WiFi Multifloor Indoor DCM Positioning

Database correlation methods (DCM) are used to locate mobile stations (MS’s) in wireless networks. A target radiofrequency (RF) fingerprint - measured by the MS to be localized - is compared with georeferenced RF fingerprints, previously stored in a correlation database (CDB). This paper focuses on the DCM positioning in multifloor indoor environments. In this scenario, the authors apply two combined techniques to reduce the search space inside the CDB, while improving the floor identification accuracy: i) unsupervised clustering using a single Kohonen Layer and ii) floor classification using committees of backpropagation artificial neural networks (ANN’s), one committee per each floor. The effects of the proposed solution on the DCM positioning accuracy are experimentally evaluated using 46200 target fingerprints and a CDB with 924 reference fingerprints, containing Received Signal Strength (RSS) values of 136 WiFi 802.11b/g networks in a 12-floor building. The correct floor is identified in 91% of the samples, and is within 2 floors in 99% of the samples. The average positioning error is 4.7 meters and is below 5.5 meters in 75% of the samples.