Path planning system for a mobile robot using self-organizing map

Abstract

In order to navigate a mobile robot using environmental information (distances to the obstacles, visual image and so on), many kinds of variables should be handled to recognize the situation of the robot, estimate the position and make its behavior. It is not easy to decide the robot's action using a lot of variables on-line and in real-time. So the feature abstraction from the obtained environmental information is inevitable and one of the most important problems to be solved in order to realize the autonomous mobile robot. The self-organizing map (SOM) is known as one of the effective methods to extract the principle feature from many parameters and decrease the dimension of parameters. These features are suitable to the path planning system of the mobile robot.. In this paper, a navigation and a position estimation algorithm based on SOM are introduced into the path planning system. The efficiency of the system is examined and discussed through the collision avoidance and position estimation experiment using the distances obtained from the ultrasonic range sensors and the magnetic field data.