Satisfying integrity requirements for highly automated UAV systems by a systems engineering approach to Cognitive Automation

The approach of cognitive automation [1] has proven its effectiveness in the implementation of complex behaviors in the field of highly automated UAV guidance applications. The need for higher levels of automation results from the anticipated dynamics and complexity of situations UAVs will have to deal with, possibly without any chance for intervention by the supervising human agency. In the first part of this paper the concept of cognitive automation and its implementation as Artificial Cognitive Units (ACU) will be presented. The paper will give some insight in the key performance features of cognitive automation and its proof of concept in a multi UAV guidance application. Next, the conventional systems engineering approach for highly integrated and complex aircraft systems will be discussed. The main focus here will be on the aspect of requirements engineering. During the process of deriving requirements on various levels, a comprehensive description of the purpose and function of a system is generated. The intention of this paper is to discuss an approach on how this knowledge can be captured and implemented into an ACU for the sake of high system integrity. Finally, an experimental approach will be presented which shall demonstrate how to develop and integrate conventional automation and cognitive automation in order to demonstrate the interdependencies of both solutions. It shall also be demonstrated how a combination of both approaches can satisfy the integrity requirements for highly automated UAV systems following the principles of systems engineering.