• Mar 8, 2010
Mechanism used to "force unintended acceleration - Click above for high-res image gallery

In a video webcast for the media that concluded a just a few moments ago, Toyota hit back at the "simulated unintended acceleration" demonstration that was shown by ABC News late last month. When we saw the original report, we postulated that the condition Gilbert produced may not be representative of a scenario that can actually happen in the real world, and judging by Toyota's findings, we were right.

As reported on Friday, Toyota went to an independent testing firm called Exponent to attempt to replicate the results from the Gilbert study. Dr. Shukri J. Souri of Exponent acknowledged that Gilbert did indeed create a scenario that produced what looked like a valid accelerator pedal signal to the electronic engine management system. But before demonstrating how the test was performed, Souri explained how the pedal sensor wiring works and showed how the connector is constructed. As we expected, Gilbert's testing methods and ABC News' report are very much in question. Make the jump for our full breakdown.



[Source: Toyota]



Tired of Toyota recall news? Try out the recall-free version of Autoblog.


The Toyota's normal pedal sensor configuration consists of two independent sensors that produce voltages in different ranges, and there should always be a correlation between the voltages. One of the ways a fault in the sensor system can be detected is if the signals does not rise or fall in the correct relationship.

The wires in the connector are in-line, and Souri's team actually sectioned a connector to show all the insulation between the six wires. Gilbert's setup required breaching three of the six wires, attaching a 200 ohm resistor between two of them and then shorting the third line with full power to the resistance circuit.

This methodology provides a voltage input to the two circuits. The presence of the 200 ohm resistor ensures that the relationship between the two signals remains within the parameters the ECU is expecting. If there is more or less resistance, the relationship between the two signals will not be maintained and a fault should be detected.

Given the nature of the connector between the wiring harness and the pedal assembly, it's highly unlikely that such a scenario could ever occur in the real world. The only way this is could occur is if the insulation on the wires in the harness itself was stripped away (exactly the way Gilbert did for his test) and just the right amount of resistance occurred only between the two signal lines. Obviously, such a scenario is unlikely.

According to Kristen Tabar of Toyota, the resistance range that will work is very narrow. The 200 ohm resistance is not something that could occur with any materials that are likely to appear in the vehicle environment. For example, water intrusion would produce resistance in the 1,000+ ohm region. A straight short of the wires would be well below 200 ohms. Either scenario would trigger a fault code.

During the webcast, Toyota and Exponent demonstrated the same scenario on a Ford Fusion, BMW 325i and Subaru Legacy (they also had a host of other vehicles on hand) and each vehicle replicated the racing engine condition without signaling a fault code, although each one required a different resistance value. Like pretty much every modern vehicle available, these vehicles use a similar type of gas pedal architecture.

So what does this all prove? First, it demonstrates that with a bit of re-engineering of the pedal circuit, any engine can be made to race independently of what the driver commands with their foot. David Gilbert was able to modify the Toyota pedal sensor circuit to make this happen within the parameters that the fault diagnosis system was expecting so that no fault was detected.

What it doesn't prove is that Gilbert's scenario is a possible cause of the issues that drivers have seen. In fact, based on how this test was conducted, it's almost impossible that this is the cause of Toyota's unintended acceleration woes.

In no way does any of this exonerate Toyota with regard to a potential software problem. It's too early tell if a errant black box is the cause and we still have no evidence one way or the other on this count. This also doesn't discount the possibility of some other electrical (such as electromagnetic interference) or mechanical problem that we aren't aware of at this time. So the mystery continues... Toyota's press release is available below.


Show full PR text
COMPREHENSIVE ANALYSIS RAISES CONCERNS ABOUT GILBERT CONGRESSIONAL TESTIMONY, ABC NEWS SEGMENT

Sequence and Nature of Artificially Manipulated Faults Unrealistic, Televised "Unintended Acceleration" Staged With Virtual Remote Throttle
Control

TORRANCE, Calif., March 8, 2010 /PRNewswire/ - Today, during a live webcast, Toyota raised serious concerns about the validity, methodology and credibility of a demonstration of alleged "unintended acceleration" in a Toyota Avalon by Professor David Gilbert of Southern Illinois University and depicted in ABC News broadcasts and on-line segments.

A comprehensive analysis conducted by a world renowned engineering group, as well as testing by Toyota, has concluded the following about Professor Gilbert's demonstration:

The vehicle's electronics were rewired and reengineered in multiple ways, in a specific sequence, and under conditions that are virtually impossible to occur in real-world conditions without visible evidence.
Toyota vehicle electronic systems were actively manipulated to mimic a valid full-throttle condition.
Substantially similar results were successfully created in vehicles made by other manufacturers.
In the demonstration dramatized by ABC on February 22, Professor Gilbert, assisted by segment reporter Brian Ross, asserted that he had detected a "dangerous" flaw in the Toyota electronic control system that he alleged could lead to unintended acceleration. The following day, Professor Gilbert offered a preliminary report of his findings in testimony to the U.S. House of Representatives Energy and Commerce Subcommittee on Oversight and Investigations.

Engineers at Exponent, one of the country's leading engineering and scientific consulting firms, as well as Toyota engineers, have reviewed and recreated Gilbert's demonstration with substantially similar results in representative vehicles of other makes.

Separately, at Toyota's request, Dr. J. Christian Gerdes, associate professor of mechanical engineering at Stanford University and the director of the Center for Automotive Research at Stanford (CARS), conducted an independent review of Professor Gilbert's testimony and the preliminary report presented to Congress.

Their findings were demonstrated today at a news conference during which the accelerator circuitry of a Toyota Avalon, as well as a sampling of well-regarded and popular competitive makes, was rewired and manipulated as Gilbert did in his demonstration.

Kristen Tabar, general manager of electronics systems, Toyota Technical Center, summarizes three of the major concerns with the artificial nature of Professor Gilbert's demonstration.

"First, an electrical circuit that has been reengineered and rewired will not behave as it was originally designed and engineered," said Tabar.

"Second, no automaker can or should be expected to design detection strategies for artificially created events in the absence of any evidence that such an event can occur in the real world.

"Third, if the artificial condition created by Professor Gilbert had occurred in the real world, it would have left readily detectable fingerprints."

Exponent and Toyota engineers have found no evidence to suggest that any of the steps of Professor Gilbert's demonstration exists in the real world. Thus, the fact that the Toyota Avalon used by Professor Gilbert did not show a Diagnostic Trouble Code after his demonstration does not indicate an undetectable safety defect. The same is true of the representative vehicles of other manufacturers tested by Exponent and Toyota.

Professor Gilbert's reengineering and rewiring of the vehicle's electrical system involves the following manipulations in a specific sequence. First, the protective insulation on two separate wires that carry the accelerator pedal position signals to the Engine Control Module must be individually cut or breached. Next, these wires are connected to each other through a 200 Ohm resistor.

This contrivance, by itself, did not cause an increase in engine speed. To cause an increase in engine speed, it is necessary to cut the insulation on a third wire, the 5-volt power supply to the accelerator pedal, and force a low resistance connection between the power supply and the secondary signal wire.

The resulting increase in engine speed is a result of the subsequent artificial and sudden application of the 5-volt power supply to this signal line with the rewired circuit. When subjected to similar unrealistic reengineering and rewiring, the competitive vehicles evaluated by Exponent and Toyota achieved substantially similar results with varying levels of resistances.

This manipulation of electrical components and a power source created artificial voltages that the engine control module, or ECM, would interpret as valid accelerator pedal signals. In essence, this test created a virtual, remote control accelerator pedal that replicated the vehicle's own normally functioning accelerator pedal.

Also contrary to statements made in the ABC News story, had short circuits of the kind artificially created by Professor Gilbert occurred in real-world driving conditions, they would have left visible evidence such as damage or deterioration of the wiring and components.

As revealed in their testimony before Congress, Professor Gilbert's Preliminary Report was commissioned by Sean Kane, a paid advocate for trial lawyers involved in litigation against Toyota and other automakers. Mr. Kane also appeared on the ABC News broadcast in support of the claim that Professor Gilbert's demonstration revealed a flaw in the electronic throttle control system that could lead to "runaway" Toyota and Lexus vehicles. The relationship between Mr. Kane, Professor Gilbert and the trial lawyers who support Mr. Kane's advocacy was not revealed by ABC News during the newscast, nor was Toyota offered an opportunity to view the demonstration or given time to respond.

Toyota believes that the public and Congressional committees have been misled by Professor Gilbert's demonstration and the dramatization of it by ABC News. This has cast unwarranted doubt on the safety of Toyota and Lexus vehicles. Toyota remains confident in the integrity of the electronic throttle control system in its vehicles and there has been no reliable evidence of any kind to the contrary presented to the media or to Congress. Toyota's electronic systems have multiple fail-safe mechanisms to shut off or reduce engine power in the event of a system failure. Extensive testing of this system by Toyota has not found any sign of a malfunction that could lead to unintended acceleration.

Toyota has commissioned Exponent to conduct a comprehensive analysis of the electronic throttle control systems in Toyota and Lexus vehicles. No limitations of any kind were imposed on Exponent by Toyota. This evaluation is ongoing. An interim report of Exponent's findings has been provided to Congress and establishes the functionality of the electronic throttle control fail-safe systems. The final results of Exponent's exhaustive analysis will be made public when completed. As with all such reliable engineering analyses, Exponent's final results will provide the data and information necessary for others to validate Exponent's conclusions.

Exponent, Inc. is a leading engineering and scientific consulting firm with expertise in over 90 technical disciplines. Exponent has a full-time staff of over 900 located in 23 international offices. Exponent's multidisciplinary organization of engineers, scientists, physicians and business consultants, addresses complicated issues facing industry and government today. The firm's consultants provide product design analysis, development, and testing; analyze failures and accidents to determine their cause and prevent their recurrence; and evaluate environmental and human health concerns to find cost-effective solutions. Exponent is certified to ISO 9001 and is authorized by the General Services Administration (GSA) to provide professional engineering services. For information about Exponent capabilities and credentials visit www.exponent.com

Visual demonstrations of Toyota and Exponent testing, as well as detailed information regarding Toyota's electronic throttle control system operation and testing for electrical and electromagnetic interference can be found at www.toyotanewsroom.com.


I'm reporting this comment as:

Reported comments and users are reviewed by Autoblog staff 24 hours a day, seven days a week to determine whether they violate Community Guideline. Accounts are penalized for Community Guidelines violations and serious or repeated violations can lead to account termination.


    • 1 Second Ago
  • 71 Comments
      • 4 Years Ago
      One word... "KAIZEN!!!"
      • 4 Years Ago
      People don't put all-weather floormats on top of regular floormats! There is a warning on most aw mats that say so. This is what happened to the Saylor family in SD. The lexus dealer (a privately-owned franchise) made that mistake and should be culpable, not Toyota in that case.
      • 4 Years Ago
      I'm not a Toyota fan (quite the contrary), never owned a Toyota (but ex wife had 2), but...

      This is no different than feeding to the ECU/ETCU the exact signals for WOT on both channels.

      The only thing that idiot Gilbert proved is that if you feed on both channels the EXACT valid voltages for WOT, the ECU will command WOT - Duh!!!

      He simply achieved that by an intricate sequence of wire cuts, resistor soldering, etc.

      Moronic. *GILBERT* (during his testimony) and Gilbert/ABC during the news video went out of their way to avoid describing what he did. They misled the public into thinking a simple short or other plausible scenario caused the WOT.

      He could have said "By an unlikely (in the real world) combination of wire cuts and resistor soldering, it is possible to cause uncommanded WOT without an error code". That would have been science and honorable.

      Shame on both.
      • 4 Years Ago
      "First, it demonstrates that with a bit of re-engineering of the pedal circuit, any engine can be made to race independently of what the driver commands with their foot. David Gilbert was able to modify the Toyota pedal sensor circuit to make this happen within the parameters that the fault diagnosis system was expecting so that no fault was detected. "

      I don't think his experiment was to show the cause of UA, but to show why after a UA incident there would be no computer fault codes to indicate anything bad had happened. Then it's your word against the automaker. If this leads to more robust data collecting then I'm ok with that.
        • 4 Years Ago
        I think they're also trying to show that this type of scenario can be done with any drive-by-wire gas pedal setup from any manufacturer. Then it's everyone's potential nightmare, not just Toyota's.
        • 4 Years Ago
        No... You have to go through this re-engineering to cause the unintended acceleration without it leaving a fault code.
        This Gilbert guy tried to force the unintended acceleration other ways but they all left a fault code in the ECU.
        Of course, ABC decided not to show any of those experiments.
        • 4 Years Ago
        +1 on this not being an explanation of why UA occurs - and I think that's something that people just aren't getting here. The e-pedal is supposed to supply two (or more) signals (inverse, in some manufacturer's setups) that are interpereted by the ECU. Whenever the two signals are not in agreement, the ECU should go into a fail-safe mode, and should raise codes.

        That the experiment contains factors that are unlikely to occur naturally is completely irrelevant. The fact of the matter is that the post-mortem work done by Toyota made the faulty assumption that their telemetry is infallible. The experiment proves that it is not.

        I'm not anti-Toyota, and this isn't a smoking gun. It should, however, be a wake-up call to Toyota to re-examine the data without making a faulty assumptions. Toyota has a large number of talented engineers, but it's tough to diagnose a problem (or come up with an effective solution) if your data is wrong or incomplete.
        • 4 Years Ago
        Well this still doesn't explain why Toyota allows full 5V as a valid signal on some models at least.

        But one thing Dr. Gilbert seemed to be hinting is that off-idle Toyota's system would loose its redundancy in case of a short. Toyota is now specifically denying that and until demonstrated otherwise, I'll have to believe them. I'm still curious as to why Toyota's system doesn't have a fixed ratio of 50% like most other systems that don't use opposing slopes.

        And it certainly doesn't explain why Toyota's sensors aren't in a sealed assembly outside the main pedal assembly like most other makes.

        Still, it looks like it'll be a bit longer before anything is proven one way or another.
      • 4 Years Ago
      "I don't think his experiment was to show the cause of UA, but to show why after a UA incident there would be no computer fault codes to indicate anything bad had happened. Then it's your word against the automaker. If this leads to more robust data collecting then I'm ok with that."

      PLUS 1!! Not to mention it has to make you question whether or not the vehicle can determine when it is accelerating in an unintended manner.
      • 4 Years Ago
      This professor is just jumping on the bandwagon to get notoriety....that's it. He wants money.
      • 4 Years Ago
      The most serious problem is David Gilbert may not have enough credentials. According to SIU-Carbondale web site, his PhD is in Workforce Development. His department is to produce Automotive Technitians, not Engineers. I seriously doubt that he has necessary credential to make his claims. In order to make such claims, in this case, one should have at least Professional Engineer level credential. The difference between technitian and engineer is similar to that of physician assistant and medical doctor.
      Thank you.
      • 4 Years Ago
      "During the webcast, Toyota and Exponent demonstrated the same scenario on a Ford Fusion, BMW 325i and Subaru Legacy (they also had a host of other vehicles on hand) and each vehicle replicated the racing engine condition without signaling a fault code, although each one required a different resistance value. Like pretty much every modern vehicle available, these vehicles use a similar type of gas pedal architecture."


      didnt we all say the same thing? You could rig any car to do the same thing. You could rig a toaster to explode, a computer to melt down or a dryer to spin out of control. Gilbert and his cronies should be ashamed of themselves.


      Look, you can take marketshare from Toyota over years and years of making great cars and inching your way up, or you could decimate Toyota with a witchhunt and catch up faster. I guess the easy way does it for Government Motors.
        • 4 Years Ago
        @neptronix: you failed at reading comprehension today. So a full and complete rebuttal is finger pointing? No manufacturer's ECUs detect artificial conditions which mimic a VALID (not faulty as you say) signal from the pedal.

        And everyone keeps mentioning the Lexus ES crash in other news threads. Well, that was a floor mat issue. Sorry, but if you can't figure out how to put the damn car in neutral, of course you're gonna crash. He had 16 miles to figure that out, instead, he rode the brakes. I encourage anyone to ride their brakes for 16 miles and see if the pads don't disintegrate and catch fire. And once that brake fluid boils, they're done.
        • 4 Years Ago
        Jason:
        I would love to see the output of the pedal. It seems strange you can divide a signal exactly in half with a single resistor and not a resistor ladder. I somewhat suspect the signal at VPA1 in the diagram is either not linear or not exactly half of VPA2 as it is supposed to be. I'd love to do the experiment myself, but I don't have a Toyota with TBW to try it with.

        It is quite possible the hack this guy made doesn't really mimic a pedal signal exactly and thus Toyota's ECU should catch this. Regardless, this particular circuit (200 Ohm bridge) is not likely to happen in the real world, so fixing any failure to throw a code in this case would be lower priority than finding the actual problems or putting in fail-safe override systems.
        • 4 Years Ago
        Not necessarily. A computer can easily be programmed to be skeptical of it's input. If it's receiving invalid numbers ( as it will when the pedal is truly broken ), it can trigger a fault. Other manufacturers may not have this safety-check in place, but they also didn't install intermittently faulty pedals in their cars.

        Toyota's just pointing the finger at other people. The test may not make the pedals fail like they do in real life ( nobody knows how that happens yet ), but it still shows that Toyota's computer cannot handle faulty input. ALL manufacturers should have computers capable of doing that.

        Toyota just happens to be in the limelight because they lied about the problem, just like ford lied about explorer tire blowouts + cruise control switch fires & got totally railed by the media then too.
      • 4 Years Ago
      He stated that his test was ONLY to show a Toyota engine could have a SUA and not record a fault code. Because there is NO code Toyota does not look to fix the real problem. Toyota has confirmed that he Gilbert was correct. Toyota saying there is no electrical problem is based on fault codes no being on the computer, since they have confirmed that the computer may not show fault codes this means they have to fix the faulty computers.
        • 4 Years Ago
        Fault codes indicate a specific malfunction. Short-circuiting the throttle is something that your car will never achieve by itself without somebody doing it deliberately, so there is no fault code for that condition.
      • 4 Years Ago
      Its about as realistic as "simulating" an old-style gas-pedal problem by putting a brick on it. Sure, it'll rev then...
      • 4 Years Ago
      Yes, and it sickens me to see them continuously dance around specifics of the issue and not take care of it.

      One hand shooting themselves & the other hand digging their grave.
      This report sounds good in their favor, but i am a programmer and know better than this.

      The computer can detect any faulty condition if it is programmed to do so.
      Any experienced programmer could have implemented a true fail-safe system in a day.

      The ECUS are likely firmware-updateable ( pretty much standard issue across all manufacturers ) , so the cost of replacing the ECU shouldn't be a factor.
      • 4 Years Ago
      We know the drive by wire helps the engineers tune performance but adding more complexity to such a important area, has been a accident waiting to happen. Throttle position sensors are bad enough.
    • Load More Comments