抄録
This paper deals with locomotion selection that is appropriate for situation of a robot. Falling Risk is derived from errors and environment information by using Bayesian Network. The robot evaluates the Falling Risk as an indicator of uncertainty. Locomotion selection during walking is modeled as Semi-Markov Decision Process and most appropriate action is selected by using the greedy algorithm. As a result, the robot moves in the environment that is difficult to travel in one action, maintaining the maximum moving efficiency.
