Research On Signal Detection Algorithm In Underwater LED Blue Green Optical Communication

As people continue to explore the ocean,the role of underwater visible light communication technology is becoming more and more important.However,the non-linear effects of various devices in the UVLC system,the background light noise,and the influence of various substances in the water on the light,cause the loss of signal data received at the receiving end of the system,which leads to the system based on Euclidean distance The signal after the decision cannot be demodulated accurately,which reduces the performance of the system.In this paper,the related research is carried out from the aspect of signal detection in UVLC system.The specific work arrangements are as follows:(1)First,it elaborates the research background and significance of UVLC,and the current domestic and foreign research status of the problems in the UVLC system and their solutions.Then the modulation technology used in the UVLC system is introduced and the propagation characteristics of light under water are analyzed.Finally,based on the Beer-Lambert model and considering the influence of other parameters such as the divergence angle and incident angle of the light in the system,an underwater optical communication channel model is established.(2)Research the signal detection algorithm in UVLC system based on machine learning.First introduced the K-means clustering unsupervised learning and support vector machine supervised learning algorithms in machine learning.Then respectively design signal detection models based on K-means clustering and SVM.The operation of these two algorithms does not depend on any underwater channel information,nor does it require a channel estimation module,so their robustness is high.Finally,K-means clustering,SVM and ED detection schemes were simulated and analyzed in different underwater environments.Since SVM detection can adjust the decision line according to the characteristics of the current received signal,and ED and K-means clustering detection are based on ED for judgment,the performance of the UVLC system based on SVM detection is better than that of ED and K-means clustering detection.(3)The signal detection algorithm is verified based on the UVLC system experimental platform.First,the experimental platform of the UVLC system is briefly introduced,and then based on the experimental platform under the set experimental conditions and parameters,different transceiver distances,turbidity,bubble flow rates and QAM modulation orders were carried out in a tank with tap water.Performance test of three detection algorithms.The results show that the effect of SVM-based detection is better than that of K-means and ED,and SVM detection can achieve better signal demodulation,replacing the system's QAM demodulation method,verifying the feasibility and reliability of the SVM detection algorithm.