| Under the background that the country is striving to achieve carbon neutrality,carbon peaking and global energy crisis as soon as possible,LED has become the most ideal light source in the 21 st century by virtue of its advantages of energy saving and low consumption.Visible light communication(Visible Light Communication,VLC)is a new technology that integrates optical communication and wireless communication.It can achieve high-speed data transmission while completing indoor lighting.It has been included in the key development target of 6G by the country.In this thesis,visible light technology will be studied,focusing on channel modeling and visible light positioning in irregular scence,and combining computer vision technology and VLC technology to calculate the optical power distribution of visible light reflected on the wall by using real wall data.The communication link of indoor visible light is divided into direct link(LOS)and non-direct link(NLOS),The influence of ideal indoor and irregular indoor scenarios on visible light communication mainly comes from NLOS link.In this thesis,through the research and analysis of the reflection law of light on the wall,it is found that under ideal indoor conditions,the normal vector of each reflective element on the wall is perpendicular to the wall itself,so the reflection angle of light is always equal to the incident angle,which is called ideal In an irregular indoor environment,each reflective element is in a different plane due to the unevenness of the wall,so its normal vector is not always perpendicular to the wall itself,resulting in irregular reflection of light,so The ideal indoor channel model will not be suitable for irregular scence,taking this as an entry point,modeling visible light channels in irregular scence.In order to reflect the wide applicability of the established channel model,this thesis uses 3D point cloud technology to scan the wall with 3D scanning equipment to obtain point cloud data.The original point cloud data consists of tens of thousands of sets of 3D coordinates,which is very large and Disorderly,first,the kd-tree technology is used to find the neighborhood of the point cloud data to determine its spatial topology,and then the principal component analysis method(principal component analysis method)is used to estimate the real normal vector of each point cloud,that is,the real wall surface.Finally,the optical power distribution of the VLC channel model established based on real data is calculated.The results show that the optical power fluctuates in the range of-7.360 d Bm to-25.431 d Bm.Traditional visible light positioning algorithms,such as TOA,TDOA,and AOA,require high sensor accuracy,complex hardware facilities,and high cost;RSSI positioning algorithms are easily affected by background light and diffuse reflection links,resulting in poor positioning accuracy.Aiming at the disadvantages of traditional localization algorithms,this thesis proposes a localization algorithm based on Genetic Algorithm-Optimized BP Neural Network(GA-BP)in irregular indoor scence.Considering the influence of the diffuse reflection link,in the 5m×5m×5m irregular indoor simulation scene,the GA-BP algorithm greatly improves the positioning accuracy compared with the BP neural network,and the root mean square error of the positioning is 2.32 cm. |
PDF Full Text Request