Data Acquisition and Analysis for the Design of a
Powered Prosthetic Leg
By Pushpak Jha
Class of 2011
Electrical/Mechanical Engineering
Honors Program
Project Advisor: Kevin Fite
This project involved the acquisition and analysis of human gait biomechanic data and
thigh electromyogram (EMG) data for the purposes of design and control of a powered prosthetic
limb for transfemoral amputees. When complete, this prosthetic leg will provide amputees with
an anthropomorphic limb with gait mechanics approaching that of able-bodied subjects. In this
multi-subject study, each participant’s left thigh was instrumented with surface EMG electrodes
(Delsys model DE-2.3) to target five muscle groups: bicep femoris, semitendinosus, vastus
lateralis, vastus medialis, and rectus femoris. Biomechanic and EMG data were then measured
for level walking, stair ascent, stair decent and sit to stand motions. The data was then compiled
and processed using MATLAB. The data processing of each variable of interest entailed
ensemble averaging, interpolating gaps in the data, normalizing to full stride, and computing
standard deviations and coefficients of variance. The EMG data was band-pass filtered, biased
about zero, full-wave rectified, and then forward and reverse low pass filtered using a 2nd order
Butterworth filter. Based on the measured data, the kinematics and kinetics for each subject’s
hip, knee and ankle were characterized. The ensemble averaged data will be used in the design of
powered knee and ankle joints for a powered transfemoral prosthesis, and the thigh EMG data
will be used in the development of an EMG-based control interface for amputee-limb integration.