Metasurface-Enabled High-Resolution Localization of Non-Cooperative Wireless Source in an Indoor Environment

It is a fundamental but challenging open question to localize narrowband wireless sources in compilated indoor environment in a super-resolution and cost-efficiency way due to the complicated multipath effects. As a matter of fact, most wireless localization methods developed so far, like, the direction of arrival (DOA), received signal strength indicator (RSSI), time of arrival (TOA), etc., relies on the assumption of line-of-sight signal propagation, and fail to deal with the complicated indoor environments with the abundantly multi-path wave phenomena. To deal with the formidable challenge, we here present a super-resolution method of localizing the non-cooperative wireless sources in an indoor environment by deploying a deep-learning-driven reprogrammable metasurface and building a proof-of-principle demo system working at around 2.4GHz, where only a pair of coherent receivers are utilized for the signal acquisition. We demonstrate experimentally that our system is capable of localizing the commodity Wi-Fi signal source with an accuracy of several centimeters and a rate of 20Hz.