washingtonpost.com
Frank Ahrens: Why it's so hard for Toyota to find out what's wrong

By Frank Ahrens
Sunday, March 7, 2010; G01

I won't lie to you: I was not a good engineering student. That's one of the reasons I went into journalism. But I managed to acquire a bachelor of sciences in mechanical engineering, and the recent Toyota hearings on Capitol Hill brought back a lot of memories. Specifically, memories about how engineers figure out why mechanical things fail.

It was made painfully clear at the hearings that a number of lawmakers do not understand the process. An exchange between Del. Eleanor Holmes Norton (D-D.C.) and Toyota President Akio Toyoda illustrated the problem.

Toyoda said that when his company gets a complaint about a mechanical problem, engineers set to work trying to duplicate the problem in their labs to find out what went wrong.

Norton said: "Your answer -- we'll wait to see if this is duplicated -- is very troublesome." Norton asked Toyoda why his company waited until a problem recurred to try to diagnose it, which is exactly what he was not saying.

Members of Congress are generally lawyers and politicians, not engineers. But they are launching investigations and creating policies that have a direct impact on the designers and builders of incredibly complex vehicles -- there are 20,000 parts in a modern car -- so there are some basics they should understand. Chief among them: The only way to credibly figure out why something fails is to attempt to duplicate the failure under observable conditions. This is the engineering method.

"It's just so difficult for people to understand the complexity of the thing," said David E. Cole, chairman of the Center for Automotive Research in Michigan and an engineer. "They don't have the background. They don't have the time to do an investigation. They want to oversimplify a thing that can't be oversimplified."

Toyota is facing an incredibly difficult task. Here's what it knows: It has received hundreds of complaints about unintended acceleration in its vehicles in recent years. People have died in these crashes. Over the same period, hundreds more have died in Toyota crashes that had nothing to do with runaway acceleration. After that, it knows nothing.

Toyota must search its data and look for patterns or similarities among the incidents. Among dozens of variables, it must consider:

-- Road conditions where accidents happened (Is it a rain/snow/sleet/temperature problem?)

-- Where the cars were made (Is it a parts or assembly problem?)

-- How the cars were drawn up (Is it a design flaw? If so, is it a mechanical or electronic flaw? Or a combination?)

-- How the cars were tested (Did the company fail to anticipate a series of events that would lead to a flaw?)

-- Whether it is a fatigue problem (Did something break down sooner than Toyota thought it would?)

Then follows a process of elimination. It's not dissimilar to a doctor diagnosing an illness: Take a thorough reading of the symptoms, then begin eliminating causes. Treat what you think is the illness. If it doesn't go away, treat your second guess at the illness.

Toyota appears from the start to have removed its electronic throttle control from the list of possible causes of the runaway acceleration and focused on two mechanical issues: floor mat entrapment and sticky gas pedals.

But the National Highway Traffic Safety Administration is expanding its search for the problem. It bought a 2006 Lexus that experienced runaway acceleration and will test it for everything it can think of -- including electronic problems. No details about the Lexus testing, which will take place at the agency's facility in East Liberty, Ohio, were available.

So I called Giorgio Rizzoni of Ohio State University, an expert in failure analysis and director of the school's Center for Automotive Research. I asked: How will NHTSA test its Lexus?

First, he said, the vehicle will be outfitted with instruments and taken to a proving ground, where it will be driven for hours to see whether the runaway acceleration can be duplicated.

"But what if you keep driving for days and days and the fault doesn't come back?" Rizzoni said. "What do you do at that point?"

Then, the car goes into the lab and is taken apart.

If the electronic throttle control is the suspected culprit, it is removed from the Lexus and set up on a test bench, where it is affixed with monitoring equipment and a power source. You give it juice and see what happens; see whether there are drops or surges in micro-voltage that could lead to runaway acceleration, for instance.

If that doesn't provide the answer, you go 21st century on the thing. You run something called a "hardware-in-the-loop simulation," where you hook up the physical throttle assembly to a virtual Lexus via a complex computer. This lets you test the throttle control up close while it is subjected to road conditions via the computer simulation. You can run tests for days on end without the expense of a test-driver.

Attention has been focused on mechanical and electronic issues with Toyotas, but Rizzoni raised another possible cause of the runaway acceleration: a software glitch.

He explained that each vehicle contains "layers of computer code that may be added from one model year to next" that control nearly every system, from acceleration to braking to stability. Rizzoni said this software is rigorously tested, but he added: "It is well-known in our community that there is no scientific, firm way of actually completely verifying and validating software."

Here's an example everyone is familiar with: You're working at your computer in Windows software and an error message pops up. It asks whether you want to report the error to Microsoft. Microsoft has exhaustively tested this version of Windows before its release, but it cannot completely predict how it will operate out in the world, subject to user demands. That's why it gathers error reports and uses them to fix the software on a rolling basis.

If you put a lot of parts together to form a complex electromechanical machine and make it talk to itself via software, it can behave, sometimes, in ways you cannot anticipate. It can fail for reasons you cannot anticipate.

That's the problem Toyota faces. And, after thorough testing by Toyota, NHTSA and garage mechanics trying to win the $1 million Edmunds.com prize, no single answer may be found. Obviously, this will not stop juries from awarding damages in the liability lawsuits already filed.

Finally, Toyota can't say this, but I can: Some of the cases of runaway acceleration could have been caused by driver error. Think about the times you've been in an accident, a near-miss or -- more to the point -- a distracted-driving situation that almost veered out of control. You remember the white-hot spike of fear that shot up your spine. You remember the shakes afterward. But do you remember what you did during those few seconds of panic? Do you remember where your feet and hands and eyes went?

Quoting from a 2009 Los Angeles Times article on runaway Toyota acceleration:

"Richard Schmidt, a former UCLA psychology professor and now an auto industry consultant specializing in human motor skills, said the problem almost always lies with drivers who step on the wrong pedal.

'When the driver says they have their foot on the brake, they are just plain wrong,' Schmidt said. 'The human motor system is not perfect, and it doesn't always do what it is told.' "

If you were lucky, your reflexes, muscle memory and driving experience -- and sheer chance -- saved you, and you emerged unscathed from your near-miss. But you could just have easily smashed your foot down on the wrong pedal or jerked the wheel the wrong way. Or hit the radio volume and scared yourself into a dangerous maneuver. Or made a dozen other mistakes.

And none of those would have been the fault of the automaker.

View all comments that have been posted about this article.

© 2010 The Washington Post Company