MERCURY: Multilateral Bargaining on Reconfigurable Intelligent Surfaces for Alternative Positioning, Navigation, and Timing

Alternative Positioning, Navigation, and Timing (PNT) solutions are important for ensuring robust and reliable communication, navigation, and timing services, especially in environments where traditional Global Navigation Satellite Systems (GNSS) may be unavailable or compromised. In this paper, we introduce the MERCURY mechanism, leveraging reinforcement learning and multilateral bargaining game theory, alongside the Reconfigurable Intelligent Surface (RIS) technology, to optimize RIS selection, allocation of RIS elements, and phase shifts to develop a novel alternative PNT solution. These optimizations aim to minimize the PNT error and accurately determine the targets, positions. Our contributions include designing an Optimistic Q-learning with Upper Confidence Bound (OQ-UCB) algorithm for autonomous RIS selection, proposing a game-theoretic model for distributed RIS elements allocation, and developing a phase shift optimization model. Through simulation-based experiments, we demonstrate the real-time applicability and superiority of the MERCURY mechanism in improving PNT system performance and mitigating targets’ PNT error.