Reference tag supported RFID tracking using robust support vector regression and Kalman filter

Regarding the status monitoring among material, equipment and personnel during site operations, much work is conducted on localization and tracking using Radio Frequency Identification (RFID) technology. However, existing RFID tracking methods suffer from low accuracy and instability, due to severe interference in industrial sites with many metal structures. To improve RFID tracking performance in industrial sites, a RFID tracking method that integrates Multidimensional Support Vector Regression (MSVR) and Kalman filter is developed in this paper. Extensive experiments have been conducted on a Liquefied Natural Gas (LNG) facility site with long range active RFID system to evaluate the performance of this approach. The results demonstrate the effectiveness and stability of the proposed approach with severe noise and outliers. It is feasible to adopt the proposed approach which satisfies intrinsically-safe regulations for monitoring operation status in current practice.