Human Movement Compensation Control for Supernumerary Limb in Overhead Support Task: A Non-Cooperative Game Theory Approach

The overhead support task is a crucial component within the realm of Supernumerary Limb (SRL) applications. However, SRL's performance in overhead support is inevitable hindered by human movement as SRL is worn on human body. Therefore, integrating human intention and develop efficient and natural forms of compensation method is essential in overhead support task. This paper proposes an adaptive controller based on non-cooperative differential game theory-a framework which can provides mathematical model among multiple agents and minimize their objectives-to compensate the human movement in overhead support scenario. This approach first constructs a non-cooperative model to represent how human and SRL execute overhead support task. Then, we estimate the control strategies adopted by human based on the human movement. Finally, the adaptive controller adjusts the feedback gain of SRL to reduce reaction force on human and accomplish overhead support task. Simulation and experiments involving human participants demonstrate that these properties coordinate effectively with human in overhead support task.