Comparing Reputation Schemes for Detecting Malicious Nodes in Sensor Networks

The Computer Journal (2011) 54 (3): 482-489

Comparing Reputation Schemes for Detecting Malicious Nodes in Sensor Networks

Partha Mukherjee and Sandip Sen*

sandip@utulsa.edu

Department of Computer Science, University of Tulsa, Tulsa, OK 74104, USA

Abstract

Remotely deployed sensor networks are vulnerable to both physical and electronic security breaches. When compromised, sensor nodes can send erroneous data to the base station, possibly compromising network effectiveness.

Authors assume that sensor nodes are organized in a hierarchy and use an offline neural network-based learning technique to predict the data sensed at any node given the data reported by its siblings in the hierarchy. This allows detection of malicious nodes even when siblings do not sense data from the same distribution.  However, speed of detection of compromised nodes depends on the mechanism used to update the reputation of the sensor nodes over time. Authors compare and contrast the relative strengths of a statistically grounded scheme and a reinforcement learning-based scheme both for their robustness to noise and responsiveness to change in sensor behavior. They first extend an existing mechanism to improve detection capability for smaller errors. They then analyze the influence of different discount factors, including unweighted, exponential and linear discounts, on the tradeoff between responsiveness and robustness. They both develop a theoretical analysis to understand the tradeoff and perform experimental verification of predicted results by varying the patterns in sensed data.

Full Text Source (Subscription or Fee): http://comjnl.oxfordjournals.org/content/54/3/482.short