Date of Award


Degree Type



Electrical and Computer Engineering

First Advisor

Simon, Daniel

Subject Headings

Prosthesis, Artificial legs, Robotics, Biomechanics, Biomedical engineering, Biomechanics Biomedical Engineering Electrical Engineering Engineering Robotics


Typical tests of prosthetic legs for transfemoral amputees prove to be cumbersome and tedious. These tests are burdened by acclimation time, lack of repeatability between subjects, and the use of complex gait analysis labs to collect data. To create a new method for prosthesis testing, we design and construct a robot that can simulate the motion of a human hip. We discuss the robot from concept to completion, including methods for modeling and control design. Two single-input-single-output (SISO) sliding mode controllers are developed using analytical and experimental methods. We use human gait data as reference inputs to the controller. When doing so we see the problems associated with the gait data that make it unfit for use as reference data. We apply a smoothing algorithm to correct the gait data. The robot is evaluated based on its ability to track the gait data. Despite proper tracking of the reference inputs, operating the robot with a passive prosthesis shows that the robot cannot adequately produce the ground reaction force (GRF) of an able bodied person. We devise a novel method to control GRF of the robot/prosthesis combination based on the way that human subjects walk with a prostheses. When walking with a prosthesis, users compensate for the deficiencies of the prosthesis by modifying their gait patterns. To simulate this we use an evolutionary algorithm called biogeography-based optimization (BBO). We use BBO to modify the reference inputs of the robot, minimizing the error between the able-bodied GRF data and that of the robot walking with the passive prosthesis. Experimental results show a 62 decrease in the GRF error, effectively showing the robot's compensation for the prosthesis and improved control of GRF