Computers and cars... what's not to like?

Just wrapped up 3 days of a great SAE Convergence 2006 conference at COBO Hall in downtown Detroit.  This is the world's premier conference about the convergence of computing and automotive engineering.  Believe it or not, the conference has been going every two years since the 1970s.  Even more surprising, back then 75% of the world's electronics were used in cars and trucks.  Wow! 

Automotive no longer dominates the embedded and electronics industry the way it used to, but the average vehicle today has about 20 electronic controllers (i.e., computers) in it.  With 60 million vehicles produced each year, that's over a billion embedded computers per year (vs. 37 million Dell's last year)!  If the industry could act in a coordinated fashion, it could still move the electronics market.  Of course, this large numbers of controllers is really one of the industry's primary problems.  After all, does it make sense that an average vehicle would have 20+ computers?  Or that a luxury vehicle like the a Mercedes E-Class or BMW 7-series would have 70+ embedded controllers (many 32 bit)?  This has actually been a problem for a long time, but one of the positive things that emerged from Convergence 2006 is that the industry now seems to recognize this is an issue -- which was not the case the last time I was at Convergence back in 2002. 

Among the complementary realizations is this: 90% of the new features and vehicle innovations in the next 10 years are expected to be implemented in software.  90%!  In software!  This means that managing this innovation effectively can not be separated from managing the software development lifecycle processes.  This is a huge challenge for the industry.  Add to that: 40%+ of vehicle warranty issues can be traced to software.  And finally, one more eye-popping factoid: both the source of today's issues and the source of future innovations are from (will be from) the interaction of what started as independent software modules. 

For example, ABS and electronic throttle control interoperating in an intelligent way form Electronic Stability Control.  Add a gyro sensor, and you can create Roll Stability Control, as Ford and Volvo did for their SUVs.  Add an active chassis system, and you can level a vehicle during hard breaking events to apply greater force to rear tires and reduce stopping distance by 15% or more.  This integration of software and hardware modules -- originally designed (and tested) as separate, unrelated entities, is where the most excitement, greatest opportunities, and potentially the biggest pitfalls, all lie. 

How can the computing and automotive industries manage this transition?  And who will be the winners and losers along the way?  These questions, and some potential answers, will guide the next few posts and be a recurring topic here. 

What do you think?