Research On Modulation Format Identification Technology Of Satellite-Ground Laser Communication System

The requirement tendency of forthcoming satellite data transmission has grown to become a critical boost to the progress of communication technology.Gigantic amounts of information and complicated channel circumstances require high data speed and flexibility at the same time.Laser communication is a kind of wireless communication with laser as the carrier.It has many advantages,such as large capacity,high speed,high security,antiinterference and so on,which are difficult to realize in traditional microwave communication.Therefore,it has aroused extensive research interest in the industry.In satellite communication system,transmission of high-order modulation format can gradually be realized,which can improve spectral efficiency and data transmission rate.The next-generation satellite-to-ground laser communication network is expected to adjust the modulation format dynamically according to the link conditions and terminal equipment configuration to meet the different requirements of the terminal system and service.To correctly identify the modulation format of the received signal,it is necessary to let the receiver specify the modulation format information.The modulation format information and effective information can be transmitted at the same time,but it will occupy a certain amount of spectrum resources,and the spectrum gain brought by the higher-order modulation format cannot be reflected.Therefore,in order to maximize the transmission efficiency,the receiver needs to have the ability to realize modulation format identification(MFI)without prior information,which is a challenging requirement.The main work and innovations of this paper are as follows:(1)A modulation format recognition simulation scheme based on Convolutional Neural Networks(CNN)in satellite-to-ground laser communication is proposed.The simulation simulates signals with four modulation formats of OOK,BPSK,QPSK,and16 QAM through dynamic time-varying channels,and uses CNN for training,verification,and testing to verify the performance of the CNN structure in modulation format identification for dynamic channels.(2)For the scenario of dynamically switching between mPSK and mQAM multi-order modulation formats in laser communication,an atmospheric coherent detection system was built using VPI simulation software.Six modulation formats were generated,namely BPSK,QPSK,8PSK,16 QAM,32QAM,and 64 QAM.Data passing time-varying channel was trained and verified in the CNN model of the MFI scheme,and the recognition results were used to select the correct demodulation method for the DSP module.It has been verified that it is also robust to the data under untrained channel conditions.(3)In the indoor coherent optical communication experimental platform,8.9kilometers of outfield test data were used to reproduce the outfield channel conditions.The received signals of six modulation formats,BPSK,QPSK,8PSK,16 QAM,32QAM,and64 QAM,were connected to a CNN trained with simulation data for blind recognition.This achieved modulation format identification in laser communication systems without prior information under actual turbulent channel conditions and verified the compatibility of the model by simulation data with actual communication link data.