A Landscape for the Development of Dependable Machines

Abstract

This white paper, developed in the framework of the IEEE P2851 standardization project, describes a landscape of activities to be performed to address the dependability of intelligent machines during their development and usage in the field. In this context, the landscape is intended as the definition of the entirety of the activities that are executed within the autonomous machine dependability lifecycle. Dependability here is intended as the property of an autonomous machine to perform reliably, safely, securely, in a time?deterministic manner, etc. It should be noted that despite autonomous machines introducing specific dependability challenges, the scope of this document is not limited to full autonomy; rather, it covers the whole spectrum of degrees of automation, from no automation to full automation. Also worth noting is that nowadays dependable machines are extremely connected. Using automotive as an example, Connected Automated Vehicles (CAV) involves interactions with other vehicles and infrastructure (V2X) and also link with the cloud for functions such as fleet management, teleoperation, maps, Over the Air (OTA) updates, etc. Therefore, the proposed lifecycle is considered end to end, covering the dependability aspects of connectivity as well. The landscape includes the definition of needs for each activity, in terms of methodologies, description languages, data models, and databases that have been identified as necessary or critical to perform those activities. The goal of this work is to provide structures and directions to allow a seamless exchange of information and interoperability between activities at the same or different level of abstraction.