A sensor array system for the estimation of 3D body and parts movements

Abstract

This paper introduces a Sensor Array System for the 3D capture and representation of users' limb and body movements. It also develops a design and implementation of a Wireless Inertial Measurement Unit (IMU) Sensor Array. By placing sensors on one's arm, hand, upper leg and lower leg, data needed by the application developers to represent the body pose of the user in a virtual environment is provided at a minimum frame rate of 100 Hz. A central processing unit fuses all data and sends it over to a PC. Data fusion and processing algorithms consider the position of the sensors using a quaternion representation. This allows the sensor array to represent coherently the sensed movements and correct the errors of the gyroscope by using magnetic sensors data. Since this system is not tethered by wires and there is no requirement for the user to stay in a camera's field of view (FOV), the user has the full freedom of motion in a 3D Virtual Environment while equipped with a Virtual Reality (VR) headset. In a more complex setup, data from the Sensor Array can be fused with data provided by a 3D camera in order to bring the real or the avatar image of the user in the virtual environment. A general architecture of the Sensor Array System is provided. The architecture is designed to take into account the integration with the Sensor Array of other devices specific to the VR field, such as infrared cameras, joysticks, VR headsets and other game controllers. The functionality of the Sensor Array is finally illustrated by results obtained from real movements of the user while exploring a virtual environment.