Abstract
Cyber-Physical Systems (CPS) must be robust against several sources of uncertainty. For high-assurance applications, robustness against uncertainty drives development and ownership costs, as well as size, weight and power. Hazard analysis and requirement generation are critical early-design steps that can help identify cost-effective solutions. As CPS evolve towards more autonomous systems, new sources of uncertainty arise which require complementing current methods with new ones. We present a modeling formalism to support the requirement generation process for autonomous CPS.