Error bound analysis of indoor Wi-Fi location fingerprint based positioning for intelligent Access Point optimization via Fisher information

Abstract

We observe that the laborious collection of location fingerprints, that could also potentially change with time and space, remains a hurdle towards the widespread deployment of indoor and outdoor positioning systems using the Wi-Fi. Motivated by this, we focus on the error bound analysis of indoor Wi-Fi fingerprint based positioning for intelligent Access Point (AP) placement optimization by using the Fisher Information Matrix (FIM) to characterize the relationship between the positioning errors and signal distributions. We first derive the closed-form error bounds with respect to different system parameters. Second, we use the Simulated Annealing (SA) algorithm to conduct the Wi-Fi AP placement optimization towards the lowest error bounds. Finally, we present the detailed discussion, extensive simulation, and experiment results.