Memory based environment recognition system for mobile robots with omni-directional images

Autonomous mobile robot needs various applications such as a control system, environment recognition, decision making, self localization, mapping, and obstacle avoidance. We developed the environment recognition system for wheeled mobile robot because these applications are mostly based on the environment recognition. In our previous research, we developed the wheeled mobile robot with passive linkage mechanism for outdoor environment and evaluate the mobility. And, we proposed environment recognition system using passive linkage mechanism data analyzed self-organizing map. This paper describes our proposed method which is the environment recognition system using the image processing. Using our proposed method, the environment recognition system could succeed self localization and the number of templates is decreased because our method using self organizing map has interpolation function.