Research On Indoor Localization Technology Based On Wi-Fi

With the popularity of smart phones and the development of mobile Internet, the abilityto navigate persons and devices in indoor environments has become increasingly demanding,especially in complex indoor environments such as airport halls, exhibition halls, warehouses,supermarkets, libraries, underground parking, mines and other environment often need toidentify the mobile terminal or the holder, the location of the room facilities. At domesticand overseas about large companies and research institutes have developed a number ofindoor wireless positioning system, mainly using infrared, ultrasonic, Bluetooth, WIFI, RFID,UWB and other short-range wireless technology.It takes much attention that WIFI based indoor wireless positioning system. BecauseWIFI networks where wireless location-based technology is widely deployed and cost less.The earliest WIFI based positioning system was developed by Microsoft which is calledRADAR system. WIFI based wireless location positioning method commonly usingfingerprint can usually achieve better positioning effects. But at this stage the position of thefingerprint in the offline stage positioning method requires a lot of manpower and materialresources, needing large number of the reference position, and in each of the reference pointin all directions collected a lot of data to establish the location fingerprint database. Usuallytakes a lot of time and labor costs, so we have great trouble in the popularity of indoorlocation system.In order to solve this problem, we propose the WIFI based indoor positioning methodusing Gaussian process regression. We mainly study how to establish the fingerprint databasewith little collection information and make great efforts to improve positioning accuracy; howto minimize the offline phase time and manpower consuming. This algorithm makes full useof the existing small number of WIFI fingerprint data, predicting unknown WIFI fingerprintsthrough training, losing little accuracy while greatly reducing the positioning system in theoffline stage time consuming. This paper also presents a RSSI based spatial clusteringtechniques to solve the problem of false targeted areas to improve positioning accuracy.We design an APP based on Gaussian process regression WIFI indoor positioning systemin Android platform. The system consists of server program and Android APP. Androidsmartphones can achieve a positioning application APP, positioning information datacollection, file upload and download, fingerprint location database matching and positioningdisplay. The windows platform combined with MATLAB mixed programming to set up IISserver to achieve fingerprint database file location The training file upload and download and fingerprint location database matching. Finally, the system was tested under differentconditions with the effects of different router deployment layout on the positioning system.