基于学习机制的自组织认知无线网络安全协议研究

Many classic issues in Cognitive radio (CR) networks (CRNs), such as spectrum sensing, spectrum sharing have gained extensively studies, however, network security as a foundation for robust and reliable networking, its research is still in a preliminary phase. Especially, for Cognitive Radio Ad Hoc Networks (CRAHNs) with features of scalability, adaptive capability and dynamic spectrum access has received fairly well attention in real applications. How to design a complete and thoroughly considered secure protocol for CRAHNs is of significant research importance. Specifically, in CRAHNs, the distributed spectrum sensing and decision, the frequently changing spectrum resources, multi-hop architecture and the dynamic nature of network topology and the requirement of control channel to negotiate communications, give plenty of chances for malicious and adversarial attackers. Its network security issues are increasingly becoming evident, and many of which are very challenging. In this project, we discuss potential security attacks and their models targeting from the physical layer and moving up to the network layer, will launch appropriate defensive countermeasures and safety solutions, and the research of security attack models. In the current increasingly complex wireless security environments, we consider CRAHNs equipped with learning mechanism, could adaptively countermeasure this sort of security attacks in the long run, and be resistant and robust to more intelligent adversaries. Preliminary result has shown to achieve reliable and secure CRAHNs.

认知无线网络的经典研究议题，如频谱感知、频谱共享等已得到充分研究。但网络安全作为实现健壮与可靠网络的基础，其研究仍处于初级阶段。特别地，以具备可扩展性、自适应能力和动态频谱接入为特点的自组织认知无线网络，已受到实际应用的关注；如何为其设计一套完善的安全协议，研究意义重大。具体讲，该网络的分布式频谱感知与决策、经常变化的频谱资源、多跳架构与网络拓扑结构的动态性、需要控制信道协商通信等一系列新特点，给恶意与敌对攻击者带来可乘之机。其网络安全问题日益突出，并富有挑战性。本项目描述了该网络典型的安全攻击问题及其特点；将开展抵御安全攻击的算法与协议设计的研究，以及安全攻击模型的研究。在当前复杂的无线安全环境下，考虑网络具备学习机制，能够长期自适应地抵御这些典型安全攻击，并对各种智能的对抗攻击者表现出鲁棒性。已展示一些初步结果，实现健壮且安全的自组织认知无线网络。

认知无线网络的经典研究议题，如频谱感知、频谱共享等已得到充分研究。但网络安全作为实现健壮与可靠网络的基础，其研究仍处于初级阶段。特别地，以具备可扩展性、自适应能力和动态频谱接入为特点的自组织认知无线网络，已受到实际应用的关注；如何为其设计一套完善的安全协议，研究意义重大。具体讲，该网络的分布式频谱感知与决策、经常变化的频谱资源、多跳架构与网络拓扑结构的动态性、需要控制信道协商通信等一系列新特点，给恶意与敌对攻击者带来可乘之机。其网络安全问题日益突出，并富有挑战性。本研究以分层形式，从物理层到网络层，描述了该网络典型的安全攻击问题及其特点。在当前复杂的无线安全环境下，考虑网络具备学习机制，能够长期自适应地抵御这些典型安全攻击，并对各种智能的对抗攻击者表现出鲁棒性。本项目已开展抵御安全攻击的算法与协议设计的研究，以及安全攻击模型的研究。已展示许多丰富的结果，实现了健壮且安全的自组织认知无线网。