aiqbench analysis
This research introduces a curriculum-guided framework that enhances reinforcement learning policies for UAVs by aligning temporal-difference error distributions to generalize against unseen GNSS spoofing attacks. By deriving a theoretical certificate for performance degradation bounds, the approach achieves near-perfect mission success rates in fixed spoofing scenarios and significantly improves efficiency under dynamic adversarial conditions. The work addresses a critical gap in robust RL, moving beyond model-specific defenses to ensure transferability across diverse attack vectors.
Curriculum-Adapted Robust Reinforcement Learning for UAV Deconfliction in Adversarial Environments
arXiv:2506.21129v2 Announce Type: replace-cross Abstract: Autonomous unmanned aerial vehicles (UAVs) increasingly rely on reinforcement learning (RL) for navigation. However, global navigation satellite system (GNSS) spoofing attacks can induce out-of-distribution observation shifts that corrupt value estimation and degrade mission performance. Existing robust RL approaches typically improve resilience against specific attack models but often fail to generalize to attacks not encountered during training. To address this limitation, we propose a curriculum-guided adaptation framework that progressively exposes a robust policy to gradient-based adversarial observation perturbations of increasing intensity while aligning temporal-difference (TD) error distributions across curriculum stages. Rather than adapting to a particular attack model, the proposed approach preserves TD-error consistency to promote transferability across attack conditions. We further derive a TD-space generalization certificate showing that if the TD-error distribution induced by a test-time attack remains sufficiently close to that of the final curriculum stage, the resulting performance degradation is bounded. The framework is evaluated in a UAV deconfliction environment with dynamic 3D obstacles under previously unseen fixed and dynamic GNSS spoofing attacks. Under fixed spoofing conditions, the curriculum-adapted policy achieved near-perfect mission success rates, compared with 20-56% for standard and robust RL baselines. Under dynamic obstacle-luring spoofing attacks, it achieved the highest episodic rewards while reducing mission completion steps by up to 45% across increasing aerial traffic densities.
Abstract:Autonomous unmanned aerial vehicles (UAVs) increasingly rely on reinforcement learning (RL) for navigation. However, global navigation satellite system (GNSS) spoofing attacks can induce out-of-distribution observation shifts that corrupt value estimation and degrade mission performance. Existing robust RL approaches typically improve resilience against specific attack models but often fail to generalize to attacks not encountered during training. To address this limitation, we propose a curriculum-guided adaptation framework that progressively exposes a robust policy to gradient-based adversarial observation perturbations of increasing intensity while aligning temporal-difference (TD) error distributions across curriculum stages. Rather than adapting to a particular attack model, the proposed approach preserves TD-error consistency to promote transferability across attack conditions. We further derive a TD-space generalization certificate showing that if the TD-error distribution induced by a test-time attack remains sufficiently close to that of the final curriculum stage, the resulting performance degradation is bounded. The framework is evaluated in a UAV deconfliction environment with dynamic 3D obstacles under previously unseen fixed and dynamic GNSS spoofing attacks. Under fixed spoofing conditions, the curriculum-adapted policy achieved near-perfect mission success rates, compared with 20-56% for standard and robust RL baselines. Under dynamic obstacle-luring spoofing attacks, it achieved the highest episodic rewards while reducing mission completion steps by up to 45% across increasing aerial traffic densities.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2506.21129 [cs.LG]
(or arXiv:2506.21129v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2506.21129 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 306 words
Related coverage
Curriculum-Adapted Robust Reinforcement Learning for UAV Deconfliction in Adversarial Environments
arXiv · Oct 15, 2026
Reflection’s Beam model hits 501 billion parameters
Quantum Zeitgeist · Oct 5, 2026
Cooperative Multi-UAV Navigation in Complex Environments via Systematic Multi-Agent Deep Reinforcement Learning
arXiv · Jul 29, 2026
Towards More Efficient, Robust, Instance-adaptive, and Generalizable Sequential Decision making
arXiv · Oct 1, 2026