Journal of Bioengineering & Biomedical Science

Towards Low - Cost Mems Imu Gait Analysis: Improvements Using Calibration and State Estimation

Abstract

Lucas Samuel Lincoln, Eric Allen Johnson and Stacy J. Morris Bamberg

Inexpensive, unobtrusive 3D motion tracking of human gait is of increasing interest for the medical and entertainment industries. Of particular interest are rehabilitative applications. For instance, being able to measure foot travel, e.g. stride length or foot clearance, would be very useful. Approaches using low-cost MEMS inertial measurement units have often been limited by requiring expensive calibration procedures and by the sensor’s inherent noise and bias drift. The authors apply two techniques to improve IMU based gait tracking: a novel calibration routine and a zero-velocity bias update algorithm. The application of these aids reduces error by an average of 99.55% over six trails. Results show a 5.96% tracking accuracy in the progressive direction, which corresponds to errors on the centimeter scale.