An Indoor Positioning Algorithm Based On WiFi And Sensor Data Fusion

Due to the popularity of WiFi infrastructure in indoor spaces and the rapid development of intelligent terminal devices, brings the new vitality of indoor positioning. But at the same time, indoor environment is complicated and various, as well as the heterogeneity of mobile intelligent terminal devices, these factors make Indoor positioning method based on WiFi achieve poor accuracy of positioning.This paper thinking about designing methods to improve indoor positioning accuracy by using indoor WiFi signal and intelligent terminal built-in multi-sensor data aiming at these problems. Firstly, we designed a stable indoor positioning method based on WiFi by using weighted RSS which aiming at the complicated indoor environment and the heterogeneity of mobile intelligent terminal devices. On this basis, we designed a novel finger vector for positioning which has the advantages of stable and calibration-free. Secondly, we Analysis the problems and deficiencies of the current indoor positioning methods based on WiFi in multi floors environment, proposed a multi floor positioning method based on floor discriminant, build a floor discriminant model with Fisher Linear Discriminant. We also predict floor change events by discriminate the states of people, such as walking, stairs and elevator. Then we experiment our method in a building of six floors. The positioning accuracy improved 6% on average. Thirdly, we proposed a hybrid multi floor positioning method on the basis of above work. Experiments in actual environment show that, our approach not only improved the positioning accuracy but also save energy consumption by reduce positioning time-consuming. Finally, we develop a multi floors navigation system based on the research above.