Pre-processing of fingerprints to improve the positioning accuracy of 802.11-based positioning systems

Since the first days of location-based services, the reliable esti-mation of a user’s position has been seen as one of the key pre-requisites for their success. While satellite-based systems can de-liver suitable results when the user is outdoors, another solution is needed in case that he is located inside a building. Here, the position estimation with the help of 802.11 and fingerprinting al-gorithms promises to fill the gap. One major drawback though is the high initial effort that is needed to deploy such a fingerprinting-based 802.11 positioning system. Therefore, in this paper, novel algorithms that pre-process the data collected for the fingerprints are presented. By this pre-processing step, the amount of data that needs to be collected and therefore also the amount of time that needs to be spent collecting the data can be reduced by about 80%. Additionally, despite the reduced amount of data, the positioning accuracy cannot only be retained but even slightly improved by about 15 % depending on the system’s setup