Research On Physical Layer Secret Key Generation For Internet Of Things

With the rapid development of Internet of Things(Io T),the security issues become more and more prominent.Traditional cryptographic protocols face key distribution and key management problems because of numerous nodes,and self-organized network.Physical layer secret key generation from the randomness of wireless communication channels is a promising technique to share cryptographic keys securely between legitimate users.It can supplement lightweight cryptographic protocols,and is an appealing solution for Io T.The existing works on physical layer security in Io T have carried on some preliminary exploration researches.However,combining the application characteristics of Io T,there are still unsolved problems,including: 1)the secret key rates are relative low in the quasi-static channels in Io T,where the channels change slowly;2)the legitimate two parties cannot generate the secret key directly in the multi-hop communication scenarios because there is no shared channel between them;3)the existing group key generation methods are based on pairwise key generation,which require many times information reconciliation,thus decreasing the group key generation efficiency.Aiming at the above problems,this thesis researches on the physical layer secret key generation for Io T,which is supported by National Natural Science Foundation of China “Research on physical layer security networking technology based on game theory” and “Lightweight physical layer secret key generation based on the random characteristics of the receiving signals”.Using the rich node resources,we first propose the relay-aided secret key generation methods with the quasi-static channels where a direct link is available between the legitimate two parties.Then the secret key generation methods based on secure secret transmission are present in the multi-hop communication scenarios.Finally,the group secret key generation approaches based on secure secret sharing are proposed in the group communication scenarios.The main research results are shown as follows:1.The secret key generation methods based on cooperative relays are proposed to improve the generated secret key rate for quasi-static channels in Internet of things,where the two legitimate users exploit the randomness of relay channels as additional random sources,thus improving the secret key rates.Firstly,a relay-aided method based on the secure network coding is proposed,which is based on Shannon's “one-time pad” perfect cipher.After that the relay employs secure network coding technique to assists the two legitimate nodes to obtain the relay channels information.Finally,the two legitimate nodes agree on a secret key from the direct and relay channels information over the direct channel.The achievable secret key rate is derived,which verifies that the method can improve the achievable secret key rate.Monte Carlo simulation results show that the achievable key rate of the proposed method dominates the amplify-and-forward method,and increases linearly with SNR,approaching the optimal rate.What's more,we present a multiple relay-aided secret key generation method to further improve the secret key rate,using the multiple intermediate nodes in Io T.And the optimal relays selection and power allocation model is established under the consideration of constant coherent time and power constraints,which is NP-hard to solve.Then we propose a suboptimal relays selection and power allocation algorithm.Monte Carlo simulation results show that increasing the relay nodes,selecting the relay with a larger variance channel can further improve the achievable secret key rate,and the achievable secret key rate of the proposed algorithm dominates that of the random relays selection and equal power allocation.2.The legitimate two parties cannot generate the secret key directly in the multi-hop communication scenarios because there is no shared channel between them.We first propose a secure secret transmission method,which is based on Shannon's “one-time pad” perfect cipher.Then a secret key generation method based on secure secret transmission is presented with one transmission path,where the relays employ secure secret transmission method to assist the two legitimate users to obtain the correlative observations of the secret channel,and the two legitimate users agree on a common secret key directly over the public channel.The achievable secret key rate is derived,which verifies that there is nothing leakage about the secret channel information to the eavesdropper.Monte Carlo simulation results show that the proposed method can improve the achievable secret key rate comparing with the amplify-and-forward method.In addition,taking the advantage of the multiple transmission paths in Io T,a multiple paths based secret key generation method is presented to further improve the secret key rate.The achievable secret key rate is derived,which verifies that the method can improve the achievable secret key rate.Monte Carlo simulation results show that increasing the wireless transmission paths,selecting the transmission path with less hops and larger variance channels can further improve the achievable secret key rate.3.The existing group key generation methods are based on pairwise key generation,which require many times information reconciliation,thus decreasing the group key generation efficiency.We present a secure secret sharing based group secret key generation method,where the group users employ secure secret sharing to assist all the users to obtain the correlative observations of the secret channel,thus only one time information reconciliation is needed.We first divide the general group netwoks to star and chain topology subnetworks.Then two secure secret sharing methods are presented for the two networks based on Shannon's “one-time pad” perfect cipher,respectively.In star topology network,the central node employs secure secret sharing method to assist all the group users to obtain the correlative observations of the secret channel.While in the chain topology network,the nodes in the middle of the chain employ secure secret sharing method to assist all the group users to obtain the correlative observations of the secret channel.Finally,all the group users agree on a common group key with one information reconciliation step.Monte Carlo simulation results show that the proposed group key generation methods in star and chain topology networks both can improve the achievable secret key rate comparing with the existing methods and the achievable group key rate in star topology network dominates that in chain topology with the same group users.Therefore,the number of users on the chain network should be decreased to improve the group secret key rate of the total network.