A Deadband-Based Method for User Effort Reduction in Human-Robot Shared Control ?

摘要

In a human-robot collaboration framework, the authority over the execution of tasks can be shared between the human user and the automatic controller. In case of disagreement, the human user may exert a significant amount of effort opposing the robotic counterpart. Therefore, a prompt assessment of user intent is of the utmost importance in shared systems. In this paper a novel strategy, inspired by anti-windup techniques, is proposed to adjust the component of automatic control according to the sensed human effort. A haptic human-robot interface is developed and the method is validated in a shared control application in which both the human user and the automatic controller can steer a simulated ground vehicle. Evidence is shown that the proposed method can reduce the required human effort, with performance comparable to the state-of-the-art.