Design And Implementation Of Fused Indoor Positioning System Based On Photodiode And Camera

At present,with the development of Internet of Things technology and the performance improvement of the mobile devices,indoor positioning technology has been widely applied to the mobile robots positioning,goods tracking in the large indoor places and other applications,which brings convenience to people's lives.In recent years,Visible Light Positioning(VLP)has attracted much attention in the field of indoor positioning due to its low cost and high accuracy.VLP systems can generally be divided into two categories:photodiode-based(PD-based)and camera-based positioning systems.However,in the PD-only systems,the visible light signal is susceptible to the interference from external ambient light,which reduced the positioning accuracy;and in the camera-based VLP systems,there is a strict requirement for the camera's resolution and viewing angle,which is difficult to meet in the actual positioning environment.Considering the problems in the previous works,a low-cost,high-accuracy,semi-real-time fusion VLP system combining both PD and a cheap camera is proposed in this thesis.In this thesis,the relationship between the ambient light and the received signal strength from PD is analyzed firstly by experimental results.A two-step modified channel model is designed to reduce the influence of ambient light and modify the Lambertian coefficient to improve the accuracy of PD-based positioning measurements.Then,the camera-based positioning measurements are extracted by analyzing information including positions and angles in the pictures captured by the camera.Such measurements are invulnerable to the ambient light and have higher accuracy than PD-based measurements.Moreover,when the camera captures a rectangular pattern on the ceiling of the positioning area,a new single-lamp positioning algorithm based on the ceiling rectangle pattern is proposed to further increase positioning accuracy.Finally,a novel Fixed-Lag Ensemble Kalman Smoother(FLEnKS)algorithm is proposed in this thesis for VLP systems,which contains a sliding window and a two-filter structure.The PD-based measurements after two-step modification and the camera-based measurements are then fused by the FLEnKS to achieve a high-accuracy and semi-real-time positioning solution.In this thesis,a real experimental environment is built and dynamic experiments are conducted in the 5×5×2.84m~3 positioning area.The experimental results show that the 2D mean positioning error of the proposed positioning solution is 8.67cm and the 3D mean positioning error is 12.15cm in the case of external environmental interference.Compared with the existing PD-only visible light indoor positioning system,the positioning accuracy has been improved by 60%~70%.The proposed fusion VLP solution based on PD and camera realizes low-cost,high-accuracy and semi-real-time indoor positioning.