Generation of image template with variable matching-time for autonomous mobile robot with vision

The vision has been focused on by many researchers as a powerful sensor for an autonomous mobile robot because of the great ability to collect the abundant information about the situation around the robot. Now, we have aimed to establish the methodology of action acquisition with the realtime search for the robot with vision. In the case of action acquisition with the real-time search, it is necessary that the computational amount of the image processing varies according to the situation around the robot. Moreover, the efficient and robust image processing is required. These are because the available time for action acquisition varies according to the situation and it is limited. In this research, we aim a new image processing method which satisfies the above two features required to the robot vision, and we propose a generation method of the image template for the self-position estimation of an autonomous mobile robot based on the above ideas as an example in this paper.