Trustworthy Autonomy through Ethics-Aware Integrity Monitoring

Abstract

The increasing reliance on autonomous systems in aerospace raises a dual challenge: ensuring that aircraft remain physically safe while also making decisions that are ethically responsible and auditable. Traditional Prognostics and Health Management (PHM) has matured in predicting and preventing technical failures, yet it operates independently from emerging mechanisms for ethical oversight. This separation leaves open questions about how to arbitrate when physical health indicators and ethical obligations point to conflicting actions, and how such decisions can be explained and trusted in highstakes environments.
This paper proposes a unified framework that integrates PHM with a runtime ethics module, treating both health and ethical risks as core aspects of system integrity. Central to the approach is an Event Manager that evaluates proposed actions against prognostic forecasts and codified ethical rules, applying a priority scheme that ranks safety of life above regulatory compliance, system preservation, and mission objectives. To ensure transparency and accountability, the system generates event-triggered explanations and records both outcomes and justifications in a tamper-evident blockchain ledger.
We demonstrate the framework through three case studies: an eVTOL facing an emergency landing where passenger survival must be balanced with bystander safety, a long-endurance UAV deciding whether to continue data collection or return before failure, and a UAV swarm navigating dynamic no-fly zones. Across Monte Carlo simulations, the integrated PHM–Ethics approach consistently reduced combined integrity losses, defined as missed failures plus ethical violations, when compared with PHM-only or Ethics-only baselines. Explanation payloads remained lightweight and blockchain logging introduced only modest latency, demonstrating feasibility for real-time aerospace operations. By showing that PHM and ethics can be brought into a single decision loop without sacrificing performance, this work provides a concrete pathway toward trustworthy autonomy in aerospace. The results highlight how transparency, auditable decision-making, and equitable risk management can be engineered into safety-critical systems, offering practical tools to support certification, regulatory trust, and public confidence in the next generation of autonomous flight.