A Verification Framework for Certifying Learning-Based Safety-Critical Aviation Systems

TL;DR

This paper introduces a comprehensive verification framework combining offline mixed-fidelity verification and online monitoring to ensure the safety of learning-based components in aviation systems throughout their development and deployment.

Contribution

It presents a novel, modular framework integrating design-time and run-time assurance methods for safety-critical learning-based aviation systems.

Findings

Effective offline verification using mixed-fidelity models

Online reachability and statistics-based safety monitoring

Framework supports continuous learning and safety assessment

Abstract

We present a safety verification framework for design-time and run-time assurance of learning-based components in aviation systems. Our proposed framework integrates two novel methodologies. From the design-time assurance perspective, we propose offline mixed-fidelity verification tools that incorporate knowledge from different levels of granularity in simulated environments. From the run-time assurance perspective, we propose reachability- and statistics-based online monitoring and safety guards for a learning-based decision-making model to complement the offline verification methods. This framework is designed to be loosely coupled among modules, allowing the individual modules to be developed using independent methodologies and techniques, under varying circumstances and with different tool access. The proposed framework offers feasible solutions for meeting system safety requirements at different stages throughout the system development and deployment cycle, enabling the continuous learning and assessment of the system product.