| In recent years,underwater wireless optical communication(UWOC)has received more and more attention due to its high bandwidth and high communication rate.However,the transmission distance of UWOC is severely limited due to the absorption and scattering of the underwater environment.In order to realize long-distance UWOC,Single Photon Avalanche Diode(SPAD)with photon limit sensitivity has been studied to detect optical signals at the receiving end.The key to realizing data demodulation at the receiving end is to recover synchronous clock and baseband signal,and the recovery of synchronous clock is the basis for realizing the recovery of baseband signal.Because SPAD outputs discrete single-photon pulses,and the photon detection process is affected by strong shot noise.Therefore,the key point and difficulty of realizing long-distance photon-counting UWOC lies in the recovery of synchronous clock,but there is currently no systematic solution for synchronous clock recovery of photon-counting UWOC.Under this background,this paper focuses on the fast recovery technology of synchronization clock for long-distance photon-counting UWOC.The main research contents and results are as follows:1.Based on the statistical distribution of the LED emission light field,the attenuation process of the underwater channel,and the non-standard Poisson random process of photon detection,the photon-counting UWOC system model is established.According to the system model,the On-Off Keying(OOK)modulation is taken as an example to study the photon-counting UWOC,the bit error rate(BER)and transmission distance of the system are analyzed by theoretical formula derivation,and the influence of relevant parameters on the BER and transmission distance of the system is studied by simulation.2.A synchronous clock recovery scheme based on multi-channel gated photon-counting is proposed,and a model of the synchronous clock recovery scheme is established.Based on Pulse Phase Modulation(PPM),Monte Carlo simulation and experiment are used to analyze the influence of the length of synchronization header sequence,the number of gated clocks,and the time slot frequency on the communication performance of the system.The experimental results show that the multi-channel gated photon-counting scheme proposed in this paper when the time slot frequency is 1MHz,the photon-counting UWOC with 0.875 photons/bit and the Symbol Error Rate(SER)of only 1.8×10-3 can be realized.3.Aiming at the problem of further improving the precision of the recovered synchronization clock,for the first time,deep learning is combined with photon-counting UWOC,and a synchronous clock recovery method based on deep learning is proposed.According to the system model,a scheme of using Monte Carlo method to make training data is proposed.Two neural network models for realizing synchronous clock recovery are designed based on regression problem and classification problem.The simulation results show that this method can effectively improve the precision of the recovered synchronization clock and reduce the BER of the photon-counting UWOC system.4.In order to verify the feasibility of the synchronous clock recovery method based on deep learning in the actual system,a photon-counting UWOC experimental platform based on deep learning is designed and built,and the influence of related parameters on the system communication performance is tested through experiments.Experimental results show that when there are only 8 photons in the time slot,photon-counting UWOC with a baud rate of 1 Mbps and a BER of 5.4×10-4 can be achieved. |
PDF Full Text Request