Research On WLAN Based Indoor Positioning Techniques

With the rise of the concept of location based service, people’s demand for radio locationis become higher and higher. In an outdoor environment, GPS has been able to meet the mostpeople’s location needs, but in the indoor environment, as the block of building walls, the GPScan’t work properly, so the study of indoor positioning technology becomes the hot topic ofresearch in the field of radio location. According to the category of indoor positiontechnology, this paper contrasts the method based on sensor technology and signalmeasurement technology, begin with the low cost and wide covering area, and selects theFingerprint Indoor Position Method based on WLAN as the research object.Centering on the research of the Fingerprint Indoor Positioning Method based on WLANmainly studied the access point selection algorithm and positioning feature pretreatmentalgorithm in off-line stage and matching algorithm in on-line stage. Selecting the fairly stableaccess point in the room by means of selection algorithm based on stable access point priority;decrease the influence on the positioning stage by the unstable access point. Experiment bynearest neighbor algorithm in the indoor environment, the experimental results show that theaccess point selected by the algorithm based on stable access point priority is more stable, andthe matching ratio increased by21%than the algorithm based on strongest signal access pointpriority, and the results are more accurate. For small probability signal will be introduced bymean method in off-line stage problem, pretreatment singular values by Pau-Ta criterion, andthe traditional Pau-Ta criterion is improved by change the form of Bessel formula so thatprocess data fast and efficient, the experimental results show that the improved algorithmwork more faster than traditional algorithm in same condition. This paper describes analgorithm based on support vector machine classification, and the matching ratio can reach90%or more. In the last part of the paper, a WLAN fingerprint indoor positioning systembased on support vector machine is designed and implemented, the system can pick outlocations about3meter effectively, and generally achieve the expected effect.