Crowdsourced Indoor Localization for Diverse Devices with RSSI Sequences

Abstract

Indoor localization is of great importance for a range of pervasive applications, attracting many research efforts in the past decades. Received Signal Strength Indication (RSSI) has received much attention due to its simplicity and compatibility with existing hardware, which has been widely used for indoor localization. However, traditional fingerprint-based localization solutions require a process of site survey, in which radio signatures are collected and stored for further comparison and matching. It’s a labor-intensive work to acquire the fingerprint database which costs much time and human resources. Meanwhile, users’ heterogeneous devices may receive different RSSIs even at the same location. To alleviate these problems, we present an efficient localization method with crowdsourced users’ RSSI sequences. We first cluster multiple users’ RSSIs to construct a logical floor plan graph, and construct a physical floor plan graph from the real floor plan. Then we map the logical floor plan with physical floor plan with a path-based method to construct the radiomap. To solve the device diversity, we propose a novel RSSI distance metric. When localizing an query user, we propose a bit encoder method to prune the RSSIs that cannot be the result. We demonstrate the efficiency and effectiveness of the proposed solution through extensive experiments with two real data sets.