Recursive penalty method to estimate a stable velocity profile for an autonomous vehicle

A driving velocity decision algorithm is used as an auxiliary means for estimating the velocity at which an autonomous vehicle appropriately follows a predetermined path, or for estimating a safe velocity of a remotely controlled robot. To find a reliable velocity decision algorithm for an autonomous vehicle moving over a rough terrain, it is necessary to satisfy the conditions such as the roughness of the terrain, the dynamic characteristics of the vehicle, and safety criteria. Conventional methods usually do not consider these conditions at the same time, and some methods do not estimate a perfect safe driving velocity due to a probabilistic velocity estimation technique. In this study, we propose a new velocity decision algorithm that satisfies safety criteria while considering the roughness of the terrain and the dynamic characteristics of the vehicle. To verify the proposed method, a driving simulation of a 4 × 4 vehicle model with various roughness characteristics was carried out, and the results of the safe velocity estimation were compared with those of the existing method.