Research On Physical Layer Security Transmission Technology With Imperfect Channel Information

The performance of the physical layer security scheme depends on the accurate estimate ofthe channel state information (CSI). If the CSI on main channel or eavesdropper channel at thetransmitter is inaccurate or partially, the security of the scheme cannot meet the expectedrequirement. Therefore, many researchers have devoted to the effect of the imperfect CSI on thesecrecy rate. However, the following questions still exist: Firstly, spatially correlation among thechannels in multi-antenna system will attenuate the interference effect of the artificial noise.Secondly, study on the imperfect CSI is not entirely, for they do not consider the inaccurate CSIon both main channel and eavesdropper channel. Thirdly, the feedback of the entire CSI from thelegitimate user will cost much time in the limited feedback system, and the partially CSI willattenuate the secrecy rate.Considering the above questions, this paper has made a research on the physical layersecurity technology under imperfect channel information. Firstly, in the case of spatiallycorrelated channel and inaccurate CSI on both main channel and eavesdropper channel, wepropose two kinds of physical layer security scenario using artificial noise. Afterwards, in thepresence of limited feedback in no reciprocity channel communication situation, we propose aphysical layer security scenario using antenna selection. The detail of the work is as follows.1. A physical layer security scenario using artificial noise in correlation MISO channel isproposed. Because of the correlation between the main channel and the eavesdropper channelcaused by transmitted antenna correlation, the interference effect of the artificial noise isdecreasing. Correlated processing to the artificial noise will improve the similarity betweenartificial noise and eavesdropper channel. For this reason, obtain the correlated matrix from thestatistics of the main channel CSI, and then process the artificial noise with the correlated matrixto increase the similarity between artificial noise and eavesdropper channel. Simulation resultsshow that, the average secrecy rate of the scenario is0.5bit/s/Hz higher than the original.2. A physical layer security scenario using artificial noise with inaccurate CSI is proposed.When the CSI on main channel and eavesdropper channel is inaccurate, original algorithm aboutpower allocation and transmitted weighting coefficient would yield mismatch. Throughanalyzing the error caused by the mismatch, proposed an improved power allocation andtransmitted weighting coefficient algorithm. Amo ng the optimal algorithm, the mismatch istreated as the noise. Then, the power allocation and transmitted weighting coefficient aboutsignal and artificial noise on transmitter are design by the algorithm. Simulation results show that,the scenario is effective to the mismatch caused by time varied and insensitive to the channelnoise and the channel estimate mismatch. The average rate could improve1bit/s/Hz and2bit/s/Hz individual for mismatch and correlated cofference caused by time varied.3. A physical layer security scenario using antenna selection is proposed. By analyzing theequal power transmitting and beamforming individual, we obtain the distributed of the receiverSNR and the average secrecy rate under different number of selected antenna. Consideringlimited feedback, choose the better transmitted scenario and the best transmitted antenna association. The transmitter transmits the signal according to the feedback of the transmittedscenario and transmitted antenna association CSI. Simulation results show that, the averageamount of secure information gained by the scenario is better than single scenario.