An Adaptive Navigation Method for Wheeled Robots Based on Online Learning Mechanism

Abstract

The invention relates to an adaptive navigation method for a wheeled robot based on an online learning mechanism. First, the wheeled robot is controlled to learn the specified path, and the precise distance information between the surrounding environment and the wheeled robot is obtained through the SICK laser range finder. The mapping relationship is stored in a tree-structured knowledge base. After learning, the wheeled robot will continuously retrieve the constructed knowledge base and return to the motion control for navigation. At the same time, it implements background obstacle avoidance based on the obstacle avoidance algorithm of autonomously escaping from the obstacle area. Once obstacle avoidance is turned on, new environmental distance information and The mapping relationship of motion control information is updated to the knowledge base in real time. The invention improves the intelligence level of the wheeled robot, so that it has the self-adaptive ability to the complex dynamic environment, and the double closed loop of the online learning algorithm and the obstacle avoidance algorithm jointly improves the navigation efficiency.

Description

technical field [0001] The invention belongs to the technical field of mobile robot navigation. Based on the SICK laser range finder, the accurate distance information between the surrounding environment and the wheeled robot is obtained, an online learning mechanism is adopted to let the wheeled robot learn the navigation path, and a An obstacle avoidance algorithm that autonomously escapes from obstacle areas is used to avoid background obstacles, thereby realizing adaptive navigation of wheeled robots. Background technique [0002] Mobile robot navigation is a hot topic in the field of robotics, and using more streamlined sensors to obtain better paths for navigation has always been a common goal in the industry. [0003] In order to realize the adaptive navigation of the specified path, the traditional navigation method needs to rely on a large amount of sensor information, and at the same time, it needs to perform data fusion processing on various sensor information, wh...

AI Technical Summary

Problems solved by technology

[0005] The present invention aims at the above-mentioned navigation method based on a single learning mechanism, which is not adaptable to the dynamic environment, and the collision risk always exists, and the problem that the obstacle avoidance navigation based on the laser rangefinder must rely on other sensors to escape from the obstacle area. Provides an adaptive navigation method for wheeled robots based on an online learning mechanism, which solves the problems of insufficient robustness and objective accompaniment of collision risks in dynamic scenes based on single learning mechanism navigation

Images