Database updating through user feedback in fingerprint-based Wi-Fi location systems

Abstract

Wi-Fi fingerprinting is a technique which can provide location in GPS-denied environments, relying exclusively on Wi-Fi signals. It first requires the construction of a database of “fingerprints”, i.e. signal strengths from different access points (APs) at different reference points in the desired coverage area. The location of the device is then obtained by measuring the signal strengths at its location, and comparing it with the different reference fingerprints in the database. The main disadvantage of this technique is the labour required to build and maintain the fingerprints database, which has to be rebuilt every time a significant change in the wireless environment occurs, such as installation or removal of new APs, changes in the layout of a building, etc. This paper investigates a new method to utilise user feedback as a way of monitoring changes in the wireless environment. It is based on a system of “points” given to each AP in the database. When an AP is switched off, the number of points associated with that AP will gradually reduce as the users give feedback, until it is eventually deleted from the database. If a new AP is installed, the system will detect it and update the database with new fingerprints. Our proposed system has two main advantages. First it can be used as a tool to monitor the wireless environment in a given place, detecting faulty APs or unauthorised installation of new ones. Second, it regulates the size of the database, unlike other systems where feedback is only used to insert new fingerprints in the database.