3DinSAR: Object 3D localization for indoor RFID applications

More and more objects can be identified and sensed with RFID tags. Existing schemes for 2D indoor localization have achieved impressing accuracy. In this paper we propose an accurate 3D localization scheme for objects. Our scheme leverages spatial domain phase difference to estimate the height of objects which is inspired by the phase-based Interferometric Synthetic Aperture Radar (InSAR) height determination theory. We further leverage a density-based spatial clustering method to choose the most likely position and show that it improves the accuracy. Our localization method does not need any reference tags. Only one antenna is required to move in a known way in order to construct the synthetic arrays to implement the locating system. We present experimental results from an indoor office environment with EPC C1G2 passive tags and a COTS RFID reader. Our 3D experiments demonstrate a spatial median error of 0.24 m. This novel 3D localization scheme is a simple, yet promising, solution. We believe that it is especially applicable for both portable readers and transport vehicles.