Neuromechanical Model-Based Corrective Torque Estimation During Weight Shifting in Lower Limb Amputees

Abstract

As a consequence of limb loss, unilateral lower limb amputees (LLAs) exert additional flexion torque in the intact limb during postural balance, leading to other secondary complications. A prosthetic device with torque delivery is required to avoid such complications. However, estimating desired torque is an acknowledged challenge in this domain. Motivated by the above, this paper focuses on estimating the corrective torque required from an assistive device to balance the individual while performing weight-shifting exercises. Further, a healthy individual’s inverted pendulum (IP) model is explored to model LLA’s dynamics during weight-shifting. The model and the control strategy are validated using forward and backward body lean angles recorded from healthy and LLA individuals. Finally, a robust Proportional Integral Derivative controller is designed to estimate the corrective torque for maintaining the postural balance during weight-shifting. Along with the IP dynamics, a suitable control law based on Coefficient Diagram Method achieves the experimentally recorded body lean angle. The proposed control scheme is validated through simulation studies to evaluate the tracking performance in two different biomechanically relevant IP postural models. The findings assist in determining the external torque required to maintain the postural balance during gait initiation and compensatory measures for pathologically reduced torque. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Gait initiation is a transition between vertical posture and gait. Since vertical stance is inherently unstable, a postural adjustment occurs before gait initiates, propelling the center of mass forward and towards the first stance leg. The absence of flexor torque in the amputated limb causes a loss of postural stability during weight shifting activities. In the case of powered exoskeletons and prosthetic devices, a prior estimation of corrective torque through a computational approach helps to select a suitable actuator that provides an appropriate joint torque while maintaining the postural balance. As a result, an efficient computational model of gait initiation that accounts for neuromuscular system components before designing assistive devices is necessary. The controllers further provide corrective action to the assistive device to obtain robust and optimal coordination between the prosthetic element and gait. Therefore, this study demonstrates the significance of neuromuscular controller and assists in determining how much corrective torque from an external device is required to maintain postural stability and adaptability during weight shifting exercises. The effectiveness of the neural controller is also tested for real-time data to compensate for the effect of external disturbance and noise, thereby achieving a proper and controlled degree of movement in amputee users. These findings could provide one of the fundamental aspects of gait initiation model paradigms of compensatory measures for pathologically diminished torque and contribute to developing low-cost prosthetic solutions for the benefit of the rehabilitation society in developing countries.