Runtime Safety Filtering for Learned Small UAS Separation Policies under GNSS Degradation

2026-07-10

AI Summary

This study addresses the safety risks posed by degraded GNSS performance in learning-enabled small unmanned aircraft systems (sUAS), where inaccurate position and velocity estimates can compromise separation assurance. For the first time, it systematically compares two runtime safety architectures: action filtering based on worst-case traffic state estimation and observation correction that preserves the autonomy of policy decisions. Leveraging worst-case state estimation under bounded observation uncertainty, discrete-time control barrier functions, and adversarial GNSS degradation scenarios, experimental results demonstrate that observation filtering—by retaining the policy’s decision authority—significantly outperforms handcrafted action constraints, reducing near-midair encounters by 90% while maintaining a robust trade-off between separation distance and closure rate. In contrast, action filtering yields negligible safety improvements.

Abstract

Learning-based separation assurance for small Unmanned Aircraft Systems (sUAS) achieves near-zero collision rates in simulation, but assumes accurate position and velocity information from Global Navigation Satellite Systems (GNSS). This assumption fails in urban environments, where multipath propagation, signal blockage, and intentional interference degrade navigation integrity. This raises a fundamental architectural question for deploying learned separation policies under GNSS degradation: should runtime safety mechanisms filter the policy's actions or its observations? This work evaluates both approaches for multi-agent sUAS separation under adversarial GNSS degradation. Both architectures first estimate a worst-case traffic state consistent with bounded observation uncertainty, then diverge: action filtering constrains policy outputs via discrete-time control barrier functions evaluated at the worst-case state, while observation filtering presents the worst-case state directly to the policy as corrected input. Experimental results show that action filtering provides negligible safety improvement, while observation filtering reduces near mid-air collisions by 90% and remains robust to the barrier function's tradeoff between separation distance and closing rate. These results suggest that, for policies with learned safety behaviors, preserving the policy's decision authority outperforms overriding its actions with hand-designed constraints.

Problem

• GNSS degradation
• learned separation policies
• runtime safety
• small UAS
• navigation integrity

Innovation

• observation filtering
• action filtering
• GNSS degradation
• learned separation policies
• control barrier functions