Analysis and Implementation of an Orientation-Aware Source Localization System with Smart Devices

Abstract

Target localization in wireless systems has experienced a great improvement during recent years given the increasing demand for location services by mobile users. Particularly, localization methods based on received signal strength indicator (RSSI) are highly attractive because hardware is readily available and cost-effective.

The RSSI-based localization literature generally focuses on the propagation environment sensitivity and overlooks a major factor in signal strength variability: the relative orientation between the source and the receiver. With current advancements in hardware, especially low-cost microelectromechanical sensors, most wireless devices are orientation-aware. This offers an opportunity to enhance the performance of multi-agent localization systems.

We propose to include orientation knowledge and the typical antenna radiation pattern asymmetries of the sensing devices into the inference task. We will gather experimental data using Android™ smartphones as sensing devices, implement the new orientation-aware algorithm and asses the improvements of our approach in simulations and real-world scenarios.

We compare the new scheme with the standard setting where the orientation is unknown. The orientation-aware implementation performs significantly better than the traditional systems in terms of accuracy. These results show that orientation-awareness capabilities should be accounted for whenever possible in tasks of statistical inference. Furthermore, this idea is likely to find applications beyond source localization.