An image based path planning scheme for exploration rover

Various kinds of space explorations have been done for decades in order to understand the origin of the solar system and life as well as to seek natural resources in space. Especially, rover missions which can perform detailed planetary surface exploration are under planning around the world. However, rover mission still has difficulty in its autonomy in spite of its potential expectation. This paper proposes a novel autonomous path planning system for efficient exploration. The proposed system contains two topics: object recognition and path planning. Rocks and craters are recognized as obstacles by image processing. In the proposed path planning, the traversability of terrain is estimated by considering the texture of image and inclination estimation using Fuzzy logic. The path is calculated by Dijkstra´s algorithm. Since the innovation of this study is computational efficiency, attempts to reduce computation are shown in every step of the proposed system. Field tests in indoor and natural environment show the effectiveness of the proposed scheme. The rover can recognize natural environment and avoid obstacles safely.