A Method of Autonomous Obstacle Avoidance for Agents Based on the Guidance of State Policy Knowledge Base

Abstract

The invention discloses an intelligent body autonomous obstacle avoidance method based on the guidance of a state strategy knowledge base, which endows the intelligent body with obstacle recognition and obstacle avoidance capabilities. The method includes five steps. Step 1 senses the obstacle avoidance environment, and judges whether obstacle avoidance measures need to be taken for the obstacle. If obstacle avoidance is required, the state of the obstacle avoidance environment is constructed; Knowledge base matching, if the matching is unsuccessful, go to step 3, if the matching is successful, then apply the knowledge base to output decision-making; step 3 based on the fuzzy neural network online output strategy for the new state, and add the state-strategy to the knowledge base; step 5 judge intelligence Whether the body reaches the target point, if not, repeat step 1, otherwise, end. The present invention is characterized in that the off-line state strategy knowledge base and the online fuzzy neural network are used to solve the obstacle avoidance problem, the knowledge base can realize incremental development during the obstacle avoidance process, and is suitable for quickly solving various obstacle avoidance problems.

Description

technical field [0001] The invention belongs to the technical field of mobile intelligent obstacle avoidance, and in particular relates to an autonomous obstacle avoidance method for an intelligent body guided by a state policy knowledge base. Background technique [0002] With the development of science and technology, mobile agents can replace human labor in the field of transportation. Although the path planning is provided in advance during the transportation process, the environment is usually complex and changeable, and there are interferences from static and dynamic obstacles. Therefore, the intelligent agent Autonomous obstacle avoidance capability is the key to ensure its smooth arrival at the destination. The autonomous obstacle avoidance of the agent is mainly divided into two steps. The first step is to judge whether obstacle avoidance is needed according to the obstacle information obtained within the perception range; the second step is to specify the correspon...

AI Technical Summary

Problems solved by technology

However, these methods are mainly based on online real-time calculations and do not use offline knowledge bases. When faced with complex scenes with dynamic obstacles, there are problems of large amount of calculation and poor timeliness.

At the same time, in the process of obstacle avoidance, the optimal route of the agent's complete travel process is not considered, resulting in a large number of redundant paths and even collisions.

Images