Adaptive Weighted Access Point Technique for Robust Dynamic Indoor Positioning

Driven by the burgeoning demand for location-based services, the quest for robust indoor positioning systems (IPSs) becomes increasingly imperative. Currently, fingerprint-based approach has emerged as the predominant technique for IPSs because of its ability to deliver accurate location estimations without necessitating line-of-sight transmission from access points (APs). This approach involves constructing a radio map based on the wireless signal properties and employing matching algorithms, such as the nearest neighbor, to predict the position of target. Despite its merits, this technique has its drawback, since received signal strength (RSS) are heavily impacted by the indoor environment, leading to potential discrepancies between real-time signal measurements from the target’s mobile device and RSS vector from the fingerprint database when environmental changes occur. To tackle this issue, this work proposes a novel adaptive weighted AP (AWA) algorithm that exploits newly acquired RSS vector from a small subset of the reference points to dynamically assign weights to different APs based on the impact of environmental changes on their RSSs. By harnessing these weights to compute the distance between the RSSs at unknown location and those in the fingerprint database, two novel IPSs are developed to effectively identify the nearest neighbors of the target and subsequently predict the target’s location. Remarkably, results reveal that the AWA-based IPSs achieve a substantial enhancement of 40% in average positioning error over the baseline counterparts without AWA in dynamic indoor environments by leveraging merely 9.37% of the reference points to measure the updated RSSs.