Enhancing Trust in Human-AI Interaction Through Explainable Decision Support Systems for Mission Planning of UAS Swarms

Kurzfassung

Artificial Intelligence systems often operate as “black boxes”, creating challenges for trust and collaboration in human-AI teaming environments. Particularly for Unmanned Aerial Systems swarm operations. This paper proposes an explainable decision support framework that integrates fuzzy-logic with reinforcement learning to enhance transparency, while maintaining adaptability. The framework processes uncertain inputs through linguistic variables and interpretable rules, generating natural language explanations alongside mission recommendations. Reinforcement learning optimizes system parameters within constraints, ensuring the decision-making process remains transparent while performance improves over time. This approach addresses key challenges in unmanned aircraft systems swarm coordination, particularly for dynamic task allocation when assets fail or environmental conditions change. By preserving explainability throughout the optimization process, the system enables operators to understand not only what decisions are made but why they are made, which is a crucial factor for establishing trust in human-autonomous teaming.