Defense Date

2019

Document Type

Thesis

Degree Name

Master of Science

Department

Biomedical Engineering

First Advisor

Dr. Dianne Pawluk

Second Advisor

Dr. Peter Pidcoe

Abstract

Intuitive prosthetic control is limited by the inability to easily convey intention and perceive physical requirements of the task. Rather than providing haptic feedback and allowing users to consciously control every component of manipulation, relegating some aspects of control to the device may simplify operation. This study focuses on the development and testing of a control scheme able to identify object stiffness and regulate impedance. The system includes an algorithm to detect the apparent stiffness of an object, a proportional nonlinear EMG control algorithm for interpreting a user’s desired grasp aperture, and an antagonistically acting impedance controller. Performance of a testbed prosthetic simulation used to controllably extrude pastes of different properties from a compliant tube was compared to that of the non-dominant human hand. The paste volume extrusion error and response time to perform the task were recorded for comparison. Statistical analysis using (GEE) and (TOST) suggests the prosthetic controller and human hand performed similarly along these metrics. Performance differences in the trials were more strongly correlated to tube type and repetition block. The results suggest that the developed controller allows users to perform a controlled squeezing task at a level comparable to the human hand with minimal training. It also suggests that a priori stiffness estimation acquired through quick palpations may be sufficient for effective control during simple manipulation. The lack of a learning curve suggests that the development of systems that automatically control aspects of mechanical interaction may offer users more advanced control capabilities with low cognitive load.

Rights

© The Author

Is Part Of

VCU University Archives

Is Part Of

VCU Theses and Dissertations

Date of Submission

5-10-2019

Share

COinS