SONAR BEHAVIOR-BASED FUZZY CONTROL FOR A NAVIGATION TECHNIQUE OF AN INTELLIGENT MOBILE ROBOT

摘要

This paper describes the sonar behavior-based fuzzy control for a navigation technique of an intelligent mobile robot. The fuzzy controllers are designed for avoid-obstacle and move-to-point behaviors. Fuzzy controller of avoid-obstacle behavior receives data from five sonar sensors and gives translation and rotation velocities as two outputs. Fuzzy controller of move-to-point behavior receives error values and change in error of distance, and sends translation and rotational velocities as the outputs. There are 46 fuzzy rules for avoid obstacle behavior and 10 fuzzy rules for move-to-point behavior. Behavior-based approach is implemented as Subsumption architecture. The highest level is a task behavior; move to point and wander behaviors. The next lower level is an avoid obstacle behavior. The lowest level is an emergency behavior, which has a highest priority than other. This technique is applied as a learning part of an intelligent mobile robot. The results from three experiments shown that mobile robot has ability to avoid obstacle and move to target successfully.