Laser Communication And Ranging Integration Algorithm And Its Combination With Neural Networks

In satellite communication,the communication capacity of traditional microwave link faces new challenges in the inter-satellite link of the global network and the satellite-ground broadband communication network.Satellite laser communication provides a alternative solution to the rate bottleneck in satellite communication.The space laser communication and ranging integration technology realizes the dual functions of communication and ranging on the same laser beam and hardware platform,which can simultaneously satisfy high-speed data transmission and high-accuracy distance measurement.The laser communication and ranging integration technology is an important research direction in inter-satellite network and satellite navigation.At present,the laser communication and ranging integration technology has been successfully implemented in the Lunar Laser Communication Demonstration(LLCD)system in the United States.But this technology in China is still in the orbital test stage,lacking mature applications.This thesis will focus on the key algorithms for the laser communication and ranging integration technology of the inter-satellite link.And the hardware implementation and experimental verification are carried out.This thesis summarizes the research status of laser communication and ranging integration technology,and proposes a high-speed laser communication and high-precision ranging integration technology.This technology is based on the principle of Dual One Way Ranging(DOWR).And BPSK modulation and DP-QPSK coherent demodulation are adopted.The communication scheme realizes the communication rate of 1 Gbps;The distance measurement is performed by inserting the ranging information into the communication frame,and the accuracy of ranging is achieved by high-precision time measurement assisted by the symbol phase.The ranging accuracy is better than 5mm.The integrated schemes in this thesis are all implemented on the Field Programmable Gate Array(FPGA).The high-speed parallel design of the symbol synchronization loop in the integrated implementation technology is carried out,and the ground implementation experiment is carried out on the overall implementation scheme.Dynamic neural network is a neural network with Tapped Delay Line(TDL).It processes not only the current input and output signals,but also the historical input and output signals.The dynamic neural network has a strong approximation fitting ability.It can play an important role in channel equalization of memory channels and time-varying nonlinear channels.The NonlinearAutoregressive Network with Exogenous inputs(NARX)network with simple structure and out-put delay feedback is an important supplement to the traditional channel equalization algorithm in the equalization of complex channels.In this thesis,the channel equalization performance of real-valued NARX neural network in fixed multipath channel is simulated and compared with the traditional Recursive Least Squares(RLS)equalization algorithm.An RVNARX-Decision Feed-back Equalizer(RVNARX-DFE)neural network model is proposed in this thesis,using decision feedback to retrain RVNARX-DFE in time-varying channels,and the Mean Square Error(MSE)with error weight is used as the loss function.The equalization effect of the RVNARX-DFE model is simulated and analyzed in the Rayleigh channel with frequency selective fading.