A hybrid floor identification algorithm based on Bayesian classification and special AP

Accurately discriminating different floors is a very important task in indoor fingerprinting localization, which can be used to reduce space search domain and improve localization accuracy. There exist some research works for floor identification at present; however, the accuracy is not high. To achieve higher accuracy, this paper proposes a hybrid floor identification algorithm using Bayesian classification and special AP. By extracting the distribution feature of APs in different floors with training data, the proposed approach can determine floor efficiently with 100% accuracy.