Obstacle-recognition Method for Soccer Robot Based on Human Selective Attention

Abstract

In the Robocup soccer competition, all participating robots must operate completely autonomously by using their own sensors to recognize information about their environment and determine what to do next to win the match. Therefore, real-time recognition of obstacles in constant motion is important. Considering the processing resources of the robot, it is difficult to achieve real-time recognition by using conventional computational methods. Therefore, inspired by the mechanism of selective attention in the human visual system, we developed an obstacle-recognition method. With this method, multiple obstacles can be recognized, and successful real-time recognition of multiple obstacles is possible by significantly reducing the required processing resources.