Design and implementation of a DSP-based control interface unit (CIU)

Abstract

This research involves the development of a human-body motion tracking system constructed with the use of commercial off-the-shelf (COTS) components. The main component of the system investigated in this thesis is the Control Interface Unit (CIU). The CIU is a component designed to receive data from the magnetic, angular rate, and gravity (MARG) sensors and prepare them to be transmitted through a wireless configuration. A simple and effective algorithm is used to filter the sensor data without singularities, providing the measured attitude in the quaternion form for each human limb. Initial calibration of the MARG sensors is also performed with the use of linear calibrating algorithms. The testing and evaluation of the whole system is performed by MATLABʼ and SIMULINKʼ simulations, and by the realtime visualization using a human avatar designed with the X3D graphics specifications. Through this research, it is discovered that the MARG sensors had to be redesigned to overcome an erratum on the Honeywell magnetometer HMC1051Z data sheet. With the redesigned MARG sensors, the testing results showed that the CIU was performing extremely well. The overall motion tracking system is capable of tracking human body limb motions in real time.