A novel WKNN algorithm for fingerprinting-based fusion of different radio communication technologies for indoor positioning

Indoor Positioning Systems (IPSs), which commonly rely on Wireless Sensor Networks (WSN), are gaining attention in various fields due to the emergence of Industry 4.0. In this paper, a novel Weighted K-Nearest Neighbor (WKNN) -based localization algorithm is proposed for the fusion of different wireless parameters. An experiment was set up for evaluation and measurements were performed in a laboratory. The essence of the novel method is that different wireless communication technologies, namely, 2.4 GHz WiFi Received Signal Strength Indicator (RSSI), 433 MHz CC1101 RSSI, Ultra-Wideband (UWB) RSSI, and UWB Time of Arrival (TOA) were combined for localization purposes. The proposed algorithm has two steps. In the first step the technologies are used separately to calculate the searched position. In the second step the average of the estimated coordinates is taken in all possible combination of the technologies. The original WKNN method considers the data together provided by each technology. The proposed WKNN algorithm provided better results in 5 cases than the original WKNN method. The improvement ranged from 0.7% to 15%. The largest improvement was achieved by using WiFi and CC1101 RSSI together, when the average error was reduced from 1.86 m to 1.58 m. The results show that it is worthwhile to adaptively select the appropriate fusion of technologies, as this can significantly reduce the localization error.