Date of Award

Spring 1-1-2020

Degree Type

Dissertation

Degree Name

Doctor of Philosophy In Engineering Degree

Department

Engineering

First Advisor

Bogert, Antonie van den

Second Advisor

Dr. Anne Su

Third Advisor

Dr. Hanz Richter

Abstract

The method of trajectory optimization with direct collocation has the potential to extract generalized and realistic motion controllers from long duration movement data without requiring extensive measurement equipment. Knowing motion controllers not only can improve clinic assessments on locomotor disabilities, but also can inspire the control of powered exoskeletons and prostheses for better performance. Three aims were included in this dissertation. Aim 1 was to apply and validate the trajectory optimization for identification of the postural controllers in standing balance. The trajectory optimization approach was first validated on the simulated standing balance data and demonstrated that it can extract the correct postural control parameters. Then, six types of postural feedback controllers, from simple linear to complex nonlinear, were identified on six young adults’ motion data that was collected in a standing balance experiment. Results indicated that nonlinear controllers with multiple time delay paths can best explain their balance motions. A stochastic trajectory optimization approach was proposed that can help finding practically stable controllers in the identification process. Aim 2 focused on the foot placement control in walking. Foot placement controllers were successfully identified through the trajectory optimization method on nine young adults’ perturbed walking motions. It was shown that a linear controller with pelvis position and velocity feedback, suggested by the linear inverted pendulum model, was not sufficient to explain their foot placement among multiple walking speeds. Nonlinear controllers or more feedback signals, such as pelvis acceleration, are needed. Foot placement control was applied on a powered leg exoskeleton to control its legs’ swing motion. Two healthy participants were able to achieve stable walking with the controlled exoskeleton. v Results suggested that the foot placement controller helped decelerate the swing motion at late swing. In Aim 3, the trajectory optimization method was used to identify joint impedance properties in walking. Results of the synthetic study showed that relatively close impedance parameters can be identified. Then, a preliminary study was done to identify the ankle joint impedance properties of two participants at two walking speeds. The identified impedance properties were close to previous studies and consistent between different participants and walking speeds.

COinS