Conditional Rician $K$-Factor Discrimination for Indoor Localization via AOA Estimation

This paper proposes conditioning angle of arrival (AOA) algorithms for pseudo-spectrum fingerprint acquisition based on line of sight (LOS) and non-LOS detection schema for optimizing indoor localization. The proposed approach merges two AOA based methods being that of the MUltiple Signal Classsification (MUSIC) algorithm and virtual MUSIC algorithm into a conditional based localization approach with a uniform circular array (UCA). The paper begins by demonstrating the environmental dependencies of the two AOA approaches based on the Rician $K$-factor metric. The $K$-factor is then exploited as an algorithm selection metric to arrive at improved localization performance in a realistic indoor environment.