Optic flow-based navigation system for planetary rovers

This paper proposes a novel optic flow-based navigation system for planetary rovers. Optic flow is a vector field of relative velocities between the camera mounted on a rover and environments, and it is often utilized for motion estimations or relative distance estimations. However, only either one of them can be estimated in general optic flow processing. The method proposed in this paper enables both of the estimations by utilizing an image segmentation technique in computer vision and applying Wide-Field-Integration (WFI) of optic flow for robust estimations. Thus, both obstacle avoidance and avoidance of getting stuck, which are key technologies for planetary exploration rovers, are realized in the method. Furthermore, since the estimation is accomplished in linear processing, a real-time estimation is possible by onboard computers. The effectiveness of the proposed method is examined through numerical simulations, considering sensor noises and the shape of obstacles.