Predictive Dynamics-Based Motion Control for the Rough-Terrain Locomotion of the Personal Vehicle Falcon-III

Abstract

A predictive control strategy aiming to reduce the impact generated on a personal vehicle during obstacle negotiation is proposed. Similarly to the motion executed by a cyclist when he/she is about to negotiate a step, the motion strategy uses a dynamic effect to temporarily reduce the load on the wheel of the vehicle that must negotiate the obstacle, thus reducing the impact force as well. Such motion control strategy was developed for the three-wheeled personal vehicle Falcon-III. The vehicle behavior using the predictive control strategy was first investigated for basic cases, such as when only one of the wheels had to negotiate an obstacle; later, the strategy was generalized by combining the motion derived for the basic cases. Robustness is achieved by the use of fuzzy logic to infer the displacement of the ZMP required to negotiate different obstacles. The predictive motion control strategy was compared with a conventional feedback attitude control strategy, and was shown to produce smaller impacts than simply using the feedback strategy.