Batch Localization Based on OFDMA Backscatter

Abstract

OFDMA Wi-Fi backscatter can significantly improve the communication efficiency and meanwhile maintain ultra-low power consumption; however, the ground-up reworking on the core mechanism of traditional Wi-Fi system revolutionizes the basis of many existing Wi-Fi based mechanisms. In this paper, we explore how localization can be realized based on OFDMA backscatter, where a batch localization mechanism utilizing concurrent communication in the OFDMA backscatter system is proposed. We present a series of mechanisms to deal with the fundamental change of assumptions brought by the new paradigm. First, we process signals at the receiver in a finer granularity for signal classification. Then we remove phase offsets in real time without interrupting the communication. Finally, we propose an extended MUSIC algorithm to improve accuracy with limited localization information in OFDMA backscatter mechanism. We implement a prototype under the 802.11g framework in WARP, based on which we conduct comprehensive experiments to evaluate our propose mechanism. Results show that our system can localize 48 tags simultaneously, while achieving average localization errors within 0.49m. The tag's power consumption is about 55-81.3μW.