SaFAD: Automotive industry says: safety first for autonomous driving
The safety of autonomous vehicles has been the subject of heated debate since Google surprised the world in 2011 with the news that it had already covered a quarter of a million kilometers autonomously.
Now several participants in the automotive industry, including many German companies such as VW, Audi, Daimler and Continental, have published a white paper in which they take a clear stance: Safety First!
The whitepaper was published via various press portals published (direct download). In the meantime, the abbreviation "SaFAD" has become established, for "Safety First For Automated Driving".
SaFAD is in no way binding. Karl Iagnemma, President of Aptiv, says:
"The paper serves as an aid to the creation of a kind of blueprint or template for coherent and comprehensive thinking on the safety of driverless vehicles." (Source: Forbes)
Safety by design
SaFAD is neither manufacturer-specific, nor is it intended to existing safety standardssuch as ISO 26262. Instead, it is intended to set higher standards and establish a common understanding. While the existing standards sometimes only refer to certain aspects of security, the aim here is to create an overarching understanding.
An important concept here is "safety by design". Safety is not only proven "after development" via tests and analyses, but is incorporated "before development" and ensures an inherently safe design.
Actually, "safety by design" should be a matter of course. But in practice, it is often astonishing (and frightening), how often this is not the case.
Verification & Validation
Of course, this does not mean that verification and validation are becoming less important - quite the opposite. A separate chapter is also dedicated to V&V. In particular, L3 and L4 systems present development with challenges that require new approaches, such as statistical verification (Challenge 1) or machine learning (Challenge 5).
Who's in - and who's not?
The paper was written by eleven companies. The authors are listed by name on page 3. They are Aptiv, Audi, Baidu, BMW, Continental, Daimler, Fiat Chrysler, Here, Infineon, Intel and Volkswagen.
Of course, the question is also exciting: which companies are actively involved in the field of autonomous driving and are not on the list? Tessla and Waymo (Google) in particular are highly active companies that are not on the list. We can only speculate about the reasons. The paper at least positions this initiative as completely open to all interested participants.
Conclusion
SaFAD is an interesting paper that is clearly aimed at development management. It is technical without getting bogged down in detail. If it is understood and implemented correctly by the relevant managers, it has the potential to have a positive impact on the industry.
Whether this succeeds depends on how far SaFAD establishes itself as a framework, whether it is used for supplementary standards and how open and flexible the initiators are. There are a lot of "ifs", but I wish this initiative every success.
Image: SaFAD
