Investigating Capturability in Dynamic Human Locomotion Using Multi-body Dynamics and Optimal Control

Abstract

An important goal in the development of advanced prosthetic devices is to enhance the stability of prosthetic gait and eventually to augment safety. For this, a fundamental understanding of stability and stabilization mechanisms of human walking motions is crucial. The objective of this work is to evaluate if the stability of human walking can be linked to the concept of N-step Capturability developed in robotics. The main idea of this concept is represented by the Instantaneous Capture Point (ICP), which indicates the position on the ground where a two-legged walking system would have to place the next step in order to come to a complete stop and the respective N-Step Capture Regions where a stop after N steps can be reached. While a human walker does not precisely step onto the ICP since this would be inefficient, we hypothesize that the location of this point can be used to parameterize the location of the foot position. Experiments were performed in a gait lab to record the kinematic data of healthy human gait on level ground. The human body was modeled as a multi-body system composed of eight rigid bodies that represent the pelvis and three-segmented legs as well as the upper body merged into a single trunk segment. Using optimal control methods, joint angle trajectories of the multi-body model were generated that best fit the experimental data while considering the kinematic and physiological constraints of the human body. The trajectory of the ICP was computed using the kinematic and kinetic data of the multi-body model that derived from the solution of our optimal control problem. The results show that the ICP is directly approached by the swing foot during swing phase and suggest a correlation between the foot placement strategy in human walking and N-Step Capturability.