Evaluation of Verification Approaches Applied to a Nonlinear Control System

As the demand for increasingly complex and autonomous systems grows, designers may consider computational and artificial intelligence methods for more advanced, re- active control. While the performance gained by such increasingly intelligent systems may be superior to traditional control techniques, the lack of transparency in the systems and opportunity for emergent behavior limits their application in the field. New verification and validation methods must be developed to ensure the output of such controllers do not put the system or any people interacting with it in danger. This challenge was highlighted by the former Air Force Chief Scientist in his 2010 Technology Horizons Report, stating \It is possible to develop systems having high levels of autonomy, but it is the lack of suitable [verification and validation] (V&V) methods that prevents all but relatively low levels of autonomy from being certified for use