摘要
由于无线传感器网络的工作环境存在不可靠性,其安全问题越来越被研究者重视。虽然数据加密和身份认证技术的应用提高了无线传感器网络的安全性,但仍然无法及时有效地检测出已对无线传感器网络造成威胁的恶意入侵行为。为了解决这一问题,文章针对Zigbee无线传感器网络的应用层的感知数据篡改、网络层的RREQ泛洪攻击、链路层碰撞攻击和物理层的信道干扰4种入侵行为,提出了一种基于相似度计算的无线传感器网络实时入侵检测算法。该算法通过马氏距离计算入侵行为和正常行为之间的相似程度,并采用线性拟合算法对可疑行为进行趋势评估,以检测异常入侵行为。最后通过搭建真实的Zigbee网络入侵检测环境验证了该算法的可用性,并能达到预期的效果。
Wireless sensor network（WSN） security issues are getting more attention by researchers due to unreliable and untrusted circumstances. Even encryption and authentication are applied into WSN, and they cannot prevent the intrusion of malicious network when they have attacked the WSN successfully. In this paper, we describe an online, sequential intrusion detection algorithm for the intrusion of data temper in application layer, RREQ flooding in network layer, collision in link layer and jamming in physical layer. The proposed algorithm is based on the similarity algorithm and linear fitting algorithm, and raises an alarm of intrusion immediately upon encountering a deviation or countertrend from the previous data. Based on the experiments results of intrusion detection system（IDS） using this algorithm, we demonstrate that our online algorithm is effective and achieving its desired results.
出处
《信息网络安全》
2016年第2期22-27,共6页
Netinfo Security
基金
国家自然科学基金青年科学基金[61101108]
关键词
无线传感器网络
相似度
马氏距离
线性拟合
wireless sensor network
similarity
Mahalanobis distance
linear fitting
