Fast WiFi access point localization and autonomous crowdsourcing

Abstract

The locations of WiFi access points (APs) are important for WiFi positioning, especially when a propagation model is used. However, AP localization is usually challenging in a new environment because it is hard to obtain parameters for the propagation model, such as the path-loss exponent without the presence of surveyed database. This paper introduces a novel crowdsourcing method for automatic AP localization and propagation parameters (PPs) estimation based on the navigation solution from the Trusted Portable Navigator (T-PN). The estimation for PPs and AP locations based on non-linear weighted least squares (LSQ) is carried out automatically when enough measurements are collected, and the results are recorded in the database for future use. The fast estimation method calculates the propagation parameters autonomously and adaptively to account for the dynamic indoor environment. The autonomous system will also reduce the labour and time costs for the pre-survey and maintenance of databases, as the crowdsourcing is always done in background processes on devices. The accuracy of AP localization is also estimated and recorded in the database, providing an important indicator when using the AP localization results. The performance of the proposed system is evaluated by both simulations and field tests, and the result shows that the average AP localization errors are less than 6 meters.