Date of Award

2019

Degree Type

Thesis

Degree Name

Master of Science in Mechanical Engineering

Department

Washkewicz College of Engineering

First Advisor

Richter, Hanz

Subject Headings

Electrical Engineering, Engineering, Mechanical Engineering, Rehabilitation, Robotics, Robots

Abstract

The CSU 4OptimX exercise robot provides a platform for future research into advanced exercise and rehabilitation. The robot and its control system will autonomously modify reference trajectories and impedances on the basis of an optimization criterion and physiological feedback. To achieve this goal, a robust impedance control system with trajectory tracking must be implemented as the foundational control scheme. Two control laws will be compared, sliding mode and H-infinity control. The above robust control laws are combined with underlying impedance control laws to overcome uncertain plant model parameters and disturbance anomalies affecting the input signal. The sliding mode control law is synthesized based on a nominal plant model due to its inherent nature of overcoming unspecified, un-modeled dynamics and disturbances. Implementation of the H-infinity control law uses weights as well as the nominal plant, a structured parametric uncertainty model of the plant, and a model with multiplicative uncertainty. The performance and practicality of each controller is discussed as well as the challenges associated with attempts to implement controllers successfully onto the robot. The findings of this thesis indicate that the closed loop controller with sliding mode is the superior control scheme due to its abilities to counter non-linearities. It is chosen as the platform control scheme. The 2 out of 3 H-infinity controllers performed well in simulation but only one was able to successfully control the robot. Challenges associated with H-infinity control implementation toward impedance control include determining proper weight shapes that balance performance and practicality. This challenge is a starting point for future research into general weight shape determination for H-infinity robust impedance control.

COinS