Studies Of Secret Key Extraction And Identity Authentication Based On The Characteristics Of Wireless Channels

With the rapid development of wireless communications,the security issue has become an increasingly concern.Recently exploiting the inherent properties of wireless channels to complement and enhance the traditional security mechanism is commonly observed,in which exploiting the properties of wireless channels to extract secret keys in the physical layer becomes a hot issue.The wireless channels are natural random sources based on the properties: short-time reciprocity,time-variation and location-specific.These properties ensure the common extracted secret keys with noise nature and further make it possible for one-time pad.The analysis of secret key capacity of different systems,the secret key generation in stationary environments,the secret key generation based on imperfect CSI and the application of secret key generation have been the key and difficult problems in the secret key generation study.This dissertation investigates these problems in depth and the main works are are as follows:1.The study of secret key capacity of orthogonal frequnecy division multiplexing(OFDM)systems.The secret key capacity which determins the greatest key generation rate,is a main evaluate index in secret key generation.Compared to single carrier system,OFDM systems can provide extra randomness in view of the use of multiple subchannels.How to make full use of the subchannels to increase the available keys is still an open issue.For this case,the dissertation firstly analyses the secret key capacity of OFDM systems under the assumption that the subchannels are independent and further derives the closed expression from the view of information theory.Then,how to increase the available secret key amount based on the channel state information(CSI)acquired by the legitimate communication parties in the channel probing phase,is investigated and further it is cast as an optimization problem of transmitting power allocation.Since the optimization problem is non-convex,we propose a power allocation algorithm based on geometric program(GP)and an underlying propagation protocol.The simulation results indicate that in comparison with the traditional equal power allocation algorithm,the GP-based power allocation algorithm obtains more secret key amount and achieves lower secret key disagreement rate(KDR).2.The study of secret key extraction in the static/quasi-static channels.Due to the slow variation of wireless channels in the static/quasi-static channels,which leads to the extracted bit sequences be unsuitable as secret keys because of their low entropy and an adversary can cause predictable secret key generation.To solve this issue,the dissertation proposes a random coefficient-moving window scheme to generate secret keys in static/quasi-static channels,which is applicable to both single-antenna and multi-antenna systems.In the scheme,the phase of the static/quasi-static channels are randomised by stochastic coefficients during channel probing and further a moving window is used to sum the channel estimates in the moving window at both parties.Thus,a new random source with great and random fluctuations is generated for extracting secret keys.The security of the scheme is evaluated by analysing the eavesdropper’s mean square error(MSE)for the worst scenario.Furthermore,in order to ensure the security of the secret keys in the MIMO systems,we propose an artificial noise-aided strategy to degrade the adversary’s MSE performance.Finally,we validate the feasibility of the proposed scheme in indoor and outdoor environmentes.The testing results show that our scheme achieves remarkable performance in KGR,and the generated keys pass the NIST test.3.The study of secret key extraction based on imperfect CSI.The imperfect CSI leads to high KDR and further brings great cost to achieve secret key agreement or eve a failure secret key extraction.For the imperfect channel reciprocity and noise,we propose a secret key generation scheme based on wavelet analysis.Firstly,the channel estimates are pre-processed by wavelet analysis to improve the correlation.Secondly,to ensure the randomness of the secret keys,an adaptive equal probability quantization approach is proposed to quantize the estimates.Furthermore,we validate the feasibility of the proposed scheme in real environments.Simulation and testing results all show that the proposed scheme can solve the secret key generation based on imperfect CSI and achieves notable performance improvement compared with existing schemes.For the channel estimation error,a secret key generation scheme based on the channel-phase is proposed.In the proposed scheme,firstly,the probe signal with random phase is adopted to probe the channel response and a node chosen stochastically acquires the preliminary secret key by quantizing the channel-phase response.Then,the preliminary secret key is mapped and pre-equalized,and sent to the other node,which achieves the secure distribution.Finally,the information reconciliation and privacy amplification are conducted to get the secure secret key.Only one legitimate node conducts the quantization,which reduces the channel estimation and quantization process,thus it improves the performance of secret key extraction.The simulation results indicate that compared with the existing channel-phase based and channel-amplitude based schemes,the secret key capacity and KGR of the proposed scheme are increased,the KDR is decreased.4.The study of identity authentication based on the channel phase in OFDM systems.The physical layer identity authentication,an application of secret key extraction,can complement or enhance the traditional authentication protocol.The performance of the physical layer authentication based on the channel amplitude may be degraded due to the channel amplitude variation.To solve this issue,the dissertation proposes a physical layer phase and pre-equalization challenge-response authentication scheme(PHY-PPCRAS).In the scheme,the legitimate parties share one string of secret key bit and complete the messages exchange through one OFDM symbol based on the channel reciprocity.The shared key is masked in the phase of the exchanged messages though pre-equalization.Finally,a binary hypothesis test is formulated for authentication and the analytical expressions of successful authentication rate and false acceptance rate are derived.The security analysis reveals that PHY-PPCRAS is immune to various attacks.The simulation results of receiver operating characteristic(ROC)indicates a high successful authentication rate is acquired even at low signal-to-noise region.Most previous physical layer authentication schemes suppose that the wireless channels are reciprocal during coherence time.However,due to halfduplex mode,noise and rapid variation of wireless channels,it is hard to ensure channel reciprocity in practical systems.For this problem,a physical layer phase challenge-response authentication scheme for nonreciprocal wireless channels(PHY-PCRAS-NWC)is proposed.In this scheme,the legitimate parties share two different strings of secret key bits.During authentication process,the secret keys are masked in the phases of the challenge and response messages,and two OFDM symbols are used to send them.Finally,a binary hypothesis test is formulated for authentication.The ROC performance reveals that PHY-PCRAS-NWC can achieve the identity authentication for nonreciprocal wireless channels.Besides,the scheme is immune to various attacks.