Kalman filter based localization and tracking estimation for HIMR RFID systems

Abstract

Fine-scale RFID motion capture and localization systems are important for emerging applications such as robotic wayfaring, virtual reality, remote motion capture, and tracking. Among all RFID localization and tracking systems, hybrid schemes have the highest accuracy (5–400 cm) and largest range (hundreds of meters) by combining a variety of sources. Recently, linear estimators have been introduced to hybrid inertial microwave reflectometry (HIMR) RFID systems for location and tracking estimation. In this paper, we add a Kalman filter to the output of the original HIMR estimator as a way to refine the precision of the technique. Simulation results show that compared to sensor outputs, fixed-gain and variable-gain versions of the Kalman filter track position much more accurately, with around 5 mm of rms error at a range of 1.2 m.