Implement and optimization of indoor positioning system based on Wi-Fi signal

Abstract

As wireless routers are used widely, indoor positioning technology based on Wi-Fi signal has drawn more attentions. The positioning process in our solution is divided into two phases: collection phase and positioning phase. In the collection phase, according to the fingerprint algorithm, data collectors (e.g. mobile phones) submit received Wi-Fi strength data at location-known points to the server. The collected locations and strength data will be saved in database. In the positioning phase, the server calculates positioning result according to the differences between Wi-Fi strength data stored in database and Wi-Fi strength data uploaded by mobile terminals request to be located. All the data are clustered using K-Means algorithm for increasing the positioning efficiency. K-Nearest-Neighbor (KNN) algorithm is performed in positioning phase. The result of experiment shows that the proposed approach can achieve high positioning accuracy with the use of filtered data and the weighted KNN algorithm.