Novel Information-theoretic Game-theoretical Insights to Broadcasting in Internet-of-UAVs

TL;DR

This paper introduces an information-theoretic and game-theoretic framework for secure and efficient broadcasting in Internet-of-UAVs, utilizing novel algorithms and dissipativity theory to address privacy and multi-user challenges.

Contribution

It presents a federated-learning Blahut-Arimoto algorithm and a dissipativity-based mean-field game approach for secure UAV broadcasting, addressing multi-user and privacy issues.

Findings

Proposed a federated-learning Blahut-Arimoto algorithm for secure communication.

Developed a dissipativity-based mean-field game model for UAV broadcasting.

Analyzed multi-user scenarios with a nested game approach.

Abstract

For the Internet-of-unmanned aerial vehicles (UAVs) some challenges in broadcasting and from new points of view are explored. In this paper, first, we investigate a single broadcast transceiver. From a control of noisy-channel viewpoint, we consider: (\textit{i}) Alice sends $\mathcal{X}$ to Bob as more \textcolor{black}{efficient} as possible while she wishes Bob not to get access to the private message $\mathcal{S}$ regarding the correlation between $\mathcal{S}$ and $\mathcal{X}$ $-$ i.e., Alice purposefully sends a \textit{turbulent-flow} of the information to Bob; and (\textit{ii}) where $\big (\Theta_1;\Theta_2 \big)$ is the control-action-pair which actualise a \textit{pursuit-Evasion}. We consider \textit{dissipativity} in our system due to the memory effect relating to the previous states. We thus propose a federated-learning based \textit{Blahut-Arimoto} algorithm while a 2-D \textit{dissipativity}-theoretic continuous-Mean-Field-Game (MFG) is proposed with regard to (w.r.t.) a joint probability-distribution-function (PDF) of the population distribution $-$ relating to a continuous-control-law. We also analyse what if Alice is owed to multiple Bobs in a multi-user scenario which we apply a bankruptcy based $3-$level nested game for.