This paper describes a method of controlling the robot speed using a probabilistic occupancy map. If the robot is not sure that the target region is free, the robot has to observe it several times. We develop a method to determine an optimal speed considering the distance to the region and the uncertainty of map and observation.
