Multiphase and Multivariable Linear Controllers That Account for the Joint Torques in Normal Human Walking.

Abstract

OBJECTIVE: The objective of this paper is to investigate whether a small number of sequentially composed multivariable linear controllers can be used to recover a defining relation between the joint torques, angles, and velocities hidden in the walking data of multiple human subjects. METHODS: We solve a mixed integer programming problem that defines the optimal multivariable and multiphase relation between the torques, angles, and velocities for the hip, knee, and ankle joints. RESULTS: Using the data of seven healthy subjects, we show that the aforementioned relation can be remarkably well represented by four sequentially composed and independently activated multivariable linear controllers; the controllers account for [Formula: see text] (mean ± sem) of the variance in the joint torques across subjects, and [Formula: see text] of the variance for a new subject. We further show that each controller is associated with one of the four phases of the gait cycle, separated by toe-off and heel-strike. CONCLUSION: The proposed controller generalizes previously developed multiphase single variable, and single phase multivariable controllers, to a multiphase multivariable controller that better explains the walking data of multiple subjects, and better generalizes to new subjects. SIGNIFICANCE: Our result provides strong support to extend previously developed decoupled single joint controllers to coupled multijoint multivariable controllers for the control of human assistive and augmentation devices.