Optimal Placement of Things in an Adversarial Internet of Battlefield Things
Abstract
In this project, we have developed a suite of tools, based on game theory and learning, to address the grand challenge of providing security decision making and system optimization in a large-scale Internet of Battlefield Things (IoBT) network with heterogeneous devices. First, we have considered the connectivity problem in an IoBT that includes a set of heterogeneous devices that sense different types of information (e.g., different types of IoBT nodes with heterogeneous capabilities). The IoBT devices must transmit their information, through intermediary local sinks, to the general sink. We have studied a scenario related to an adversarial IoBT in which an attacker is interested in inducing disconnections to the network by choosing to compromise one of the IoBT nodes at each time epoch. Meanwhile, the IoBT operator acts as a defender that strives to maintain the connectivity of the IoBT network by either deploying new IoBT nodes or changing the roles of the nodes. We have formulated the problem using dynamic game theory and shown that sufficient conditions can be derived under which the IoBT system will remain connected at the gameÕs equilibrium. Simulation results have shown that the expected number of disconnected IoBT sensors, when the proposed dynamic game solution is used, decreases significantly compared to other, baseline game-theoretic and non-game-theoretic solutions. Second, we have developed a new approach for analyzing the security of a massive IoBT system that incorporates a large number of nodes, as is expected in future battlefields. Here, our key goal was to merge techniques from graph theory and game theory to analyze how an IoBT can stop the spread of malicious misinformation. To this end, we have developed a novel approach to control misinformation propagation in the IoBT. Using the proposed, distributed misinformation control approach, each IoBT node can decide whether or not to accept received battlefield information at each time instant, in presence of potential malicious misinformation that is being spread. Here, each IoBT node will seek to limit the propagation of misinformation. Due to the heterogeneity of the IoBT nodes in terms of connectivity, we have modeled the IoBT as a random graph in which the nodes have heterogeneous degrees that follow a predetermined distribution. We have formulated this IoBT misinformation propagation problem as a finite-state mean-field game with multiclass agents whose players are the IoBT nodes each of which is seeking to determine its probability of accepting the received information. Mean-field games were shown to be suitable for this IoBT problem since they handle an infinite number of players which is the case for a large-scale IoBT and capture the presence of several types of populations. Analytical and simulation results show that the proposed mean-field approach can enable an IoBT network to effectively control any potential spread of misinformation across its large-scale graph network. Third, we have analyzed the role of human players in IoBT security. In particular, we have studied a scenario in which a soldier seeks to accomplish a time-critical mission by traversing a battlefield within a certain amount of time, while maintaining its connectivity with an IoBT network. In this IoBT, an attacker seeks to find the optimal opportunity to compromise the IoBT network and maximize the delay of the soldierÕs IoBT transmission link. To study this problem, novel tools from psychological game theory are developed to capture the soldier and the attackerÕs psychological behavior pertaining to the soldierÕs and attackerÕs intentions to harm and frustrate one another are considered in their utilities. To solve this game, a novel learning algorithm based on Bayesian updating is proposed to find a psychological self-confirming equilibrium of the game. Simulation results showcase the effect of psychological factors on adversarial IoBT scenarios.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 19, 2019
- Source ID
- W911NF1710021
Entities
People
- Walid Saad
Organizations
- Army Contracting Command
- United States Army
- Virginia Tech