ParkLoc

Abstract

Locating a vehicle indoors (e.g., underground parking garages) has been a difficult problem to tackle, due to the unavailability of GPS and/or Wi-Fi signals. Current GPS-free indoor localization efforts often rely on infrastructure supports such as Wi-Fi or BLE beacons, whereas the smartphone-only proposals mostly require significant data collection and training efforts per garage. In this context, we propose ParkLoc, a novel lightweight smartphone-only solution for vehicular localization in GPS/Wi-Fi-deprived environments such as indoor parking garages. ParkLoc exploits the inherent planar graph structure of the navigable paths in parking facilities, in order to match a vehicle trajectory onto a sub-section of the map, by modeling these as sparse directed graphs. Exploiting an approximate graph matching method, ParkLoc is able to track a vehicle in real-time with a median error of 4.8m and localize a parked vehicle with a median error of 2m from the nearest parking space. Furthermore, ParkLoc adopts the popular GraphSLAM algorithm from robotics research; it learns the map graph from the observed trajectory graphs and a given set of bootstrap (seed) landmark nodes, in a semi-supervised manner. A key benefit of our approach is that ParkLoc works off-the-shelf without any expensive on-site training or sensor data collection per garage. A comprehensive evaluation of ParkLoc through extensive experiments performed in 4 different parking facilities reveals the promising performance of our graph-based approach for both localization and mapping.