Design, implementation, and testing of a real-time software system for a quaternion-based attitude estimation filter

Abstract

Human limb segment angle tracking requires a system which can track through all orientations. The major problem addressed by this research was to develop a real time inertial motion tracking system based on quaternions to overcome the singularities of Euler angle filters. This work involves clarification of the theory behind quaternion attitude estimation and development of a system capable of determining the orientation of an object in world coordinates. System sensors were built using miniature accelerometers, rate sensors, and magnetometers. The software system was designed by using Unified Modeling Language (UML) with object oriented design techniques. The actual implementation created a real time orientation tracking system. The system was tested with dynamic tilt table experiments. Test results showed that the quaternion attitude estimation filter system can track human limb segments in real time within 1 degree of accuracy in an orientation and with a 55 Hz update rate.