Thigh-Based Normalized Time Estimation for Adaptive Swing Control of Prosthetic Knees

Abstract

This study introduces a novel state machine–based controller for powered prosthetic knees. The controller relies on estimating a normalized time, computed from the residual thigh motion, to regulate the swing motion of the prosthetic knee. This thigh-based normalized time (TNT) estimation method is lightweight and robust, requiring only a minimal set of parameters and enabling seamless adaptation to different walking speeds. Experiments were conducted with three able-bodied participants walking at speeds ranging from 0.1 to 1.3 m/s, comparing TNT with a state machine–based controller using a preferred fixed duration (PFD) for the swing phase. The results showed that TNT naturally scaled swing duration and velocity with walking speed, reducing duty-cycle and step-length asymmetry compared with PFD. Prosthetic knee joint power analysis further revealed reduced peak demands across walking speeds with TNT, suggesting that the controller better leverages natural pendular dynamics. Participants generally preferred TNT, especially at speeds deviating from their preferred walking pace. Overall, these findings highlight TNT's potential to enhance adaptability, energy efficiency, and user comfort in powered prosthetic knees.