Research On Key Technologies Of Physical Layer Security In Wireless Communications

With the rapid development of wireless communication technology,the problem of communication security is becoming more and more serious.Wireless physical layer security technology uses the intrinsic attributes of wireless signals to establish a security mechanism,which provides a new idea for solving the communication security problem,and can be used as a supplement to the traditional security systems to enhance the security of wireless communication systems.Wireless channel key generation technology and radio frequency fingerprint identification technology are hot issues in the field of wireless physical layer security research.The reciprocity,randomness and uniqueness of wireless channels provide shared random sources for key generation.Radio frequency fingerprint(RFF)is an inherent feature of transmitters and can be used to distinguish different wireless devices.However,in practical systems,the wireless channel and RFF are usually overlapped,which will interfere with key generation and RFF identification.Therefore,1)the separation of wireless channel and RFF,that is,the estimation of RFF,2)the identification of RFF,3)the high efficient wireless channel key generation,are the key issues to be solved in the research of wireless physical layer security.Aiming at the field of wireless communication physical layer security,this paper studies the RFF estimation and identification method in multi-antenna orthogonal frequency division multiplexing(OFDM)systems and the efficient key generation method for multi-input multi-output OFDM(MIMO-OFDM)systems,the specific research contents and innovative research results are as follows:1.Proposed a transmitter RFF estimation algorithm for uniform linear array(ULA)in OFDM systems.In the wireless signal received by the detection device,the transmitter RFF to be identified is superimposed with the wireless channel response.Wireless channel response will interfere with RFF estimation.The existing RFF recognition method has not considered the separation of RFF and wireless channel.We propose an ESPRIT based transmitter RFF estimation(ERFFE)algorithm,which uses the subcarrier of OFDM systems as a probe to obtain the in-band frequency response of the transmitter,and the RFF of the transmitter can be obtained by using the channel estimation of a single OFDM symbol.The advantages of ERFFE algorithm are: 1)does not depend on channel reciprocity;2)channel state information(CSI)does not need to be fed back through wireless channels;3)does not need to deal with multiple symbols cooperatively.Simulation results show that the RFF estimation performance of ERFFE algorithm is significantly better than that of CRbased method.Under the same conditions,compared with CR-based method,the RFF estimation accuracy of ERFFE algorithm is improved by more than one order of magnitude.2.Proposed an energy selection based transmitter RFF estimation algorithm.In view of the limited number of receiver antennas,an energy selection based ERFFE(ES-ERFFE)algorithm is proposed by analyzing the influence of multipath energy distribution on RFF estimation.The ES-ERFFE algorithm determines the number of paths to be estimated according to the energy threshold,which can effectively reduce the number of antennas required for RFF estimation,and can achieve better estimation performance than ERFFE algorithm when the number of antennas in the receiver remains unchanged.Simulation results show that the RFF estimation performance of ES-ERFFE algorithm is better than that of ERFFE algorithm,and the ES-ERFFE algorithm can save 1/3 of the number of antennas under the simulation scenario in this paper.3.Proposed a time diversity based transmitter RFF estimation algorithm.In view of the low successful identification rate of RFF in the case of low signal-to-noise ratio,a time diversity based ERFFE(TD-ERFFE)algorithm is proposed.In the TD-ERFFE algorithm,each independent RFF estimation is obtained by sending multiple sets of pilot sequences,and three RFF joint identification methods are used to obtain the time diversity gain of RFF identification.Simulation results show that the TD-ERFFE algorithm can significantly improve the successful identification rate of RFF in the low SNR region.For example,when SNR?13d B,the RFF successful identification rate of TD-ERFFE algorithm can be improved by about 20% under the simulation scenario in this paper.In addition,combining the TDERFFE algorithm with the ES-ERFFE algorithm,the TD-ES-ERFFE algorithm can further improve the RFF identification performance.4.Proposed a transmitter RFF estimation algorithm for uniform rectangular array(URA).In large-scale MIMO systems,receiver antenna arrays are usually designed as two-dimensional arrays.In order to get closer to the actual system,we propose a RFF estimation algorithm,RD-ERFFE algorithm,which is suitable for URA under the 3D-MIMO channel model.In the RD-ERFFE algorithm,URA is divided into several sub-arrays according to rows and columns,independent RFF estimation is obtained from each subarray by dimensionality reduction processing of the arrival angle(AOA),and then the sub-array gain of successful identification rate is obtained through RFF joint identification.Aiming at the problem of angle fuzzy caused by AOA reduction,two angle fuzzy filtering methods are given to reduce the influence on RFF identification.Simulation results show that the successful identification rate can be improved significantly by RFF joint identification and angle fuzzy filtering,and the RFF identification performance of RD-ERFFE algorithm is better than that of CR-based method.Under the same number of antennas,The RD-ERFFE algorithm in URA has better RFF identification performance than the ERFFE algorithm in ULA.5.Proposed a new wireless channel key generation method for MIMO-OFDM systems.In the quantitative preprocessing stage,aiming at the problem of information redundancy in frequency domain impulse response(CFR),a coarse data extraction method based on correlation time is designed.In order to improve the consistency of CFR,a fine data extraction method based on limited feedback of packet mean is designed.In the channel quantization stage,a direct quantization method and its auxiliary method are proposed,which can effectively improve the efficiency of key generation.Finally,simulation analysis is carried out in LTE-A system to verify the effectiveness of the proposed key generation method.