中文摘要：

基于属性的密码体制是基于身份密码体制的泛化和发展,它将身份扩展为一系列属性的集合,具有更强的表达性,并且拥有相同属性的成员自动组成一个环,便于隐匿签名者身份。通过对现有的基于属性门限环签名方案的深入分析,发现这些方案虽然满足匿名性要求,但拥有互补属性的恶意用户可以通过合谋伪造出有效签名。为弥补上述缺陷,首先给出基于属性门限环签名的不可伪造性、不可区分性及抗合谋攻击性的形式化定义,然后给出一个基于属性的抗合谋攻击可变门限环签名方案,其安全性可归约为CDH（computational Diffie-Hellman）困难问题。所提方案通过在用户属性密钥中引入互不相同的秘密随机因子的方法,防止合谋攻击者利用组合私钥的方式伪造签名。在随机预言机模型下,方案被证明能够抵抗适应性选择消息的存在性伪造及合谋攻击,并具有相同签名属性集用户间的不可区分性。与同类方案相比,新方案还具备更高的运算效率。

英文摘要：

Attribute-based cryptography is a generalization and development of the identity-based cryptography, which extends the identity to a set of attributes. In the attribute-based cryptosystem, different members with the same attributes can form a signature ring automatically, with the actual signer＇s personl idnetity easily hided. By in-depth analysis of several attribute-based threshold ring signature schemes proposed by the earlier reseachers, it is concluded that the malicious users with the complementary attributes can conspire to forge a valid signature. In order to compensate for the defect, the proposal first gives the formal definitions of the security characteristics in attribute-based threshold ring signature scheme, such as unforgeability, indistinguishability and anti-collusion attack. Then a new attribute-based alterable threshold ring signature scheme is presented, whose security is proven equivalent to the computational Diffie-Hellman problem. By introducing an random secret parameter in each user＇s private key, the proposal can resist the collusion attacks. It is proven that the proposal can satisfy existential unforgeability and anti-collusion against the adaptive chosen message attack in therandom oracle model, and can meet the requirement of indistinguishability among the users with the same attributes. Besides, property analysis shows that the proposal has high efficiency.