A methodological framework to support the design of safe and secure autonomous systems

Advances in robotics, artificial intelligence, and communication technology are enabling increasingly autonomous systems. Different systems have different levels of autonomy and different levels of openness, which introduce different risks for safety and security in different industries. These risks are not static. Rather, risks change as organizations seek to bring together new devices and machines into use within wider systems. Currently, however, there are only general guidelines and limited domain-specific standards for designing safety and security into autonomous systems. Even in these, the focus is on providing general performance guidelines instead of prescriptive design requirements. Hence, there is a gap between the current standards base and technologies being developed. This gap leaves technology developers with an increasing responsibility for ensuring safety. Consequently, there is need for technology developers to be able to take a holistic view of safety and security issues throughout the systems engineering process. In the first part of this paper, we discuss the implications of different levels of autonomy and openness of the use environment, and review selected key safety issues in autonomous systems development. In the second part, we discuss the risk analysis methods for autonomous machines development, and propose a preliminary methodological framework that can support product development of safe and secure autonomous systems.