The Research On WIFI Indoor Location Algorithms Based On Position Fingerprint

Recently years, with the rapid development of wireless communication and computer technology, indoor positioning technology have got great attention from all sectors of society, people’s requirement for location service becomes more and more eagerly.The deployment of WIFI hotspots indoor is extensive and intensive now, and using the WIFI network to location becomes one of the main ways to realize the indoor positioning. WIFI positioning method based on the position fingerprint as itself requiring little hardware, low cost, high positioning accuracy, which becomes one research hotspot of positioning technology.The article makes full use of the serious multipath effect indoor which caused the phenomenon the different location receives different signal strength value, and the signal intensity vector can as the only fingerprint indentify position. Through the comparison of various algorithms and practical observation, verified the efficiency of location fingerprint algorithm, and the positioning accuracy can reached more than 90% when the small area can meet the signal coverage. Paper’s main work is as follows:First, choose and deploy hotspots according to the area where we want to make orientation, build a database based on the relationship between signal strength and sampling point, which is the database of location fingerprint.Second, processing the data, after analyzed the change of signal strength with the time at the same location, treatment of a variety methods of signal processing, finally uses the selection of singular value processing, Kalman filter and synovial membrane average filtering algorithm, use the high probability method to the signal which after filtering before positioning, then resave the fingerprint data which after processing in database.Finally, proceed the real-time positioning, analyzed the defects of original Nearest Neighbor algorithm then optimized it, proposed Weighting Nearest Neighbor positioning algorithm. Comparing the positioning effect of Nearest Neighbor algorithm, Weighting Nearest Neighbor algorithm and the fuzzy matching algorithm, the experimental results shows that the positioning effect of Weighting Nearest Neighbor positioning algorithm is the best method. In addition, adding the process of data processing makes the higher positioning precision and accuracy, reached that we expected only using the least common hardware almost purely software, which highly decreased our cost.