A Graphical Bayesian Game for Secure Sensor Activation in Internet of Battlefield Things

TL;DR

This paper models secure sensor activation in IoBT systems as a graphical Bayesian game, proposing a best response algorithm to optimize secrecy and energy efficiency, with simulations showing significant energy savings and improved security.

Contribution

It introduces a novel graphical Bayesian game framework for secure sensor activation in IoBT, incorporating neighbor effects and providing an efficient equilibrium-finding algorithm.

Findings

Up to 98% energy reduction compared to baseline

Effective tradeoff between information transmission and secrecy

Proven convergence of the proposed algorithm

Abstract

In this paper, the problem of secure sensor activation is studied for an Internet of Battlefield Things (IoBT) system in which an attacker compromises a set of the IoBT sensors for the purpose of eavesdropping and acquiring information about the battlefield. In the considered model, each IoBT sensor seeks to decide whether to transmit or not based on its utility. The utility of each sensor is expressed in terms of the redundancy of the data transmitted, the secrecy capacity and the energy consumed. Due to the limited communication range of the IoBT sensors and their uncertainty about the location of the eavesdroppers, the problem is formulated as a graphical Bayesian game in which the IoBT sensors are the players. Then, the utilities of the IoBT sensors are modified to take into account the effect of activating each sensor on the utilities of its neighbors, in order to achieve better system performance. The modified game is shown to be a Bayesian potential game, and a best response algorithm that is guaranteed to find a Nash equilibrium of the game is proposed. Simulation results show the tradeoff between the information transmitted by the IoBT sensors and the desired secrecy level. Simulation results also demonstrate the effectiveness of the proposed approach in reducing the energy consumed compared to the baseline in which all the IoBT sensors are activated. The reduction in energy consumption reaches up to 98% compared to the baseline, when the number of sensors is 5000.