Hybrid Indoor Localization System Combining Multilateration and Fingerprinting

Abstract

Indoor localization systems enable object tracking, e.g., related to retail, logistics and mobile robotics during their life cycle. This paper describes a real-world scenario implementation, based on Bluetooth Low Energy (BLE) beacons, evaluating a hybrid indoor positioning system (H-IPS) that combines two RSSI-based approaches: Multilateration (MLT) and Fingerprinting (FP). In addition, a Kalman Filter (KF) was employed to decrease the positioning errors of both techniques. Furthermore a track-to-track fusion (TTF) is performed on the two KF outputs to maximize the performance. The results show that the accuracy of H-IPS overcomes the standalone FP in 21%, while the original MLT is outperformed in 52%. Finally, the proposed solution demonstrated a probability of error <2m of 80%, while the same probability for the FP and MLT were 56% and 20%, respectively.