| Wireless optical communication has the advantages of large capacity,wide frequency band,low power consumption,good secrecy and flexible installation.It is considered as an effective method to solve the“last mile”problem of communication.However,the various scattering elements in atmosphere will disturb the transmission light beams,which makes the receiver end contains direct and multipath scattering signals.Compared with the direct signal,the multipath scattering signal will generate times delay due to the different transmission paths,which will lead to inter-symbol interference.The performance of wireless optical communication will be greatly reduced in the rain,snow,sand dust and other inclement weather conditions especially.In order to support optical communication,channel equalization and signal detection are one of the effective measures to improve system performance.However,during the equalization and signal detection,channel state information is usually needed,which requires the accurate estimation of channel state parameters.How to estimate and the accuracy of parameters are directly related to the improvement of system performance.In this thesis,based on this consideration,the fading statistical characteristics of the atmospheric channel are analyzed.On this basis,the influence of different factors on estimation accuracy and system bit error rate of implicit training sequence based channel estimation are discussed emphatically for intensity modulation/direct detection?IM/DD?modulation.?1?The complexity of atmospheric channel will have a great influence on the performance of communication system.Thus,the scheme of training sequence based channel estimation in the sand dust environment is researched.The impacts of the number of scattering path and the length of training sequence on estimation performance are analyzed.The results show that the number of scattering path is less than 5,and the length of m sequence is 63,which leads to the better estimation performance under the MMSE criterion for the low signal-to-noise ratio channel environment.The research provides a theoretical basis for constructing the implicit training sequence based channel estimation scheme.?2?A channel characteristic method based channel estimation scheme is proposed.The scheme is to figure out a solution to the problem of direct current?DC?bias by using implicit training sequence.On this basis,the mean square error?MSE?,bit error rate?BER?and computational complexity of the scheme are analyzed.Moreover,it is proved that the power allocation factor is the key element influencing the estimation performance.The optimal power allocation factor is derived through the criterion of maximum output signal-to-noise ratio.The results show that the MSE of the proposed scheme can be reduced by one half approximately compared with the conventional implicit training based estimation method.When the BER is 4×10-3,the signal-to-noise ratio can be reduced by about 2 dB and 6dB while the DC bias is 0.2 and 0.4,respectively.However,the computational complexity is only increased by 14.2%.?3?To further improve the estimated precision of DC bias,and to eliminate the influence of data information on implicit training sequence based estimation,an excellent date dependent superimposed training?DDST?based correlation matching algorithm?CMA?scheme is proposed.Among them,the DDST method is used to eliminate the interference of superimposed data information,and the CMA method is used to eliminate the interference of DC bias.The results show that the MSE of the proposed scheme is closer to the situation of non DC bias interference compared with the conventional DC bias elimination method.When the BER and the DC bias are 2×10-33 and 0.4 respectively,the proposed scheme has more than2 dB performance improvement.However,the computational complexity is only increased by22.8%.The research results of this thesis have the significances to improve the performance of wireless optical communication system using implicit training sequence based channel estimation over hostile channel environments in practical engineering applications. |
PDF Full Text Request