Abstract
In Japan, when the Great East Japan Earthquake occurred, a lot of robots were mainly used in order to cope with Fukushima nuclear accident and investigate disaster areas, and since then the development of rescue robots has been focused on further. We have paid attention to the mechanism of peristaltic crawling, and the rescue robot having this mechanism is more suitable for movement in narrow space or on irregular ground than others. In this research, we suppose that the peristaltic crawling robot which we have developed moves in a narrow pipe, and we propose a pattern for the fastest forward movement there and estimate its validity by using the simulation of Q-learning and Genetic Algorithm.
