【摘要】 无条件安全密钥协商一般包括初始化、通信和决策三个阶段．该文基于纠错码理论提出了一种认证方案．该方案利用通信双方在初始阶段所获得的相关信息对通信阶段的通信内容进行认证．如果初始阶段中通信双方及敌手所获得的初始信息是由一个二元对称信源通过二元对称信道广播所得到的，该研究结果认为在敌手的信道比通信双方的信道都差的条件下，总能够找到一种（Ｎ，Ｋ，ｄ）线性码来实现我们所提出的认证方案，并使得收方接受合法消息的概率至少为１－ １／ｓ２ ，且拒绝非法消息的概率至少为１－ １／ｓ′２ ，其中，ｓ和ｓ′均为大于１的安全参数． 同时分析了如何选择编码参数以达到认证所要求的安全性能，并指出了所需的认证符长度的下界与初始化阶段信道的误比特率、认证方案的传信率及安全参数间的关系．更多还原

【Abstract】 An unconditionally secure secret key agreement protocol usually consists of three phases, Initialization phase, communication phase and decision phase. For a scenario where two communicants and an adversary receive the output of a binary symmetric source over independent binary symmetric channels as their initial information, an authentication scheme is proposed based on coding theory, which uses the correlated strings between the two communicants to authenticate the messages over the public channel. This paper illustrates with an example that when the adversary’s channel is noisier than the communicants’ channels during the initialization phase, a ( N,K,d ) linear code can always be found to implement the authentication scheme, ensuring that the receiver accepts the sender’s legitimate messages with probability at least 1-1/s 2 while rejects Eve’s fraudulent messages with probability at least 1- 1/s ′ 2, where s and s′ are secure parameters. At last, this paper shows that the lower bound of the required authenticator length is closely related to the secure parameters s,s′ , code rate of the authentication scheme K/(K+N) and bit error probabilities of independent noise channels in the initialization phase.更多还原