Multi-Step Optimization of Indoor Localization Accuracy Using Commodity WiFi

Abstract

We present a Multi-step optimization Localization Algorithm (MoLA) for improved accuracy of indoor localization by WiFi. Using consumer-grade WiFi devices, we can determine the location of an object carrying WiFi device with higher accuracy than the comparable popular systems. MoLA removes the phase error by using calibration, then estimates the angle-of-arrival (AoA) of the signal by combining I-MUSIC algorithm and the minimum description length (MDL) equation. The line-of-sight (LoS) path is identified using our novel estimator function. In the last step, the object location is estimated. Extensive measurement experiments have shown that MoLA can achieve improvement in both median localization and point localization error in a 290 m2 environment having adequate LoS with a single receiver.