Indoor Localization Based on PDR and Signals of Opportunity from Ambient Generic BLE Devices

Indoor positioning is essential for consumers with smartphones in GNSS-denied environments to be guided and navigated through complex facilities such as shopping malls, airports and stations. Although the Bluetooth Low Energy (BLE) beacons are promising for indoor positioning, it is cumbersome to setup and maintain the devices at a huge scale. This paper presents a novel approach to leverage Pedestrian Dead-Reckoning (PDR) and ambient generic BLE devices such as lighting apparatus, digital speakers and signages for locating the smartphones. The proposed method uses a build-and-refine approach to construct the map of the BLE devices. At the initial condition, the constraints of the skeleton structures in the floor map and the walking speed are utilized to estimate the absolute trajectory from the PDR-based relative trajectory, from which the first-generation BLE maps are generated. Once the BLE maps are obtained, PDR-based trajectories with BLE scan data are combined to estimate absolute trajectories which in turn are exploited to generate the next-generation BLE maps. Evaluation conducted on the office environment demonstrates the accuracy of mapping and localization.