Abstract: Aiming at the problem of obstacle avoidance for mobile robot with the platform of ASR mobile robot, a novel approach based on behavior for dynamic obstacle avoidance was put forward, which divides the robot behavior during the entire run into four behaviors, including tendency to target behaviors, obstacle avoidance behavior, going along the wall and emergency obstacle avoidance behavior of four acts, including obstacle avoidance behavior. A multi-sensor- based autonomous mobile robot obstacle avoidance mechanism was designed. A recursive median filter-based method was proposed to avoid crosstalk by grouping sampling or interval sampling technologys. The data was improved in both space and time continuity. I}he ultrasonic signal crosstalk and other interference random signals were effectively re- duced. Experiments based on both simulation platform and ASR robot verified the effectiveness of the proposed approach.

Key words: ASR mobile robots, l3chavior analysis, Fuzzy logic, Dynamic obstacle avoidance, Path planning