An EEG Based Dynamic Shared Control Method For Online Teleoperation

Shared control is a common method for human and machine cooperation in the teleoperation. Existing methods scarcely consider the effect of the performance of operators. This paper proposes a dynamic shared control approach based on the operator's states. A new indicator named Performance of Teleoperator (PoT) is introduced. By using the BP Neural Network, a model of PoT is trained, which represents the performance of the human teleoperator from the view angle of the brain EEG signals. Meanwhile, a dynamic shared controller is designed. PoT is used to balance the control weight between instructions from the teleoperator and the autonomous controller. The experimental results demonstrate the effectiveness and advantages of the proposed method.