Research On Optimization Of Spinal Codes' Coding And Decoding In Free-space Optical Communication

Free space optical communication(FSOC)is an emerging communication technology which uses light to transmit the data in free space.Compared with the RF communication and optical fiber communication,it has the advantages of unlicensed spectrum,high data rate transmission,high security,etc.,which has been widely studied in recent years.However,when the optical beam is transmitted through the atmospheric channel,the signal is attenuated due to the factors like atmospheric turbulence,scattering and absorption and other factors,among which atmospheric turbulence has the greatest influence on the optical beam.Atmospheric turbulence will disturb the light in the transmission process and seriously depress the communication quality.Therefore,it is necessary to adopt effective error control technology to ensure the stability and reliability of free-space optical communication system.Proposed by J.perry et al.,Spinal codes are a new class of rateless codes which can achieve the Shannon capacity over both additive white Gaussian noise(AWGN)and binary symmetric channel(BSC)models.Its encoding and decoding structure is simple,and the decoder complexity is low.The continuous use of pseudo-random Hash function in Spinal codes coding process gives it strong anti-interference performance in low SNR condition.Moreover,Spinal codes are capable of generating an infinite number of code symbols,which is truly rateless compared with Raptor codes and other rateless codes that limit the maximum number of code symbols.These characteristics make it has a good application prospect in the free space optical communication system.In this paper,the factors that attenuate the optical beam in atmospheric channel are introduced,especially the atmospheric turbulence,and the corresponding system model are introduced.Next,the encoding decoding strategy of Spinal codes are briefly introduced,in view of the Spinal codes transmission in turbulent channel is subject to outburst error which leads to system performance degradation,we put forward interleaving into Spinal codes' encoding and decoding process.After the passes produced by the encoder are interleaved by the interleaver,they are sent to the receiver,the information received by the receiver are deinterleaved and decoded.Considering the parallelism of the decoding cost calculation in Spinal code decoding process,this procedure is parallelized on GPU to reduce the time complexity.Finally,the outburst error correction strategy for Spinal codes based on an interleaving technique is simulated and analyzed.The results show that compared with the original Spinal codes,the Spinal codes based on interleaved technique has higher burst error correction capability in additive white gaussian channel and weak turbulent channel than the former.The number of Passes involved in successful decoding of interleaved Spinal codes is less than that of the original Spinal codes,and the average bit rate is also about 0.4 bit per symbol to 3.8 bit per symbol.In addition,the performance fluctuation of the interleaved Spinal codes in atmospheric turbulent channel is more than half that of the original Spinal codes.