Towards Practical Probabilistic Location Inference for Indoor Environment

In this work, we highlight the truncation effect in Received Signal Strength Indication (RSSI) distributions. The effect is often overlooked when applying probabilistic methods such as Multivariate Gaussian Inference (MGI), for location inference. Towards practical and accurate probabilistic location inference, we propose Multivariate Truncated Gaussian Inference (MTGI) to deal with the beacons with persistent or sporadic packet losses due to signal decay or collisions.