Indoor Localization Method Based On WiFi Fingerprint Refinement

Indoor localization is a prerequisite for LBS(Location Based Services), is research hotspot in recent years, and received widespread attention from academia and industry. With the popularity of smart phones, many kinds of indoor localization method which is based on WiFi and Smart Phone have been proposed. Compared with the previous localization method, indoor localization based on WiFi and Smart Phone has advantages, but still exits some problem such as the cost of manual collection is higher.This thesis summarized and compared from the key factors for large scale apply of indoor localization, the basic method of indoor localization based on WiFi and smart mobile phone and classification of indoor localization, and then pointed the issues and the future trends in this field. Then, we propose a location method based on fingerprint refinement according to the WiFi indoor distribution characteristics and the divide and conquer technique. This method can achieve idea positioning accuracy while greatly reduces the cost of collecting samples. Finally, this paper implements a simulation system for indoor positioning, it can be used to visual analyze and process sampler data, and also can be used to refined sample data.In this thesis, we use refinement data and simulation system to do the experiment. The experiment show that the higher precision sample data caused the lower average distance, when use the data refined to 0.25 meters and the K-nearest neighbor matching algorithm(K=4) the average positioning error distance is 1.82785 meters. Thus show that use our method can achieve idea positioning accuracy while greatly reduces the cost of collecting samples.