Research On Low-overhead Wi-Fi Rf Fingerprint Positioning Technology

In recent years,with the rapid development of mobile Internet,the widespread use of wireless networks and mobile smart devices,Wi-Fi-based indoor positioning technology has been widely adopted.However,in the current research based on Wi-Fi indoor positioning technology,there are still many challenges:the location fingerprint database constructed in the offline phase relies on manual on-site fingerprint data,which leads to high deployment cost.The location fingerprint database cannot update the indoor environment changes.The positioning accuracy is lowered,the fingerprint signal strength is unstable,and the like.Therefore,for the problem of location fingerprint database construction and location fingerprint database update cost,this paper studies the location fingerprint database reconstruction method based on Bayesian Compressive Sensing theory in crowdsourcing production environment.Firstly,this paper makes a detailed analysis of Wi-Fi indoor positioning technology.Due to the multi-path and shadow phenomenon caused by the complex indoor environment,the wireless signal strength fluctuates greatly.Therefore,in the offline training phase,the Kalman Filtering method is used to preprocess the position fingerprint data.Secondly,for the difficult problem of location fingerprint database construction,the Bayesian Compressive Sensing theory is used to reconstruct the location fingerprint database.At the same time,this paper draws on the crowdsourcing idea for data collection,and uses the neighborhood propagation clustering method to cluster the location fingerprint data.Then,the combination of Compressive Sensing theory and Bayesian theory is used to analyze the signal strength change value collected by a small number of position reference points,and then the offline fingerprint database is reconstructed by using the change value.Finally,the paper simulation realizes the location fingerprint database reconstruction scheme of Bayesian Compressive Sensing theory.The simulation results show that the position fingerprint database reconstruction scheme using Bayesian Compressive Sensing theory can reconstruct the localized location fingerprint database in the case of a small number of data samples under the premise of keeping the indoor positioning accuracy deviation small.Effectively solving the problem of large cost of location fingerprint database update,to some extent,reduces the overhead cost of manually maintaining the fingerprint database.