Cyber Physical Systems couple computational and physical elements, therefore the behavior of geometry (deformations, kinematics), physics and controls needs to be certified using many different tools over a very high dimensional space. Because of the near infinite number of ways such a system can fail meeting its requirements, we developed a Probabilistic Certificate of Correctness (PCC) metric which quantifies the probability of satisfying requirements with consistent statistical confidence.
PCC can be implemented as a scalable engineering practice for certifying complex system behavior at every milestone in the product lifecycle. This is achieved by: creating virtual prototypes at different levels of model abstraction and fidelity; capturing and integrating these models into a simulation process flow; verifying requirements in parallel by deploying virtual prototypes across large organizations; reducing certification time proportional to additional computational resources and trading off sizing, modeling accuracy, technology and manufacturing tolerances against requirements and cost.
This process is an improvement over the V-cycle because verification and validation happens at every stage of the system engineering process thus reducing rework in the more expensive implementation and physical certification phase. The PCC process is illustrated using the example of “Safe Range” certification for an UAV with active flutter control.