A modified dynamic window approach in crowded indoor environment for intelligent transport robot

Abstract

We propose the modified dynamic window approach for indoor transport robot. The dynamic window approach is the popular collision avoidance algorithm that considers the heading direction of the robot, the distance to the obstacle, and the velocity of the robot. The objective function of the DWA thinks faster speed is better but it is very dangerous when there are many people around. They will feel discomfort with fast vehicle around them. This paper proposes that when the robot is in crowded place which means there are many dynamic obstacles around the robot, the robot move slow down to watch out the collision with them. It is implemented by using several fuzzy rules to determine the possible actions of the robot in DWA. Experiment results are given to validate our proposed method.