Unscented transform-based dual adaptive control for mobile robots : comparative analysis and experimental validation

Abstract

Adaptive control involves both estimation and control, which are generally interdependent and partly in conflict.

Yet, the majority of adaptive controllers separate the two by assuming that certainty equivalence holds, even if this is not the case. In contrast a dual adaptive controller, based on the idea postulated by A. Fel’dbaum in the early 1960s, aims to strike a balance between estimation and control at all times. In this manner, the control law is a function of the estimates’ uncertainty, besides the estimates themselves, thereby leading to improved

control performance. Few such controllers have ever been implemented and tested in practice, especially within the context

of intelligent control, and to the best of our knowledge none on mobile robots. This paper present two novel dual adaptive neural control schemes for the dynamic control of mobile robots in the presence of functional uncertainty. Furthermore, by means of realistic Monte Carlo simulations and real-life experiments, a thorough comparative analysis is performed. A notable novel contribution of this work is the use of the unscented transform within the context of dual adaptive control, aimed at improving further the performance of the system.