Research On Algorithms Of Location Fingerprinting Positioning Technology Based On WiFi For Application

Position location systems are becoming increasing important as add-ons to today’s pervasive wireless technology. As signals of GPS and BDS cannot reach the indoor environments.Wi Fi positioning technology has become popular as a kind of wireless location technologys.Location fingerprinting technique using WiFi infrastructure have been proved to be the most accurate positioning algorithm compared with other techniques like time or direction of arrival technique used in WiFi location systems.This technique does not require any specialized hardware other than a moble device(such as smartphone, IPad)with wireless nework interfaces which could receive RSSI measurement. It has the advantages of low cost, hign precision and broad applicability.In this paper, WiFi location fingerprinting positioning technology was studied. Firstily analysis the research status of the indoor positioning technologys.Then analysis advantages of WiFi indoor positioning technology in present popular indoor positioning technologys.At the last, analysis the feasibility of using WiFi location fingerprint positioning technology for application.The work has been done in this paper can be shown as follows:The first section. analyze the statistical properties of RSSI, and according to the special environment of the parking lot(the height of user’s mobile phone is near to the average height of cars), creatively found the relationship between RSSI and the height of signals acquision states.The second section.Use the Eclipse software to develop the Android application, using mobile phone with the Android system to obtion WiFi signals at the reference points which locations have been known.Put the obtained RSSI values belong to dfferent WiFi APs in the mobile phone’s database.The third section. Research on the filtering algorithms: gaussian filter, logarithmic filte, kalman filter,adaptive kalman filter and particle filter, then make the adaptive kalman filter fusion particle filter and compare with those filtering algorithms, finally select the best filter to filter the obtained RSSIs,which used to build location fingerprinting database.The fourth section.Study and improve the clustering algorithms in fingerprints: based on the strongest AP clustering algorithm in combination with K-means clustering algorithm to partition the fingerprint data stored in the database,this new clustering algorithm could reduce the complexity of matching appropriate reference points, thus achieve the effect of fast matching.The fifth section.Comparing the several kinds of positioning methods: use several different deterministic algorithms respectively to calculate the location at undetermined site, then analysis the corresponding positioning errors.In this paper, the research results have been verified can improve the accuracy of Location fingerprintingsystem based on WiFi,which has broad application values.