Dedis Lab

Michael F. Nowlan

Human Identification via Gait Recognition using Accelerometers and Gyroscopes

In Fall 2009 I worked on a class project to identify humans based on their gait. This work was for CPSC-536 Networked Embedded Systems and Sensor Networks. The project used the Wii Remote to capture the forces of a person's walk in 6 different directions. The gyroscope of the Wiimote turned out to provide the most valuable information. I used this signal to match a person to their gait cycles. The project report is here.

In the Spring 2010, we continued working with the Wii Remote trying to find a more robust motion signature. Previously, the goal was to identify a person by measuring their gait by putting the remote in the pants' pocket. Now, we are looking at whether a set of body sensors can be identified as belonging to the same person. We gathered data from multiple sensors placed on the body, but the project has stalled as we've become busier with other projects during the semester. If we start up with this work again, I'll post here.

Last Update: 20100504