A Wi-Fi Fingerprint Positioning Method Based on RLWKNN

Wireless fidelity (Wi-Fi) fingerprint positioning technology has the benefits of affordable hardware and easy operation, making it a popular choice in the realm of indoor positioning. However, this technology’s positioning accuracy tends to be on the meter scale, with room for further improvement. This article proposes a novel indoor positioning method based on range limit weighted k nearest neighbors (RLWKNN) algorithm to improve the accuracy of indoor positioning. First, in order to improve the robustness of device heterogeneity, this article designs a fusion distance that combines the Euclidean distance and the cosine distance; it integrates the spatial distance and signal pattern similarity of location fingerprints. Subsequently, to avoid the hassle of setting the k value manually, the adaptive k value is designed, and it automatically determines the k value based on the maximum difference in fusion distance. Finally, to further filter location fingerprints, a range limit is introduced, and it determines whether to filter the location fingerprint by checking the distance between the location fingerprint collection location and the last predicted user location. Experimental results demonstrate that the method proposed in this article surpasses conventional positioning methods across a range of indoor environments and exhibits higher performance.