Detection of Misbehaving Nodes in Malicious Organization Dominant Networks in Internet of Things

Abstract

Local advantage is easy to established for malicious organization in sense networks in Internet of Things by deploying mass sensoring nodes with low cost such that attacks performed by its nodes are covered by other conspirators' false reports. In this paper, we propose two game approaches which achieve Bayes equilibrium in both normal networks and in the malicious organization dominant(MOD) network. We focus on the relationship between nodes and their organizations, trust of unknown nodes can be inferred from their organization reputation and previous interactions by node cooperation, which is reinforced by deduction feedback. Our simulations show that the improved cooperative game approach efficiently improves successful deduction rate and decreases forged reports significantly in all scenarios, the deduction feedback mechanism overcomes knowledge deviation in the cooperative game caused by observation limitations and reduces malicious events further.