The effect of model accuracy and thruster saturation on tracking performance of model based controllers for underwater robotic vehicles: experimental results

This paper reports an experimental investigation of the effect of two specific errors on a class of model based controllers for the low-speed maneuvering of fully actuated underwater vehicles. First, we review previously reported studies and a commonly accepted simplified plant model that has been experimentally validated for this class of vehicles. Second, we review a family of associated model-based nonlinear controllers, both fixed and adaptive, for trajectory tracking, as well as the commonly employed proportional-derivative (PD) controller. Third, we report an experimental evaluation of the performance of this entire family of controllers in the presence of two commonplace modelling errors: plant parameter mismatch, and thruster saturation. In the presence of plant model parameter errors, the fixed model based controllers often perform worse than PD control, and the adaptive controllers "learn" the correct parameters to provide asymptotically superior performance. In the presence of unmodelled thruster saturation, the adaptive controllers performs worst, and the model-based controllers perform no worse than PD.