It looks like Toyota’s reliability problems will be with them – and us – a while longer, so what, if anything, can we learn from the situation?
It seems clear that electronic content in cars will continue to increase due to the entertainment and comfort features we want as well as the need for better fuel economy (reducing weight by replacing mechanical subsystems with electronics).
That means more electronic control units (ECUs) and correspondingly greater communications complexity, since the ECUs need to talk with each other. Mentor Graphics VP Serge Leef says that for any seemingly simple function – hitting the brakes, for example – “a multitude of signals have to travel across multiplexed communications fabric and arrive at the right places at the right times.”
He adds, “When you have so much software, so much communication, and so much underlying computing complexity, there are millions of possible scenarios in the operation of a vehicle.”
Leef and others believe that, in general, automakers are not paying sufficient attention to real-time communication issues and, in particular, are not performing the kinds of tests needed to identify and resolve corner cases – the kinds of pesky, irregularly occurring issues that Toyota is trying to nail down.
Antal Rajnak, chief scientist in Mentor Graphics’ automotive business, adds, “The industry-wide general lack of respect for and understanding of real-time issues, and how to ensure and prove deterministic behavior in distributed systems, is most likely a factor in the (Toyota Prius) braking issue.”
That’s stunning, in my opinion, but it’s understandable if most senior managers in car companies worked their way up the ladder either through mechanical engineering or finance. The mindset needs to change, fast, but it’s obviously easier to say that than to do it.
Is AUTOSAR the best hope for the future? Mentor Graphics and other tools vendors contend that proprietary tools are available now for resolving real-time communication issues. “Some automakers are content to wait until AUTOSAR tools are sufficiently mature; others are willing to assume the risk of adopting AUTOSAR tools as they currently exist – or will opt for proprietary tools,” Leef says. “As electronic content in vehicles continues to increase, it is clear that something must be done.”