Foundational Practices for AI for Autonomy

When designing autonomous systems, specifically uncrewed vehicle platforms (UxVs), a system-level approach is essential for integrating AI components. These components offer new opportunities that include sensor fusion, enhanced perception, rapid adaptive control, and real-time mission planning. They also present complexities in several areas: explainability, traceability, compute requirements, and operational support once deployed. A strategic approach based on proven practices is crucial for managing these complexities and ensuring optimal outcomes throughout the system’s life.