Maintenance of Wi-Fi Fingerprint Database by Crowdsourcing for Indoor Localization

Wi-Fi fingerprint localization has been well accepted to solve the indoor localization problem, while the most important part in this method is the construction and maintenance of the fingerprint database. A timely update of the fingerprint database is required due to the fluctuation of the received Wi-Fi signal strength. However, having professionals updating the fingerprint database is costly in terms of time and effort. In this paper, a crowdsourcing approach is proposed where the users themselves can evaluate and correct the localization result, thus training and using the localization system at the same time. Particularly a clustering based error detection method is adopted to detect the erroneous input location data, which can effectively avoid the contamination of the fingerprint database. Finally an indoor localization system is developed and practical experiments in real indoor environment show that the proposed approach achieves satisfying localization accuracy over a long time.