Evaluation of model predictive control and adaptive model predictive control for a selected generalisable daily task including uncertainties

Robots need to interact with their surroundings, sometimes without having complete knowledge about it. In this research, different daily activities have been studied to find a general case that encompasses most of the situations. Subsequently, according to the selected scenario, Model Predictive Control and Adaptive Model Predictive Control methods are chosen and tested in simulation. The performance of the controllers is evaluated in different trials including model uncertainties and disturbances. Finally, the controllers are validated in a real case using a Franka Emika robot.