An EMG-based force control system for prosthtic arms

This paper proposes a novel methodology for the control of prosthetic arms. The proposed controller predicts the intended forces at the tip of the prosthetic arm using an artificial neural network fed with electromyogram signals from the residual arm muscles and kinematic data from the prosthetic arm joint. The desired predicted force is then implemented at the prosthetic arm tip using a classic robot force control strategy. The proposed controller is evaluated on healthy subjects using a master-slave telerobotic test-bed.