Research On WiFi Fingerprint Positioning And Tracking Techniques

In recent years, with the popularization of WiFi technology in wireless network application, WiFi indoor localization has become a hot topic research. A common method for indoor positioning is location fingerprinting technique based on Received Signal Strength (RSS). It has attracted much attention from researchers because of its strong feasibility and low cost.The fingerprint-based method can locate the target accurately in the indoor environment. However, when the number of fingerprints increases, the complexity of positioning algorithm will increase as well. High complexity will lead to a poor real-time performance. A fine real-time performance is very important for locating, especially tracking the target. The traditional positioning methods cannot meet the requirement. Therefore, this paper discusses the RSS based positioning and tracking techniques in depth. The main jobs are as follows:First, it builds an experiment platform in the actual environment and collects a large amount of RSS data. It analyzes the key factors that affect indoor WiFi positioning, including the relation between RSS and position, people’s impact on RSS, the number of detected AP, probability distribution of RSS and so on. It provides a reference to further study.Second, in order to reduce the complexity of positioning algorithm, a hierarchical method is used in this paper. How to divide the locating area is very important. This paper proposes a feasible method, edges-overlapping method. The region of interest is determined by the method of compressed sensing and the minimum Euclidean distance. Experiments are conducted in both the simulation scene and the actual environment. It also compares the edge-overlapping method with affinity propagation clustering. The experiment results show that edge-overlapping method can locate the target more accurately.Last, the tracking technique is studied. It mainly studies the particle filter algorithm. The hierarchical fingerprinting technique is combined with particle filter to implement target tracking, which can not only ensure the localization accuracy, but also improve the real-time performance.