HMM Based Indooor Floor Localization Research

With the rapid development of information technology, mobile positioning hasbecome a great demand of people’s daily life. Research for outdoor GPS positioning isincreasingly mature, people are keen to realize the "seamless joint" of indoorpositioning and outdoor positioning for convenient life and work. In recent years,researchers at home and abroad have done a lot of indoor positioning work, putforward various of different schemes based on signal carrier and positioningalgorithms. But the floor localization in indoor localizationin field is still at the initialstage. Many research ignored to the determination of floor, or artificially specifiedfloor directly, or use hardware devices for elevation information. Floor localizationhas its special applied value and the user requirements, so carrying out research usingsoftware solutions, algorithm of floor in the future is necessary.Aiming at floor localization, this paper first summarizes the limited literature athome and abroad in recent years, puts forward a FLOC floor localization schemebased on RSSI signal analysis, and abstract floor localization into data analysis model.In this paper, the main contribution and innovation are as follows:(1) As for the phenomenon of lack of study in floor localization, this papercomprehensively analyzes the existing indoor and floor positioning, analyzes itstechnical characteristics, and points out the insufficient.(2) Combining with the characteristics of the WiFi signal RSSI values across thefloor, this paper quantitatively analyzes the floor layer plate blocking effect, andproposes a FLOC solutions based on signal analysis;(3) In this paper, we creatively abstract floor localization into data analysisproblem, apply hidden markov model, and cluster analysis to floor localization forthe first time, improve forward-backward algorithm to fit the FLOC, evaluate themodel performance through comparative analysis of four groups of parameters,thepositioning precision can achieve100%under the specific parameters;(4) Validate FLOC on the Android platform, implement through collectingcrowdsourcing and using off-the-shelf access point facilities, which is non-site surveyand the cost is low.(5) Develop a fire-escape APP which employs the FLOC scheme,the experiment shows combining floor positioning in locating application is more flexible andintelligent than traditional manual settings.