Research On Machine Learning Aided Signal Detection Technology Of Indoor Spatial Modulation Visible Light Communication System

Visible Light Communication(VLC)systems have become an emerging research field of optical wireless communication by using Light Emitting Diodes(LEDs)as transmitters for signal transmission.In view of the shortage of traditional radio frequency communication spectrum resources and high communication power consumption,the VLC system can be used as an effective supplement to the existing RF communication by combining lighting and communication.As a new type of multi-antenna technology,spatial Modulation(SM)is used for data transmission by activating different transmitting antennas in each time slot.Because of its high spectral efficiency and low complexity,it has received extensive attention and applications.At the same time,SM's special transmitting structure is particularly suitable for indoor VLC systems,which can effectively improve the transmission efficiency of the system.In addition,the partial activation feature of SM can also effectively reduce the search space of the receiving end machine,making it easy to combine with machine learning algorithms,reducing the complexity of signal detection,and effectively improving the working efficiency of the system receiver.In view of the above problems,this article focuses on the SM signal detection problem in indoor VLC systems,and fully considers the system structure of SM-VLC and the characteristics of machine learning algorithms.By combining machine learning algorithms with SM technology,and from the perspective of optimizing system performance and reducing algorithm complexity,a new type of high-efficiency signal detection algorithm is proposed.The main work of this thesis is as follows:Firstly,it introduces the basic theory of VLC system structure,channel model and noise,and analyzes the modulation principle and system model of SM technology and its corresponding mainstream extended model,which provides the theoretical basis for the following application of SM technology in indoor VLC system.In addition,this thiesis also systematically analyzes several existing common detection algorithms for SM signals,and carried out simulation analysis to summarize the advantages and disadvantages of different algorithms.Secondly,for the indoor SM-VLC system,the SM signal detection algorithm based on supervised learning is discussed in depth.In order to reduce the complexity of the algorithm and improve the detection efficiency,based on the Support Vector Machine(SVM)classification detector,the duality theory is used to optimize the SVM classifier.This thesis proposes an efficient SVM classification and detection algorithm suitable for indoor SM-VLC systems,and analyzes the computational complexity of the algorithm.Based on this type of algorithm,signal detection algorithms for two special SM modulation systems,Generalized Space Shift Keying(GSSK)modulation and Generalized Spatial Modulation(GSM),are respectively given.The simulation results show that the signal detection algorithm proposed in this thesis has lower computational complexity on the basis of the bit error rate performance close to the optimal detection algorithm.Thirdly,based on the unsupervised detection algorithm of SM signal in RF communication system,combined with the channel characteristics of indoor SM-VLC system,this thesis proposes a corresponding SM signal detection algorithm based on unsupervised learning.At the same time,from the perspective of bit error rate performance and algorithm complexity,in the indoor VLC system model,the performance simulation analysis and comparison of the VLC system based on GSSK and GSM technology are carried out by selecting different unsupervised learning algorithms,which proves the effectiveness of the proposed algorithm.Fourthly,a detailed theoretical analysis of the channel correlation of the indoor VLC system was carried out,and the influence of system parameters on the channel correlation was verified by simulations,which provided support for the construction of the indoor VLC system model.In addition,the influence of channel correlation on the the proposed signal detection algorithm is analyzed by simulations.