EMG-driven model-based knee torque estimation on a variable impedance actuator orthosis

Lin Liu, Markus Lüken, Steffen Leonhardt, Berno J.E. Misgeld

2017 · 16 citations · 13 references

Concepts

Abstract

Development of movement assistance devices used in rehabilitation must take into account the variable human-machine interaction. To this end, electromyography (EMG) sensors detecting the muscle activation are regarded as a feasible method. In this paper, we focus on torque estimation of the human knee joint. The joint motion analysis was captured through the developed EMG sensor, and subsequently torque estimation by using an EMG-driven model. For calibrating the EMG-driven model, the experimental torque was computed by employing the inverse dynamics that have been ensured within a well-established leg-orthosis system. Here, an impedance-controlled variable stiffness actuator has been implemented. To realize a wide range of calibration tasks, we compared the knee torque by the EMG-driven model and the inverse dynamics through setting different impedance controller gains. The proposed calibration results will provide a support towards the applications in the bio-feedback based impedance control.

References

13