As a device-free approach, radio tomographic imaging (RTI) is ideally suited for low-cost indoor localization in context-aware Internet-of-Things applications. However, the fundamental RTI algorithm relies on shadowing of the line of sight (LOS) links and therefore, conventional RTI implementations using 2.4 GHz sensing networks (microRTI) fail to accurately localize users in multipath-rich indoor environments. The localization accuracy is further degraded by external human movement that affects the signal propagation. In this paper, we propose mmRTI, a novel RTI approach based on a highly-directional, millimeter-wave sensing network, that aims to improve indoor localization by utilizing the LOS-dominant nature of millimeter-wave signal propagation. We experimentally evaluate mmRTI, operating at 60 GHz, with and without human movement around the sensing network, in two indoor environments, and compare its performance against the conventional microRTI approach. We observe that mmRTI achieves a 90%-ile localization error of 0.07 m-0.25m, an improvement of 2.41 m-2.60m compared to microRTI, while remaining unaffected by external human movement, which degrades the microRTI localization accuracy by up to 1.2 m.
TEKNIK OCH TEKNOLOGIER -- Elektroteknik och elektronik (hsv//swe)
ENGINEERING AND TECHNOLOGY -- Electrical Engineering, Electronic Engineering, Information Engineering (hsv//eng)