TagTone: Scalable RFID communication through multi-frequency analysis

RFID tags are inexpensive wireless sensors that harvest energy from the excitation signal sent by a reader. These tags are widely used in industrial settings for several applications. These applications require large-scale dense deployment where packet collisions are unavoidable. The major hurdle limiting these applications from scaling is the lack of suitable protocols that can decode collided packets from a large number of tags. We present TagTone, a new protocol that utilizes the wide bandwidth for RFID communication and the independence of the channels to decode colliding packets from a large number of tags. TagTone is a scalable protocol that can decode more tags as the bandwidth or the number of antennas increases making it an ideal solution for dense deployment. In this paper, we present a thorough analysis to show that RFID channels are independent across frequencies and antennas. Packets from the tags are collected using an excitation signal comprising of multiple complex sinusoids (tones) at different frequencies. TagTone includes a novel Moving Window Packet Separation (MWPS) algorithm which modifies traditional Independent Component Analysis (ICA) for superior decoding. TagTone implemented with USRP N210s can decode 7× more colliding tags and provides 7× better throughput than the existing protocols.