中文摘要：

消息认证是合法发送方Alice传输消息M给合法的接收方Bob并向Bob认证M的交互过程。为了防止敌手Eve的攻击,Alice和Bob通常共享了一个安全密钥。该文考察如下认证框架：Alice首先通过无噪声信道将消息M发送给Bob;Alice接着利用消息M和安全密钥K生成一个认证标签;Alice再将认证标签转化为码字Xn;最后,Alice通过窃听信道模型将码字Xn传输给Bob。该文定义了固定标签率下的安全认证信道容量,并证明该认证信道容量等于H（X｜Z）。特别地,证明了文献[15]提出的协议在该文的认证模型中是可达容量的。

英文摘要：

Message authentication is an interactive procedure that allows a legitimate sender Alice to send and authenticate a message M to a legitimate receiver Bob. To prevent the attacks from an adversary Eve, Alice and Bob usually share a secret key K. In this paper, we study a novel authentication framework as follows. Firstly, Alice sends a message M to Bob over a noiseless channel; Secondly, Alice generates an authentication tag with the message M and secret key K; Thirdly, Alice encodes the tag into a codeword Xn; Finally, Alice transmits the codeword Xn to Bob over a wiretap channel. This paper defines an authentication channel capacity under a fixed tag rate, and show that it equals to H（X｜Z）. Specifically, we prove that the authentication protocol proposed in Ref. [15] is capacity-achievable under our authentication model.