Research And Implementation Of Indoor Positioning Algorithms Based On Wi-Fi

With the rise of smart city construction and the rapid development of mobileInternet, people’s demand for location based services is growing fast, which requiresit to achieve accurate positioning and tracking. Currently, positioning technology iswidely used in various fields, and the success application of outdoor positioning basedon the GPS provides an incentive to the research and development of indoorpositioning systems. However, due to its weak signal reception, GPS systems can notbe effectively used inside buildings and dense urban areas. The current indoorpositioning systems mainly use technologies such as computer vision, infrared, RFID,ultra-wideband, wireless sensor networks, AGPS, but these systems often requiredeployment of additional facilities, which limit the applications. Wi-Fi based indoorpositioning system can take advantage of existing infrastructure, and be used toprovide indoor location services, without the need to deploy additional equipment.And smart phones also have built-in Wi-Fi modules, which make Wi-Fi based indoorpositioning becomes possible. Therefore, the use of smart phones take advantage ofWi-Fi signal throughout buildings to position has become a potential indoorpositioning technology.Based on sufficiently study of the Wi-Fi indoor positioning technology, thispaper proposed corresponding improvement algorithms for the inadequacies of theexisting fingerprint-based Wi-Fi positioning algorithms, then design and implementthe Wi-Fi indoor positioning system. Firstly, Wi-Fi received signal strengthcharacteristics and influencing factors are analyzed, including RSSI probabilitydistribution and the relationship between RSSI and distance, and the influence ofbody orientation and different devices. Then, start from the construction andworkflow of a positioning system, to explore the data collections to build the locationfingerprint database, preprocessing of real-time positioning stage, the reference APselection, signal distance, nearest neighbor selection and positioning resultscalculation. In the results calculation phase, the densest nearest neighbor algorithm isproposed, which can get better positioning accuracy compared to traditional centroidmethod. After that, the role of the Kalman filter in dynamic positioning and tracking isanalyzed. By phone sensors, we can to determine the changes of user motion state,and effectively improve the dynamic positioning and tracking results by using different Kalman parameters. Next, by studying the positioning with phone sensors, afusion algorithm combines Wi-Fi positioning with sensor positioning is proposed, wecan effectively compensate for the lack of Wi-Fi positioning by the use of sensorswhen Wi-Fi positioning is not possible. It can also make better corrections whileexists high fluctuations in the normal Wi-Fi positioning process, which effectivelyimprove the dynamic positioning and tracking results. Finally, a fingerprint Wi-Fiindoor positioning system is designed and implemented on Android mobile phoneplatform. Several modules of the system are analyzed. Experimental results show thatthe proposed algorithms can achieve high-precision Wi-Fi indoor positioning moreeffectively.