Towards robustness and energy efficiency of cut detection in wireless sensor networks

Abstract

Reliable, full network connectivity in wireless sensor networks (WSN) is difficult to maintain. Awareness of the state of network connectivity is similarly challenging. Harsh, unattended, low-security environments and resource-constrained nodes exacerbate these problems. An ability to detect connectivity disruptions, also known as cut detection, allows WSN to conserve power and memory while reducing network congestion. We propose ER-CD and LR-CD, protocols that detect cuts while providing energy-efficiency and robustness to attack. Using distributed, cluster-based algorithms, ER-CD recognizes and determines the scope of disrupted connectivity while examining available data for evidence of an attack. For more resource-constrained networks, LR-CD enhances security through the use of a robust outlier detection algorithm. Extensive simulations and a hardware implementation provide experimental validation across a range of network sizes and densities. Results indicate that energy-efficiency can be improved by an order of magnitude in denser networks while malicious nodes are detected at deviations of 1% from expected behavior.