INCOSE/GfSE Webinar 8: Safety assurance under uncertainty
Click to download GfSE - INCOSE Webinar 8 Invitation!
Meeting Title: Safety assurance under uncertainty
Presenter Name: Prof. Simon Burton
Date: Wednesday, 1 June 2022
Time EDT: 11:00 am- 12:00 pm EST (4:00 pm – 5:00 pm UTC; 5:00 pm - 6:00 pm CET)
Abstract
Assuring the safety of modern, highly automated systems presents huge challenges to existing Systems Engineering processes. Such systems are becoming increasingly complex. That is, they exhibit emergent behavior, coupled feedback, non-linearity and semi-permeable system boundaries. These drivers of complexity are further exacerbated by the introduction of AI and machine learning techniques. The result of these developments is an increase in uncertainty in our ability to argue their safety.
Using examples from the field of automated driving, this webinar introduces sources of uncertainty and emergent complexity in the safety assurance process of such systems and discusses why existing safety approaches are reaching their limits of effectiveness.
Simon’s talk will illustrate how defining sources of uncertainty and acknowledging the impact of these risk factors at the governance, management and technical levels of the system is key to developing effective safety assurance strategies. State-of-the-art and current research in this area is summarized. Simon’s talk will conclude with the hypothetical question of whether such systems will ever by “safe enough” and how we could go about arguing such a claim.
Take-Away Key Message
- To understand the path to safe, highly automated AI-based cyber physical systems it is essential to acknowledge the root causes of complexity and uncertainty within the safety assurance process. This webinar will discuss how a holistic view of the system as well as dedicated measures during design and operation can be used to argue the safety of such systems de-spite these challenges.
Register in advance for this webinar at:
https://incose-org.zoom.us/webinar/register/WN_4fxs9diOSfGffFdLBJ4IPQ