Bridging the Performance Gap Between Two-Way and One-Way CSI-Based 5 GHz WiFi Ranging

Indoor Localization is gaining increased importance due to numerous location-based services in healthcare, logistics, and security, to name few, that are expected to be provided by next-generation wireless networks. Such services are characterized by stringent accuracy requirements, short response time, and lower cost which makes the localization problem more challenging and deserving of attention. A key element of the localization process is distance estimation (also known as ranging). In this paper, we design and analyze an efficient decimeter-level two-way ranging scheme for ubiquitous WiFi networks in the 5 GHz frequency band whose accuracy approaches ideal one-way ranging with no phase mismatches. We investigate the idea of channel frequency response (CFR) stitching across non-contiguous WiFi channels and how two-way CFR measurements help in achieving the CFR coherency necessary for accurate ranging. In addition, we quantify the decrease in ranging accuracy of two-way compared to one-way ranging due to SNR degradation, Line-of-Sight (LoS) component shrinkage, and doubling the multipath delay spread. Furthermore, We design a novel scheme to bridge the performance gap between two-way ranging and ideal one-way ranging which operates in three main steps: square-root of the two-way CFR, followed by phase unwrapping, and finally deep fade detection and phase errors correction. Our proposed scheme achieves significant performance gains over two-way ranging with only a slight performance gap from ideal one-way ranging. Moreover, our proposed scheme enjoys robustness as it preserves the ranging accuracy gains in various WiFi communication scenarios when operating at different SNR levels, different multipath channel models, and different CFR bandwidths, as well as operating under system impairments such as Sample Timing Offset (STO). The accuracy gains achieved by the proposed schemes are demonstrated using both simulations and an in-house WiFi testbed. Finally, we quantify the added complexity of our proposed scheme and show it to be insignificant compared to that of the MUSIC super-resolution ranging steps which confirms the practical viability of our proposed scheme.