Filtering Effect on RSSI-Based Indoor Localization Methods

Abstract

Indoor positioning systems are used to locate and track objects in an indoor environment. Distance estimation is done using received signal strength indicator (RSSI) of radio frequency signals. However, RSSI is prone to noise and interference which can greatly affect the accuracy performance of the system. In this paper Internet of Things (IoT) technologies like low energy Bluetooth (BLE), WiFi, LoRaWAN and ZigBee are used to obtain indoor positioning. Adopting the existing trilateration and positioning algorithms, the Kalman, Fast Fourier Transform (FFT) and Particle filtering methods are employed to denoise the received RSSI signals to improve positioning accuracy. Experimental results show that choice of filtering method is of significance in improving the positioning accuracy. While FFT and Particle methods had no significant effect on the positioning accuracy, Kalman filter has proved to be the method of choice in for BLE, WiFi, LoRaWAN and ZigBee. Compared with unfiltered RSSI, results showed that accuracy was improved by 2% in BLE, 3% in WiFi, 22% in LoRaWAN and 17% in ZigBee technology for Kalman filtering method.

Keywords: RSSI, WiFi, Bluetooth low energy, LoRaWAN, Zigbee