Research On Tightly-Coupled VLC/INS Indoor Positioning Technology

In recent years,with the significant breakthrough of VLC technology in communication distance and capacity and the popularity of green light source LED,VLC positioning system based on visible light communication system has gradually attracted the attention of researchers.Although the existing VLC positioning system can provide centimeter-scale positioning services indoors,there are strict requirements for the layout of indoor LED light sources.On the one hand,the LED layout should not be too sparse to ensure that there are at least four LEDs in the sensor field-of-view(FOV).On the other hand,the distance between two LEDs should be above a threshold.In addition,when the LED in the FOV is aligned in a straight line,the VLC positioning algorithm will not work,also.The inertial Measurement Unit(IMU)is the sensor that can provide Acceleration and angular velocity measurements,which can be the core of an inertial navigation system that can provide location service alone.However,due to its own dynamic drift,errors will gradually accumulate,so it cannot work alone for a long time.The combination of visible light and IMU can effectively solve the problems faced by the above system when working independently.On the one hand,IMU can provide navigation and positioning results for the system in the short period of VLC positioning system failure.On the other hand,the absolute information of VLC positioning system can dynamically correct IMU drift and effectively improve the accuracy of the system.Firstly,this paper describes the background of indoor positioning,then introduces the research status and challenges of VLC positioning algorithm,and proposes to solve the difficult problems of existing VLC positioning algorithm from the perspective of multi-sensor complementary fusion.The modeling of VLC positioning problem is discussed.Lie group Lie algebra and factor graph model are introduced.The single VLC positioning algorithm and integrated positioning algorithm are emphatically expounded from the perspective of graph optimization.The essential differences between the two algorithms and the possible performance differences are analyzed.A VLC/INS integrated positioning system is designed,and two algorithms are validated on simulation and real platforms respectively.Among them,the performance differences of the two algorithms under different LED layout and different LED number scenarios are discussed.The experimental results show that the single VLC localization algorithm can only work when there are four or more LEDs in the FOV,and it cannot work under the linear layout of the LED.Compared with the single VLC localization algorithm,not only can it continue to work when there are only two LEDs,but also has low sensitivity to the layout of the LED and positioning criteria.It shows that the integrated algorithm can effectively improve the stability and robustness of the current single VLC positioning system,but the cost is a small amount of accuracy.