Study of Indoor Positioning Method Based on Combination of Support Vector Regression and Kalman Filtering

Against the problem that indoor positioning suffers quite large errors and irregular user location movement, this paper adopts Support Vector Regression (SVR) for initial positioning and Kalman filtering for filtering of the positioning results so as to improve the accuracy of the positioning system. The experimental results show that against the real WLAN environment, SVR positioning results processed by Kalman filtering indicates the root mean square error is decreased by 16%, and 73% of positioning accuracy within 2 meters is increased to 83%.