Accuracy Enhancement of RSSI-based Distance Estimation by Applying Gaussian Filter

Abstract

RSSI values of Bluetooth Low Energy (BLE) beacon are unreliable to use localization. To cope with this problem, we propose a new localization algorithm that enhances the accuracy of RSSI value. The proposed algorithm applies Gaussian filter to RSSI values from BLE Beacons, and then uses weight value based on filtered RSSI quality. Background/Objectives: Friis formula is used to calculate distance between a BLE beacon and BLE scanners using triangulation scheme. Finally, Gaussian filtering is applied twice to the location values for accuracy improvement. Findings: Experiments are performed in indoor environment, and experiment data is calculated by MATLAB to make graph and chart for easily comparing location result. Experiment result shows that DGF algorithm shows more accurate and reliable localization result than commonly used Kalman filter algorithm. Furthermore, DGF algorithm is very effective to calculate not only distance but also location. Improvements: DGF algorithm indicates excellent performance when we adapt weight value. The proposed system can be used to tracking the location of BLE beacon in real-time.