Performance Study Of Underwater Vortex Optical Multiplex Communication System Based On Neural Network Blind Equalization Algorithm

Vortex optical multiplexing technology can effectively improve the channel capacity of underwater wireless optical communication systems.However,the existence of ocean turbulence will cause cross talk between modes of vortex beams,resulting in a decline in the error code performance of communication systems.In order to overcome the crosstalk caused by ocean turbulent channels,this thesis introduces a blind equalization algorithm based on neural networks.Compared with traditional blind equalization algorithms,the former not only performs significantly better,but also has characteristics such as faster convergence speed and lower average error.Therefore,this thesis investigates the compensation effect of blind equalization algorithms based on different neural networks on the performance of vortex optical multiplexing communication systems in ocean turbulent channels.The main work includes:1.Analysis of the propagation characteristics of vortex beams in ocean turbulence.Based on the ocean turbulence theory and laser propagation theory,the propagation model of vortex optical in the underwater turbulence channel is established.The influence of ocean turbulence on vortex optical transmission,as well as the influence of ocean turbulence on the channel capacity and error code performance of the underwater vortex optical communication system are analyzed through simulation,which lays the research foundation for the following chapters.2.The performance of underwater vortex optical multiplexing communication system based on Back Propagation(BP)neural network blind equalization algorithm was studied.A four channel vortex optical multiplexing transmission system was built,and the compensation effect of the equalization algorithm on the communication system was studied,Numerical simulations were conducted to compare and analyze the constellation recovery,algorithm convergence speed,and system error rate improvement under different factors,with and without the use of this algorithm.3.The blind equalization algorithm based on Radial Basis Function(RBF)and Long Short Term Memory(LSTM)neural network is studied,and it is applied to the vortex optical multiplexing communication system in the ocean turbulence channel.Through simulation analysis,the equalization effects of BP equalization algorithm,RBF equalization algorithm and LSTM equalization algorithm under different turbulence intensity,transmission distance,mode interval,etc.are compared.4.Introduced the blind equalization algorithm of Convolutional Recurrent Neural Network(CRNN)based on Convolutional Neural Networks(CNN)equalization algorithm and the combination of CNN and LSTM for Convolutional Recurrent Neural Network(CRNN).Through simulation,compared and analyzed the performance improvement of the above five equalization algorithms on vortex optical multiplexing communication systems in marine turbulent channels.The results of the study show that all the studied neural network-based blind equalisation algorithms can effectively improve the performance of underwater vortex optical multiplexing communications.However,the selected training sample set is the key to the system performance.When using the neural network blind equalisation algorithm,it is important to combine the specific scenarios and select the appropriate training samples,so as to choose a more effective neural network blind equalisation algorithm to achieve better improvement results.The research results of this thesis provide a theoretical reference for the development of vortex optical multiplexing based communication systems in ocean turbulence,and also accelerate the practical development of underwater optical communication.