in the transportation field.