Virtual Muscle Reflex Control
Powered prosthetic limbs and exoskeletons are actuated by single-joint electric motors. In the human body, muscles are the actuators which produce this torque and consist of nonlinear properties and internal dynamics capable of actuating multiple joints simultaneously. If the torque produced by the electric motors of these powered prosthesis can be controlled with "virtual muscles", then the resultant movement would mimic the actual mechanical and dynamical properties of the human system. Though several control strategies have been proposed for prostheses, a Virtual Muscle Reflex (VMR) controller has shown promising results in both simulation and hardware.
Inputs to the model are the muscle lengths (determined by joint angles) and the muscle excitation signal generated by muscle reflexes. The muscle model produces a force, which can then be multiplied by the moment arm to obtain the torque that would drive the electric motors:
To mimic able-bodied locomotion, the control parameters can be optimized using a cost function which minimizes the difference between the torque exhibited by human test subjects and the torque generated by the model. We use, where subjects were longitudinally perturbed with Gaussian white noise within +/- 10% of the nominal belt speed [1,2]. Perturbations were used to evaluate the ability of reflexes to describe variations within and between gait cycles.
The goal is to simulate virtual muscles and reflexes in computer software and ultimately control the Parker INDEGO exoskeleton using VMR control at the hip and knee joints.
- Virtual muscle and reflex controllers are capable of describing human gait and responses to perturbation. Dynamic Walking, Camp Ohiyesa, MI. (abstract, poster, slides)
- Real-time virtual muscle control for prostheses and exokseletons. ISB Biannual Conference, Glasgow, UK (abstract, presentation)
 S. K. Hnat and A. J. van den Bogert. Inertial compensation for belt acceleration in an instrumented treadmill. Journal of Biomechanics, 47(15):3758 – 3761, 2014. (link)
 J. K. Moore, S. K. Hnat, and A. J. van den Bogert. An elaborate data set on human gait and the effect of mechanical perturbations. PeerJ PrePrints, 3:e1248, 4 2015. (link)