Polarization Code And Orbital Angular Momentum Modulation In Space Optical Communication

Free space optical communication technology has the advantages of both wireless communication and optical fiber communication,and can flexibly provide wireless access and large-capacity data transmission.In space optical communication,the use of orbital angular momentum beam transmission has the advantages of high capacity and confidentiality.However,due to the influence of atmospheric turbulence,the optical signal is prone to intensity attenuation and phase disturbance,which greatly reduces the reliability of space optical link transmission.For the impact of turbulence in space,on the one hand,error correction coding can be introduced to suppress interference to the channel.As an encoding technology with reachable channel capacity,polarized codes can be applied to provide highly reliable transmission in space optical communications.On the other hand,it can also improve the detection of distorted optical signals and enhance the ability to receive and demodulate optical information.At the same time,deep learning technology has revolutionized the fields of image analysis and audio processing,and has continued to merge with other disciplines to generate new methods and ideas.It has theoretical significance to apply deep learning technology to coding and demodulation in space optical communications.And application prospects.In this thesis,the optical communication scheme combining polarized code and OAM is researched,a complete communication system model is constructed,and a phase screen transmission simulation model is established to analyze the influence of polarized code on the performance of OAM communication system.Based on this,the entire communication scheme is optimized,and a deep learning-based OAM demodulation method is proposed to reduce the complexity of OAM demodulation detection.The system performance of polarized codes combined with deep learning demodulation OAM is analyzed.Finally,the polar code encoding and decoding research is carried out,and a deep learning-based polar code encoding and decoding scheme is proposed,which can realize the low complexity and high reliability of the entire communication scheme.The main results are as follows:(1)The communication scheme combining polarization code and OAM modulation is studied,the system composition is analyzed,the transmission simulation model of OAM beam is studied,and the performance of polarization code in OAM beam transmission system Analysis.(2)The OAM demodulation detection method based on deep learning is studied,and the OAM intensity image data set is constructed,and the influence of different mode combinations,image sizes,network structures,etc.on demodulation performance is analyzed.An efficient OAM demodulation detection model.Finally,the performance of an OAM transmission system with polarized codes combined with deep learning demodulation is simulated and analyzed,and digital images are transmitted to verify the feasibility of the system.(3)The polar code coding and decoding algorithm based on deep learning is studied.Based on the judgment of the turbulence intensity of atmospheric channels,an adaptive coding algorithm combined with channel estimation is proposed to adjust the coding length to adapt to different turbulent states and improve the system.Transmission performance.The DNN model and RNN model are used to analyze the decoding of polar codes.The effects of different network structures on the decoding performance are compared.It is found that the RNN model can approach the performance of the SC decoding method while reducing the complexity of decoding.The end-to-end encoding and decoding form is studied,and the maximum mutual information amount after encoding passes through the channel is used to train the network to optimize the performance of the polarized code at a low SNR.