Cluster-Based JRPCA Algorithm for Wi-Fi Fingerprint Localization

Indoor localization services are emerging as an important application of the Internet Things, which drives development related technologies in indoor scenarios. In recent years, various algorithms for different scenarios have been proposed. The algorithm based on fingerprints has attracted much attention due to its good performance without extra hardware devices. However, occurrence fingerprint mismatching caused by complexity and variability is unneglectable, degrades accuracy. this article, combining weighted kernel norm L2,1-norm, a joint-norm robust principal component analysis (JRPCA brief) assisted proposed, can improve accuracy through aggregating reference points (RPs) conducting feature extraction clustering. More specifically, one-way hierarchical clustering termination method proposed obtain reasonable RP clusters adaptively according preset RPs. A two-phase matching JRPCA further increase difference between similar RPs, thus facilitating rapid identification reinforcing To validate algorithm, extensive experiments conducted real experimental results confirm that cluster-based outperforms other existing terms robustness