在叢聚式無線感測網路下以統計分析為基礎之選舉式攻擊偵測系統

Abstract

在無線感測網路的環境之下，節點遭受到攻擊者入侵並且加以控制的問題，往往對於整個網路的正常運作造成很大的威脅。其中一項較有名的攻擊手法叫做偽造資料攻擊。攻擊者會在無法預期的狀況之下假造資料回傳給主伺服器，造成感測資料可信度失效。目前傳統的解決方案最有效的，只能在離攻擊者許多路由距離之外才能偵測出攻擊事件的發生。如此只會在攻擊者已經對網路造成傷害之後才開始解決問題。本論文所提出的方法，能夠在攻擊發生的同時侷限傷害範圍。使其能夠在距離攻擊來源一個路由距離內及時發現攻擊事件，並且在網路遭受到傷害之前加以阻絕。經過實驗分析以及結理論證明，我們發現此研究成果確實可以在這些威脅之下提供無線感測網路一個高度安全的可靠環境。

Node compromise poises security problems to wireless sensor networks (WSNs), such as attacks based on fabricated reports or false votes on real reports. Most conventional methods address these problems at a location several hops from the attacker, which results in high resource consumption and the spread of damage across the network. We propose a statistical voting scheme to solve this problem within a 1-hop cluster while minimizing the use of extra resources. Through statistical analysis, we compute reasonable data ranges of each sensor to pinpoint inside attackers. After analyzing the probable behavior of compromised nodes, our scheme can limit the damage caused by compromised nodes within a cluster. Through both analysis and simulation, we demonstrate that the statistical voting scheme can provide strong protection against these critical threats.