Multi-sensor robot navigation method based on virtual environment and reinforcement learning

Abstract

The invention provides a multi-sensor robot navigation method based on virtual environment and reinforcement learning and belongs to the technical field of robots. According to the method of the invention, environmental information obtained by a radar, a camera, a compass and a speed measuring code disc sensor is adopted as input, and dimensionality reduction is performed on the image data captured by the camera through a convolution technique; a navigation strategy is pre-trained in a virtual environment, so that problems such as time-consuming and labor-consuming performance and insecurity during a robot physical training process can be avoided; and the navigation strategy which enables a robot to autonomously navigate, smoothly avoid obstacles, and efficiently reach a target position inthe virtual environment is obtained.

Description

technical field [0001] The invention belongs to the field of robot technology, in particular to a multi-sensor robot navigation method based on virtual environment and reinforcement learning. Background technique [0002] With the continuous progress of science and technology and the continuous development of intelligence, robots have been used more and more, and they have been more involved in daily production and life to assist or replace manpower. In the field of robotics, robot autonomous navigation technology is particularly indispensable and is one of the core technologies of mobile robots. Its purpose is to efficiently reach the target point without collision, human participation and other guidance. [0003] At present, autonomous navigation robots mostly rely on map creation and map planning, or traditional navigation based on sensor fusion. SUMMARY OF THE INVENTION [0004] In order to solve the problems that the algorithm is too complex in the robot navigation p...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problem that the algorithm in the robot navigation process is too complex and the training of the learning network takes too long in the prior art, a multi-sensor robot navigation method based on virtual environment and reinforcement learning is proposed.

Images