An intelligent obstacle avoidance method and system for an unmanned vehicle

Abstract

An unmanned vehicle obstacle avoidance system based on reinforcement learning, characterized in that the system includes a perception part, a decision-making part, a control part and an execution part. The perception part detects obstacle areas through single-line lidar to obtain obstacle information. The decision-making part determines whether the unmanned vehicle needs to avoid obstacles at the next moment based on the obstacle detection situation. If necessary, the control part inputs the state information of the sensing part into the reinforcement learning model. After trial and error learning, the model converges to a stable state and calculates the unmanned vehicle. The steering wheel angle required by the driver and the vehicle is executed by the autonomous vehicle execution part to achieve the obstacle avoidance function.

Description

technical field [0001] The present invention relates to an unmanned vehicle obstacle avoidance method and system, more particularly, refers to an unmanned vehicle intelligent obstacle avoidance method and system based on reinforcement learning. Background technique [0002] With the development of science and technology, unmanned vehicle technology has been in-depth development and research, realizing autonomous obstacle avoidance in complex and unknown dynamic environments is a key part of the practical application of unmanned vehicles. The traditional obstacle avoidance method uses expert experience programming, obtains its own position and attitude information and external obstacle information through on-board sensors, and then compares the vehicle's own information with external obstacles to calculate the steering wheel angle and speed required by the unmanned vehicle. These methods require certain expert knowledge, and their adaptive ability in complex and unknown envir...

AI Technical Summary

Problems solved by technology

[0019] However, the neural network has problems such as slow convergence speed, easy to fall into local minimum, poor selection of the number of hidden layers of the network, etc., and the learning effect is often related to the initial value, and it is difficult to converge to a unique optimal decision, which restricts reinforcement learning. further application of

Images