Generating And Updating Of Wi-Fi Fingerprint Database Based-on Crowdsourcing Data

Location-based Service(LBS)attracts more and more attention with the rapid development of smartphones and mobile Internet applications recently.In outdoor scenarios,many satellite-based localization technologies,such as Global Navigation Satellite System(GNSS)and Beidou Navigation System(BDS),are mature enough to provide stable localization service.However,in indoor environments,the satellite-based navigation system can't provide reliable service due to the signal fading and multipath effect.With the expanded size and complex layout of indoor environments,indoor localization service has become a popular demand,and the indoor localization technology has become a hot issue in the field of navigation.Wi-Fi indoor localization is one of the most widely used indoor localization scheme,because of the low cost,easy deployment,etc.Conventional Wi-Fi indoor localization systems consist of two phases: offline phase for fingerprint database generating,and online phase for realtime locating.The offline phase of Conventional Wi-Fi indoor localization systems consumes heavy workload and brings a lot of restrictions to the collectors,which impedes the spread and application of Wi-Fi localization system.To overcome this problem,we propose a technology of database generating based on the indoor graph and crowdsourcing data.In our method,we choose crowdsourcing data as our source,which is wide available and sufficient in quantity,we utilize the graph-based Pedestrian Dead Reckoning(PDR)to get the location of the corresponding sampled Wi-Fi signal.It's an unsupervised database generating,which reduces the workload and relaxes the requirements on collectors.Most researches of Wi-Fi indoor localization technologies focus on the fingerprint training and real-time localization algorithms,the research on fingerprint updating is very scarce.In fact,the updating of fingerprint database is very necessary in Wi-Fi localization systems.The importing error from database generating and the change of Wi-Fi environments will influence the localization system,if the database remains static or the updating is not timely or accurate,the accuracy of the location system may decrease.To overcome this problem,we propose two updating methods: Fingerprint updating based on the similarity between fingerprints and Fingerprint updating based on the fingerprints' reliability model.The application of these two methods will reduce the importing errors from database generating to the greatest extent.When facing the change of the Wi-Fi environments,the database will be updated in time,and it will effectively reduce the loss of accuracy.