Three-day blended short course on the systems engineering and safety assurance of autonomous systems that use AI, from hazards and machine learning assurance to human-autonomy teaming and the autonomous systems safety case.
Addresses the safety of autonomous systems that employ artificial intelligence. The course covers systems engineering and safety assurance for autonomous systems, including how AI is used and assured for autonomy, for decision support and in support of the safety assurance process itself. It examines the disruption autonomy brings to established system safety practice, the elements of autonomy and the mitigation of their hazards, the basics of offline supervised and reinforcement learning, legal, social and ethical issues, human factors and human-AI teaming, post-deployment assurance, regulation and risk acceptance, and a framework for building a safety case for an autonomous system, with an ethics case for such projects.
Taught in a blended format: three days of lectures and extensive group case-study exercises on the University of York campus, plus 15 hours of self-study before and after. It can be taken on its own for continuing professional development, with an optional coursework assessment of about 65 hours that counts towards York's postgraduate awards in safety critical systems engineering.
Facilitator not identified in public sources checked.
Autonomy opportunities and assurance challenges · Elements of autonomy and mitigation of autonomous system hazards · Guidance for developing an autonomous system safety case · Basics of offline supervised and reinforcement learning · Assuring the contribution of AI to autonomy, decision support and safety analysis · Legal, social and ethical issues · Human factors and human-AI teaming · Post-deployment assurance
Design accelerated tests to estimate how products perform over their service lives.
Two-day hands-on course on system and design FMEA using the AIAG-VDA seven-step approach, from boundary and P diagrams to risk analysis, optimisation and design verification planning (DVP&R), with an optional certification exam.
Five-day systems-oriented architectural design course covering physical and logical architecture, MBSE, design decision-making, interfaces and design for reliability, safety, maintainability and producibility.
Course presenters may include: John Fitch