An Annulus Local Search Based Localization (ALSL) Algorithm in Indoor Wi-Fi Environments

With growing demands on location-based services, the indoor localization has attracted great attention in both academia and industry. This paper focuses on improving indoor localization performance based on Wi-Fi signal processing. Specifically, we propose an annulus local search based localization (ALSL) algorithm, which combines the distance fitting and fingerprint-based techniques. First, ALSL derives the relationship between the Received Signal Strength Indicator (RSSI) and the Euclidean distance for a Wi-Fi Access Point (AP) based on multinomial fitting. Then, an annulus construction scheme is proposed to reduce the online search space for fingerprint-based localization. On this basis, a probability model is derived to estimate the probability of an online point (OP) which is located at certain reference point (RP) based on Bayes theorem. Finally, K-Nearest-Neighbor (KNN) algorithm is adopted to compute the final estimated location of an OP. For performance evaluation, we build a prototype of the indoor localization system and implement the ALSL algorithm. We conduct a series of experiments in real-world scenarios, which show that ALSL is able to effectively reduce the computational overhead during the online localization phase with the annulus based local search policy. In addition, our field testing also demonstrates that ALSL can outperform existing competitive solutions in terms of improving the localization accuracy.