The WSJ in a subsequent story identified George Person, recently retired head of the recall division at NHTSA, as the source.  (see No Black Box Exoneration for Toyota and Lawsuits Fill in Outline of Toyota Sudden Acceleration Cover-Up)

Person’s leak appears to have prompted the briefing and, based on the briefing memo, Person accurately described the findings.  But what is missing from the WSJ – and most other reports about the data – is context.

(And while we’re on the topic of Person, we failed to mention that we do appreciate Mr. Person’s concern about government transparency – Person told the WSJ “When I asked why it hadn’t been published, I was told that the secretary’s office didn’t want to release it.”  But he lost us when he claimed “The agency has for too long ignored what I believe is the root cause of the unintended acceleration cases … It’s driver error.  It’s pedal misapplication and that what this data shows.”  Sorry George, but, we still hold fast that your statement tells more about your bias than it does the facts.)

Let’s start with the data, what is it and what has NHTSA disclosed about it:

NHTSA claims it conducted 58 field inspections and the vehicles it selected for inspection were crashes in which:

“There was an allegation of unintended acceleration or the possibility of unintended acceleration based on preliminary incident information; the vehicle was still available with the EDR intact; the vehicle contained an EDR with pre-crash data; and the owner of the vehicle was willing to allow NHTSA to read the EDR.  It is also important to note that most Toyota models manufactured before 2007 were not equipped with EDRs capable of pre-crash data.”

NHTSA noted that it gleaned the following breakdown from the data in the 58 cases:

35 showed no brake application

14 involved partial braking

9 involved braking late in the crash

3 involved early braking

2 involved mid-event braking

1 event was said to have involved pedal entrapment

1 event showed both brake and accelerator application

1 case the EDR contained information related to a separate incident

1 case NHTSA is still working to resolve inconclusive data from the EDR

5 cases resulted in no EDR activation at all

How has NHTSA characterized these findings:

“At this early point in its investigation, NHTSA officials have drawn no conclusions about the additional causes of unintended acceleration in Toyotas beyond the two defects already known – pedal entrapment and sticking gas pedals.”

NHTSA made it clear that these data are only a small piece of the puzzle, and while driver error is a likely cause of some SUA events, the EDR data don’t make for a compelling case that this is all that’s happening – particularly as independent experts and Toyota continue to document events in which the vehicle diagnostic systems fail to detect unwanted acceleration events.

NHTSA has included in the 58 cases events that have “the possibility of unintended acceleration based on preliminary incident information.”  How many is not known.

We have examined many possible SUA crash events too, and those “possible” SUA events tend to involve single vehicle run-off-the-road crashes with no explanation and no witnesses.  While these events need to be investigated, they are far from the more typical events in which no EDR is activated, the driver, passengers, other witnesses, or in some cases Toyota dealers report first hand engines racing that can’t be explained by mechanical interference or driver error.  Some of these cases were detailed in the Multi-District Litigation complaint against Toyota which cited dealer reports produced by Toyota.  (see Lawsuits Fill in Outline of Toyota Sudden Acceleration Cover-Up)

The NHTSA EDR data don’t address pre-2007 Toyota / Lexus model vehicles as those vehicles are not equipped with EDRs that capture pre-crash data – yet many of the pre-2007 vehicles (including the Camry and Tacoma) have the highest SUA complaint rates and are not part of any recall.

Safety Research & Strategies has already addressed the murkiness and accuracy of Toyota’s EDR’s (see EDR: Toyota’s Electronic Doubt Receptacle).  And since most of the news reports failed to mention it, we would like to remind you again of an important bit of context: Toyota has always stated that the accuracy of the black boxes has never been scientifically validated. In fact, the company fights to keep the data from being used in litigation because it says the EDR data aren’t reliable.

Now, NHTSA’s briefing yesterday noted the agency has done its own EDR validation testing on two 2007 Camry’s and one 2008 Highlander and they have found “Toyota EDR data the same as the data produced by the NHTSA test equipment.” So, what’s the take-away from this set of data:  The Toyota EDRs report accurate information when the vehicle is functioning properly.  What happens when a vehicle experiences an unintended event, sets no error code, and a crash ensues?

In summary, upwards of 95 percent of the SUA reported incidents don’t involve EDR activation.  The recent NHTSA EDR data summary is interesting and worth further examination, but it clearly cannot be extrapolated to the thousands of incidents of SUA in which an EDR didn’t activate and witnesses report racing engines without driver input or mechanical interference.

The MDL, filed this week on behalf of Toyota and Lexus owners alleging that the automaker’s SUA defect has caused their vehicles to lose value, shows that Toyota has known, at least since May 2003 that its Electronic Throttle Control had a “dangerous” unintended acceleration problem with an unknown cause. That civil action, and a second one claiming damages for Toyota and Lexus owners who were injured or killed in crashes alleged to have been caused by SUA, cite six incidents which occurred between 2003 and 2010, witnessed by Toyota technicians, dealers and others. The e-mails also show that Toyota spent considerable energy trying to divert NHTSA from looking too closely at the issue. Here are some highlights from the class-action complaint:

“On May 5, 2003, in a ‘Field Technical Report’ Toyota acknowledged the fact that “[s]udden acceleration against our intention,’ was an ‘extremely serious problem for customers.’ The technician reported a UA incident and stated: ‘We found miss-synchronism between engine speeds and throttle position movement.” The probable cause was unknown, but ‘(e)ven after replacement of those parts, this problem remains.’ The author requested immediate action due to the ‘extremely dangerous problem’ and continued: ‘[W]e are also much afraid of frequency of this problem in near future.’

“In May 2004, a 26 Forensic Technologist and MSME examined a vehicle in New Jersey that had experienced a UA event. The report was forwarded to Toyota on January 13, 2005. It concluded that the vehicle’s ETCS was not operating correctly. This report was not provided to NHTSA.

“In a February 27, 2007 email sent by Michiteru Kato to Christopher Santucci, Mr. Kato decided against sending his most knowledgeable ECU (Engine Control Unit) engineer to an ECU demonstration being conducted for NHTSA, in order to avoid questions regarding ECU failures: ‘…I thought that 3 guys from TMS is too many (two at most), and if the engineer who knows the failures well attends the meeting, NHTSA will ask a bunch of questions about the ECD. (I want to avoid such situation).’

In a Field Technical Report dated December 12, 2008, a technician stated: ‘After traveling 20-30 feet the vehicle exhibited a slight hesitation then began to accelerate on its own. Engine speed was estimated to have gone from 1500 rpm to 5500 rpm at the time of the occurrence…Probable Cause =Unknown.’

The examples cited in the MDL complaint are similar to many of those reported to Safety Research & Strategies by Toyota owners. (see Toyota Real Stories)

George Person, erstwhile chief of the Recall Management Division, who just retired after 27 years at the agency, has no reservations about the causes of Toyota SUA. He opined that DOT officials had withheld a report, showing that EDR data in 23 crashes showed a wide-open throttle and no evidence of braking before the crash, out of embarrassment.

“It has become very political. There is a lot of anger towards Toyota,” Person said in a subsequent WSJ story published on July 30. Transportation officials “are hoping against hope that they find something that points back to a flaw in Toyota vehicles.”

“The agency has for too long ignored what I believe is the root cause of these unintended acceleration cases,” he told the WSJ, as his head disappeared into the hole George was digging. “It’s driver error. It’s pedal misapplication and that’s what this data shows.”

Person lobbed what looked like a boulder three weeks ago, when the WSJ reported that NHTSA had Event Data Recorder (EDR – aka, “Black Box”) from “several dozen” crashes and that all of them showed accelerator application with no braking at the time of impact:

“The Toyota findings appear to support Toyota’s position that sudden-acceleration reports involving its vehicles weren’t caused by electronic glitches in computer-controlled throttle systems, as some safety advocates and plaintiffs’ attorneys have alleged.”

The front-page story was picked up, with some outlets blaring headlines of the “Toyota-is-Exonerated” variety. The pushback was swift: NHTSA said that it was in no way close to drawing any conclusions about what has caused tens of thousands of Toyota Sudden Unintended Acceleration complaints. (The agency is currently examining Toyotas with the National Aeronautics and Space Administration and is awaiting a report it commissioned a National Academies of Science panel to write about SUA.)  And, after a spate of such stories, USA Today captured the entire charming episode under the headline: “Feds, Toyota, deny they said what nobody said they said in acceleration fracas.”

First – while we are deeply disappointed that this bears repeating, we repeat – Toyota EDR data alone is unreliable, by the automaker’s own admission. Second, there are so many details of these 23 crashes still not public. How were these crashes selected? Are these single-vehicle, run-off-the-road fatal crashes with no witnesses that someone has surmised could have been SUA related?  If they are how relevant are these crashes to the SUA problem? (See No Black Box Exoneration for Toyota) Most of the Toyota SUA events have not triggered the EDRs. They are low-speed events that may have little or nothing to do with the higher-speed crashes.  Further, those much more frequent, lower-speed events have had many witnesses.

The e-mails submitted as exhibits in the MDL make Person’s boulder look like a pebble.

Toyota’s efforts to place the story with the Journal seem to be paying dividends –  literally. The automaker’s stock rose 1 percent on the news and reporters scrambled to repeat the Journal piece with no independent sources.

So, does this bring the curtain down on Toyota’s SUA woes and concerns about their electronics?  Hardly.

Recall that Toyota reported to Congress in January that the company identified 37,900 customer contact reports “potentially related to sudden unintended acceleration” analyzing “dozens” of data recorders from the thousands of complaints doesn’t extrapolate to a driver error problem.  Nor does it explain the large jump in complaint rates when Toyota moved to Electronic Throttle Control (ETC).  Most complaints and crashes do not actually activate the EDR, which only records data in crashes severe enough to deploy an airbag or, in some instances, in near-deployment events.  The actual pool of unintended acceleration claims in which an EDR was activated is very small.

Safety Research & Strategies has already addressed the murkiness and accuracy of Toyota’s EDR’s (see EDR:  Toyota’s Electronic Doubt Receptacle).  And since the Journal failed to mention it, we would like to remind you of an important bit of context: Toyota has always stated that the accuracy of the black boxes has never been scientifically validated. In fact, the company fights to keep the data from being used in litigation because it says the EDR data isn’t reliable.

In order to extract and read the data stored on a Toyota EDR, proprietary equipment that downloads, analyzes and generates a report based on the data is required. This equipment is not available to the public.  The EDR must be downloaded and any reports must be generated by Toyota or NHTSA, which has recently been provided with readout tools.  Until March 3, 2010, when Toyota delivered one readout tool to the NHTSA, Toyota claimed that it had a single prototype tool in the U.S. that could extract the data and that it would only download data if requested by law enforcement, NHTSA or the courts.  The details of the quantity and quality of the Toyota EDR data have been shrouded in secrecy.  No one, other than Toyota, knows exactly what data is recorded, retrieved and how it is processed and analyzed to produce a report.

According to Toyota, the type of data recorded varies depending on which generation of EDR is in the vehicle. Toyota doesn’t disclose prior to the download which generation of EDR is installed on specific vehicle makes, models and years and what data is available on each version. The owner of the vehicle does not know what is being recorded, and when data are downloaded they have no way to determine whether the data downloaded is complete, how the data are being processed or the accuracy of the translation.

Toyota now wants the public to believe that an analysis of “dozens” of black boxes out of the thousands of unintended acceleration complaints somehow translates to an exoneration of electronics.

Toyota Fault Detection Capabilities in Question

Dr. David Gilbert the Southern Illinois University Carbondale professor of automotive electronics was the first to point to the Toyota fault detection capability as an area of interest.  Gilbert’s testing found that Toyota’s could lose the signal redundancy at the accelerator pedal sensors – an important failsafe that is in place to prevent unintended acceleration – without any detection.  Once the failsafe was lost, introducing voltage to the sensor would cause the throttle to open.  While Gilbert’s study didn’t pinpoint the root cause of SUA, his findings pointed squarely to problems with the electronic fault detection that could help explain why an engine could race without driver input and leave no error codes – a common complaint.

Since Gilbert disclosed his findings in testimony to Congress – and several independent experts have validated Gilbert’s findings, more evidence that the fault detection system shortcomings in Toyotas continues to mount.  The SUA issues affecting Toyotas don’t appear to be due to a single root cause, rather, they appear to be facilitated by the lack of a robust diagnostic and fault detection strategy. So, when something does go wrong – whether mechanical or electronic – problems can go undetected and the engine can race without driver input or setting an error code.

An examination of a 2004 Toyota Camry by in Ft. Lauderdale, Florida revealed yet another hole in Toyota’s fault detection strategy. The vehicle had experienced an unintended acceleration event, and exhibited an intermittent mechanical throttle sticking condition. Independent experts and Toyota technicians witnessed the engine race to nearly 3,000 RPMs when the throttle plate stuck.  But the electronic controls failed to detect the stuck throttle and limit engine speed to ensure a safe condition.

Dr. Todd Hubing, a Clemson University professor of vehicle electronics, presented similar findings to the National Academy of Sciences panel charged by NHTSA with examining unintended acceleration. Hubing was able to replicate Dr. David Gilbert’s work and obtain wide-open throttle without the fault detection system setting an error code – but with only a single fault.  Gilbert’s analyses found that first a loss of signal redundancy at the accelerator pedal sensor was needed followed by a voltage spike to create an unintended wide-open throttle.  Hubing found that many of the faults created invalid signals that sometimes would be detected, other times not.

The Real World

Simple explanations are nice, but they just don’t square with all the facts. Toyota owners have complained about sudden unintended acceleration in a variety of scenarios – some have been long duration, highway speed events, where the driver was already using the accelerator. Some have experienced multiple events. Some have experienced multiple events in the same vehicle with different drivers at the helm.

Within six weeks of leasing a 2010 Camry from Campbell Toyota, both Debrah and Dave Auger of Chatham, Ontario experienced three bouts of sudden acceleration.

Two incidents occurred when Debrah was at the wheel. The first took place in the parking lot of a small strip plaza. Debrah was parking, and had shifted into reverse, when the vehicle “jolted back quite hard,” Debrah recalled. She threw the transmission shift into park. The incident was puzzling, but the Augers passed it off as a one-time glitch.

Dave Auger, an experienced law enforcement officer, was driving at about 30 miles per hour on a four-lane highway in Port Huron. He had just taken his foot off the gas, when the Camry suddenly surged forward. “It felt like a hand suddenly shoved the car forward,” Auger said. He quickly applied the brakes and pulled the vehicle to the side of the road.

The third and final experience occurred as Debrah Auger was pulling up to a one-way stop in the Augers’ subdivision. She had made a complete stop, and had taken her foot off the brake, when the Camry took off. Debrah jammed on the brake, but she couldn’t stop it before the vehicle had surged into the intersection.

And don’t forget the Haggerty case, the first significant incident that pointed squarely to electronic issues:

Kevin Haggerty, owner of a 2007 Avalon, experienced five different SUA events.

They could not be blamed upon floor mats: Haggerty did not have accessory floor mats, and his OE mats were secured in place.

They could not be blamed on a sticky pedal: Several times, the vehicle accelerated without his foot on the gas pedal. The engine would sometimes return to idle after driving a few miles or after the Avalon shut down and restarted or was stopped and put into park.  On December 28, 2009 Haggerty, a volunteer firefighter and salesman, was driving to work on a highway when the car began to accelerate without his foot on the gas pedal. He was unable to stop the car with brakes and shifted into neutral to slow the car down. Only a couple of miles from his local Toyota dealer, Haggerty decided to call the service manager to let him know he coming in. He managed to drive the vehicle by alternating from neutral to drive and pressing very firmly on the brakes.

Heck, Haggerty’s incident couldn’t even be blamed on pedal misapplication: When Haggerty arrived at the dealer he shifted into neutral and exited the car with brakes smoking and the engine’s rpms racing.  The service techs examined the car and found no pedal interference or sticking and could provide no explanation or any computer error codes.

Toyota’s regional representative in Caldwell, NJ later inspected the vehicle, but did not provide details of this inspection to Haggerty. Instead, Toyota Motor Sales authorized replacement of the throttle body and accelerator pedal assemblies and sensors and paid for the $1700 repairs and rental car costs. The Toyota dealer told Haggerty that they were unsure whether the repairs would fix the vehicle.  (Toyota later tried to blame the event sticky pedal, despite service technicians who stated they pulled back on the pedal to no avail.)

While there may be dancing in the streets of Torrance, we think it’s a little early to cue the music.

Law enforcement based this in part on the results of the Camry’s Event Data Recorder (EDR) – aka, “black box” – readout, which appeared to show that the driver Barbara Kraushaar never hit the brake in the five seconds before her Camry struck a Ford Taurus, and killed driver Colleen A. Trousdale.

A news report in Syracuse’s Post-Standard quoted Auburn Police Lt. Shawn Butler, thus:

“What that showed us is the five seconds prior to the crash, that her vehicle was steadily increasing in speed and that there was no application of brake at all. And the accelerator was in both the off and on position, fluctuating, in the last five seconds.”

Because of her condition (unspecified) police did not charge Krushaar in the crash. But did the EDR give an accurate account of the driver’s actions? Would Toyota swear to it in a court of law? We’re guessing not, since Toyota has always maintained that the data obtained by its EDRs is unreliable. (Unless it points to driver error.)

Last month, with great fanfare, Toyota announced that it would make 150 Event Data Recorder readout devices available, with some going to NHTSA and Transport Canada and was furiously training field techs on how to use them. Toyota also announced that it was “developing new policies and procedures for responding to direct customer requests for EDR readouts and data hand-off to help ensure a smoother, more informed process for all parties involved” and was “actively developing plans to transition to a commercially available EDR readout device and software package.”

“By increasing the number of Event Data Recorder readout devices and training more staff across the country, Toyota is better prepared to respond to customer concerns quickly and address their needs more effectively,” said Steve St. Angelo, Toyota Chief Quality Officer for North America. “We have delivered on this pledge to our customers and to Congress as we continue working hard to set a new standard of customer care at Toyota.”

Whew! We got tired just reading about it. Toyota was forced to get busy a few months ago, after admitting to the House Energy and Commerce Committee that Toyota owned the only readout tool in the entire US capable of reading the data from its Event Data Recorder.

It’s too bad that Toyota doesn’t think much about accuracy of the black box results. Toyota has staked out this position numerous times in the press and in litigation.

For example in response to questions by the LA Times on its EDRs, Toyota said:

“Given the fact that the readout tool is a prototype and has not been validated, it is Toyota’s policy not to use EDR data in its investigations. However, Toyota has used the readout tool under certain circumstances.”

In James McAlonan v. Toyota Motor Corporation, et al, Toyota argued that the EDR data retrieved from a 2003 Toyota Echo in November of 2007 should be excluded as evidence because of the lack of reliability of the data. In the affidavit of Toyota Design and Technical Analysis Manger Mark Kastis, he specifically points out that the readout is unreliable because there were errors and anomalies in the readout; that those errors have not been explained by controlled crash test results; repeatable laboratory test results; or extensive field experience; how were they resolved by specific corroboration from physical evidence. He also conceded that the readout tool has never been validated:

“The prototype readout tool used to perform the readout in this litigation had never before been used to read out data from a 2003 Echo EDR involved in a real world crash. The readout tool has not been validated as a reliable device to accurately convert the data contained in this EDR to the form presented in the readout report,” Kastis said.

Similarly, the general information section in Toyota’s SRS Event Data Recorder Operation Manual specifically states:

“The accuracy of the memory of Toyota’s Event Data Recorder (“EDR”) is still being validated, and the readout tool for the EDR is still in the prototype stage. Toyota cannot verify the complete reliability of such information, unless such data can be independently corroborated, e.g., through physical evidence, etc.”

According to publicly-available sources, Toyota has been installing airbag Event Data Recorders in its vehicles since the 2001 model year, and vehicle stability control EDRs since the 2000 model year, both focusing on frontal crashes.  In 2002, Toyota expanded capabilities to include rollover events. In 2004, it developed technology to incorporate side impact collisions.  However, according to a deposition of Toyota engineer Motoki Shibata, Toyota has actually been able to record and download vehicle data as far back as 1997.  In addition, Toyota’s Hybrid vehicles can report Operation History Data which records special operations performed by the driver and the number of times abnormal conditions that have been input into the Hybrid Vehicle (HV) control ECU.  The history recorded includes accelerator and brake application information. This data is retrieved using a Toyota tool called a Techstream. Unlike the EDR readout tool, this is available to the public for purchase.

The details of the quantity and quality of the Toyota EDR data have been shrouded in secrecy.  No one, other than Toyota, knows exactly what data is recorded, retrieved and how it is processed and analyzed to produce a report.

According to Toyota, the type of data recorded varies depending on which generation of EDR is in the vehicle. Examples of the data that can be recorded include engine speed, whether the brake pedal was applied or not, vehicle speed, to what extent the accelerator pedal was depressed, position of the transmission shift lever, whether the driver and front passenger wore seat belts or not, driver’s seat position, SRS air bag deployment data and SRS air bag system diagnostic data. What Toyota doesn’t disclose prior to the download, is which generation of EDR is installed on specific vehicle makes, models and years and what data is available on each version. The owner of the vehicle does not know what is being recorded and when the data is downloaded, nor do they have any way to determine if the data downloaded is complete and accurately translated.

Lest you think that we are picking on Toyota – other manufacturers have been less than honest about the capabilities of their EDRs in the past – and their EDRS have recorded inaccuracies – or at least large discrepancies between what the readout says and what the physical evidence shows.

In a study entitled, Evaluation of Event Data Recorders in Full Systems Crash Tests, the authors examined Ford and General Motors EDRs. They concluded:

“The majority of the EDRs examined in this study did not record the entire event. In one-third of the GM tests (10 of 30), 10 percent or more of the crash pulse duration was not recorded. In two of the four Ford tests, the last 100 ms of the crash pulse was not recorded. A data loss of this magnitude would prevent a crash investigator from using an EDR to even estimate the true delta-V of a vehicle.”

In a presentation, GM’s then Executive Director of Vehicle Safety, Robert C. Lange noted that when using EDR data, one must always account for and correlate data with physical information. In Mary Acker v. General Motors, the EDR recorded an air bag deployment but the air bags did not actually deploy. In Prudencia Sanchez v. General Motors, the EDR download indicated that the driver’s seat belt status was buckled in the non-deployment event and unbuckled in the deployment event that occurred immediately afterwards.

What is disturbing in Toyota’s case is how readily Toyota will use it to impeach the testimony of consumers and how law enforcement officials, who aren’t aware of the unreliability of Toyota EDR data and murkiness of its download process accept the results.

So, to review: Toyota’s EDR data is unreliable; the type of data recorded for each make and model is known only to Toyota; the readout tool has never been validated.

One might logically ask: if Toyota’s EDR is so useless, why make so many readout tools available? For that, you need the readout tool for Toyota press releases, developed by Safety Research & Strategies. Don’t worry: anyone can have one and no special training is required! Just substitute the words “public relations” anytime you see the word “quality” or the acronym “SMART,” as in the automakers new rapid response “SMART” teams.

Works pretty good.