Neurobionics Lab

Sensorimotor Control of Impedance

Example of a subject’s “learning curve” trying to match the target stiffness value (red line).
Subjects don’t have access to live feedback and must rely solely on information from the previous trial to make adjustments, just like playing darts.


Currently available lower-limb wearable robotics, while functional, often fail to accurately replicate human biomechanics.  For example, joint impedance, which is a fundamental characteristic of safe and accurate mechanical interaction, is largely absent from market-available assistive technologies.  Joint impedance is a property of the human muscular system which describes joint behavior as a function of inertia, damping, and stiffness.  At present, the importance of impedance in human neuromotor control is not completely understood, but it is thought to play a fundamental role in optimizing environmental interaction.  The objective of this work is to determine the extent to which healthy subjects can consciously control and modulate their own joint impedance.  The accurate characterization of this ability in able-bodied subjects will inform the development of more robust and naturalistic prostheses and other wearable devices.


The purpose of this work is to elucidate the accuracy and precision of mechanical impedance modulation in able-bodied subjects.  We can quantify this ability by using our lab’s custom dynamometer to calculate subject’s joint stiffness as they repeatedly attempt to match their own stiffness to a target stiffness value.  Stiffness can be adjusted by co-contracting the muscles surrounding a joint.  Due to the mathematics involved, the stiffness values cannot be rendered in real-time, so subjects are not able to make continuous adjustments.  Instead, discrete adjustments are made from trial to trial, in a paradigm known as feed-forward regulation.  The “learning curve” we are able to generate from the complete series of trials provides insight into subject’s ability to actively match and maintain the desired target stiffness.  This use of this feed-forward model is justified by applying the same technique to torque and position matching tasks, which have successfully been the object of more standard live-feedback protocols.

Contibutors: Alexander Wind, Elliott Rouse


Wind, A. M., and Rouse, E. J. (2020) “Neuromotor Regulation of Ankle Stiffness is Comparable to Regulation of Joint Position and Torque at Moderate Levels.” Scientific Reports, 10(1), 1-9.