Exploiting Acoustic Echoes for Smartphone Localization and Microphone Self-Calibration

We present a novel approach for locating a smartphone in mixed non-line-of-sight (NLOS) and line-of-sight (LOS) environments when the positions of the sensor nodes are unknown. The reception times of acoustic signals emitted by the smartphone are used to estimate the position of the target and the microphones in three dimensions. The acoustic reflections from walls are treated as unknown virtual receivers, which facilitate the task of finding the sender. The observed time differences by the microphones and the virtual receivers are tracked by a joint probability data association filter. Afterward, the trajectory of the target and the receiver positions are estimated using a far-field approximation approach.