Optimization and performance analysis of Wi-Fi indoor fingerprint localization algorithm

With the development of the Internet of Things (IoT), location estimation has become an essential and key factor in various monitoring applications for the IoT. Location fingerprint localization based on Wi-Fi has gradually become a hot spot in the field of localization technology. It uses different fingerprint-matching algorithms for user location estimation. However, many interfering factors can affect the transmission of the signal, resulting in biased fingerprint data. Such as indoor multipath, shadows, noise, and other interference. Therefore, to enable various applications of IoT, the need for accurate indoor location services has become very important. In this paper, different fingerprint-matching algorithms in the localization system based on Wi-Fi Received Signal Strength Indicator (RSSI) signal are compared and analyzed. In the analysis, we compare the Nearest Neighbor algorithm (NN), K-Nearest Neighbor Algorithm (KNN), and Weighted K-Nearest Neighbor Algorithm (WKNN) for user location estimation. This paper proposes an improved localization algorithm for WKNN, which not only optimizes the number of RSSI acquisition and localization range but also optimizes the selection of K. Thereby improving the accuracy of indoor fingerprint localization. The performance of these fingerprint-matching algorithms is discussed from the aspects of average localization error and probability distribution of localization error. Experimental results show that the WKNN fingerprint-matching algorithm has higher localization accuracy compared with other fingerprint-matching algorithms.