価値反復を用いた移動ロボットによる屋外ナビゲーション

抄録

We examine our software package for mobile robot navigation in the real world. This software uses value iteration for updating a decision making policy in the whole state space. Though the computational cost of this method is too large for current computers, we expect that it will be standard in future. In the experiments, we use this package with a laptop computer on a mobile robot. Though the calculation time is still a problem, it works on the mobile robot for the first time.