Modeling And Signal Detection Research Of Multiple Input And Multiple Output Underwater Wireless Optical Communication System

ompared with underwater acoustic communication,Underwater Wireless Optical Communication(UWOC)has the advantages of large bandwidth,small delay and high transmission rate.However,UWOC is faced with many challenges,such as absorption,scattering,communication distance,misalignment of transmitter and receiver,etc.,which will affect the performance of UWOC system to some extent.Therefore,this paper aims to model the underwater optical channel,study the characteristics of the underwater optical channel,and study the relationship between the underwater optical channel and water quality,transmitter and receiver parameters,and then optimize the simulation method for large attenuation.Using Multiple-Input Multiple-Output(MIMO)technology,UWOC system can achieve high data transmission rate and high reliability.However,this technology greatly increases the hardware cost,space complexity and power consumption of the system.Optical Spatial Modulation(OSM)technology,as a new MIMO technology,only activates one transmitter in each symbol transmitting period.Both the Modulation symbol and the space position of the activated transmitter carry information,so the data transmission rate is improved.It has the advantages of reducing system power consumption and avoiding inter-channel interference,but its data transmission rate is greatly limited.Therefore,the classical MIMO technology and OSM technology can’t achieve the balance of data transmission rate and channel interference.Therefore,this paper introduces the OSM technology based on the combination of transmitters,which activates multiple transmitters at the same time in each symbol period and transmits information through the combination of transmitters and digital modulation symbols,thus further improving the transmission rate of the UWOC system.However,this method inevitably introduces inter-channel interference.The Bit Error Rate(BER)performance deteriorates.This paper mainly aims at improving the signal detection performance of UWOC system.Firstly,the MIMO underwater wireless optical channel is studied,and based on this,the OSM technology based on transmitter combination is further studied.Finally,the UWOC signal detection technology based on deep learning is studied.The main work is as follows.(1)According to the optical characteristics of seawater,the modeling of underwater wireless optical channel is studied.Firstly,monte Carlo method is used to simulate the underwater optical channel,and the characteristics of the underwater optical channel are studied in detail,including the characteristics of receiving intensity and time domain impulse response.In addition,based on the severely attenuated channel,the method of increasing the aperture of the receiver appropriately is proposed to shorten the monte Carlo simulation time.Finally,multiple transmitters and receivers are deployed at the transmitter end and receiver end respectively,and multiple channel matrices are obtained by changing the placement of the transmitter and receiver,providing data sets for subsequent research on underwater wireless optical signal detection.(2)In order to improve the data transmission rate of UWOC system,this paper introduces the UOSM technology,and further proposes the UOSM technology based on the combination of transmitters,Underwater Optical Generalized Spatial Modulation(UOGSM)technology.This technology combines the advantages of spatial modulation and spatial multiplexing,solves the problems of low reliability and low data transmission rate of single-input single-output system,and alleviates the problems of complex hardware links and large power consumption of MIMO system.However,this technology is at the cost of signal detection complexity and sacrifice of BER performance.For the UOSM system with transmitter combination,the traditional Maximum Likelihood(ML)detection algorithm has high complexity,while the forced zero detection algorithm and the minimum mean square error detection algorithm have low complexity,but the BER performance is not high.Based on deep learning,this paper proposes a neural network structure for underwater optical signal detection,which can achieve the same BER performance as ML detection,but its time complexity is much lower than ML detection.