A Compact and Robust 6-DoF Pose Tracking System Using Magnetic-Inertial Sensors and a Single Uniaxial Electromagnetic Coil
Shuda Dong,Heng Wang
DOI: https://doi.org/10.1109/jsen.2023.3348243
IF: 4.3
2024-02-16
IEEE Sensors Journal
Abstract:Magnetic pose tracking, which is noncontacting and occlusion-free, is increasingly used in virtual/augmented reality, rehabilitation, and computer-assisted medical interventions to track the motion of human body and various tools and devices. Most existing magnetic tracking systems employ multiple magnetic sources and multiple magnetic sensors, making the system bulky and energy consuming. Complex signal processing circuits and algorithms are needed to distinguish the magnetic signal from each magnetic source. In this article, a compact 6-DoF magnetic-inertial tracking system is developed using only a single stationary uniaxial electromagnetic coil and a single magnetic-inertial sensor board attached to the moving object. Magnetic measurement of the electromagnetic source, geomagnetic measurement, and inertial measurements is fused using an extended Kalman filter (EKF) framework to provide accurate, drift-free, and robust pose estimates. An ON—OFF switching strategy for electromagnet control and a difference magnetic measurement technique are used to cancel out potential ambient magnetic disturbances. Magnetic measurement of the electromagnetic source is mainly responsible for 3-DoF position estimation, while geomagnetic field measurement is used to correct the yaw angle drift in inertial orientation estimation. Since geomagnetic measurement is susceptible to magnetic disturbances, it is switched off adaptively once detecting disturbances to maintain the robustness of the pose tracking system. Experimental results show that the root-mean-square (rms) Euclidean position error is 2.84 mm and the average rms orientation angle error is 1.22 °C even in the presence of magnetic disturbance of a permanent magnet.
engineering, electrical & electronic,instruments & instrumentation,physics, applied