THALIS: Trust-Based Heterogeneous Autonomous Localization and Information Spreading

This paper presents THALIS, a novel trust-based framework for autonomous localization and information dissemination in GPS-denied environments. The proposed THALIS solution addresses the critical challenges of reliable peer-targets selection, energy-efficient collaboration, and information exchange by introducing a reinforcement learning-based Confident Adaptive Q-Learning algorithm, which dynamically optimizes the peer-targets selection by balancing the signal strength, trust metrics, and energy constraints. Additionally, a multi-component trust model is introduced by combining the direct experiential assessments and the network-wide reputational data to evaluate the peer-targets reliability. Also, a multilateral bargaining protocol is designed to ensure the fair information spread among the peer-targets during the mission-critical operation. The THALIS solution is validated through in-field experiments performed at the U.S. Army Combat Capabilities Development Command Army Research Laboratory demonstrating the superior performance of the proposed solution both in terms of localization and energy efficiency.