Social learning against data falsification in sensor networks

Sensor networks generate large amounts of geographically-distributed data. The conventional approach to exploit this data is to first gather it in a special node that then performs processing and inference. However, what happens if this node is destroyed, or even worst, if it is hijacked? To explore this problem, in this work we consider a smart attacker who can take control of critical nodes within the network and use them to inject false information. In order to face this critical security thread, we propose a novel scheme that enables data aggregation and decision-making over networks based on social learning, where the sensor nodes act resembling how agents make decisions in social networks. Our results suggest that social learning enables high network resilience, even when a significant portion of the nodes have been compromised by the attacker