Measurement of angles in time-of-flight data for the automatic supervision of training exercises

Abstract

In this paper we present a novel automatic system to supervise and support rehabilitation and fitness exercises. Goal of the system is to enable autonomous training of rehabilitation patients to support their recovery from movement restrictions as well as the prevention of age-related diseases such as osteoporosis. The system uses time-of-flight technology in order to measure movement angles in real-time. This novel setup is compared to a standard tracking system using visual markers. The measurements of both systems correlate with r values between 0.99 and 0.97 depending on the task domain. The mean error of the angle measurements is between 13 and 17 degrees.