Sudden acceleration is also described by a variety of other terms including  "sudden unexplained acceleration", "uncontrolled acceleration", "unintended acceleration", "uncommanded acceleration", "runaway vehicles", and,  in an NBC TV report (NBC News/MSNBC, Feb. 10, 1999), as  "vehicles taking off on their own". The picture presented is of a vehicle event in which the throttle moves, apparently uncommanded, of its own accord, rather mysteriously, to the fully open position, without the driver pressing the accelerator pedal. In such a situation the driver appears not to be able to exercise any kind of control, except by switching off the ignition or braking, which itself may be dangerous. 

The US NHTSA [National Highways Traffic Safety Administration ] provided the following rather narrow definition of sudden acceleration from near standstill in their 1989 Sudden Acceleration Report:

The NHTSA definition appears to leave out of consideration the following types of incidents :

• any incident that might occur while the vehicle is already in motion, for example:
• a sudden acceleration from 20 to 60 mph, or 40 to 80 mph.
• speed oscillations (hunting) about the set speed.
• overshoot beyond a new set speed with the cruise control on.
• speed drift ( either up or down) with the cruise control on.
• sudden unexplained deceleration.
• lurching
• sudden accelerations when vehicle is pulling in to a parking bay or stopping at traffic lights
  
• incipient incidents -"near miss" incidents - when the driver switched off  the ignition, or managed to apply sufficient brake pressure and brought the vehicle to a halt safely.
• alleged incidents of sudden acceleration in vehicles with manual gearboxes. [These may result in a racing engine, but it is likely that the driver's natural reaction will be to disengage the clutch and then bring the vehicle to a halt by controlled braking.]

• some of the alleged sudden acceleration  incidents in Ford Explorers reported in the UK by the Channel 4 programme in the Autumn of 2000 appear to be outside the NHTSA definition, since (a) they mostly took place with the vehicle already moving at speed and (b) in some cases the vehicle was brought to a halt without causing an accident and it could be argued that the driver had managed to retain control of the vehicle.
  
• Many sudden acceleration incidents reported in vehicles fitted with electronic throttles occurred at speed and therefore would be outside the NHTSA definition. e.g the incident on 28th August 2009 when a 2009 Toyota Lexus E350 driven by California Highways Patrol Officer Mark Saylor experienced a serious malfunction that caused  the vehicle to accelerate when already at road speed and which resulted in an accident in which Saylor and the three other family members travelling with him were killed.

In my opinion, this is a totally inappropriate basis for the diagnosis of intermittent electronic malfunctions. I know of no similar assumption being made anywhere in the electrical and electronic engineering industries as part of a fault diagnosis strategy.  I know of no electrical engineer or technician cognisant in any degree with electronic systems who would dream of assuming that if he or she could not reproduce an intermittent failure therefore it could never have occurred.

This  particular assumption,  as far as sudden acceleration diagnosis goes, has the effect of re-classifying any  intermittent electronic faults that fail to reproduce as  proven not to exist and, ergo, by fallacious argument, as an instance of driver malfunction or "pedal error". 

Since any sudden acceleration incidents that could not be reproduced either through in-vehicle tests or laboratory bench tests were deemed by NHTSA not to be due to intermittent electronic failures, the following conclusion in the 1989 Sudden Acceleration Report is scarcely surprising:

“For SAI in which there is no evidence of throttle sticking or cruise-control malfunction, the inescapable conclusion is that these definitely involve the driver inadvertently pressing the accelerator instead of, or in addition to the brake pedal.”

 Absence of proof is not proof of absence. In the event of failing to reproduce a sudden acceleration then the investigator must remain agnostic as to cause and certainly not start reaching "inescapable conclusions" as to the event being caused by the driver.

Since 1989 when the NHTSA report was written, sudden accelerations have continued to occur in sporadic clusters. Manufacturers have continued to shelter behind the report’s finding and blame drivers, or loose floormats, or stuck throttles, for sudden accelerations. Sudden accelerations still occur, sometimes with disastrous results. In cases where there is death or injury, by default the above "logical assumption" regarding reproducibility leads inexorably the reversal of the burden of proof, so that the vehicle, as a general rule is, presumed ‘innocent’ i.e. not to have malfunctioned, unless the driver can prove otherwise. 

 Why the NHTSA should have adopted such a limiting definition of a sudden acceleration is not at all clear.  To an outside observer, it appears that unexplained acceleration in the automobile environment is being investigated on a somewhat ad-hoc basis, with insufficient reference to known art in control engineering theory and to standard practice in control system fault diagnosis.

For example, if a competent electrical or control engineer were investigating unexplained behaviour in a P & I controller in an industrial application, they would first seek to establish whether or not the system response was within the design specification. To do this they would of course have to know the key system design parameters, which would have been provided to them, as a matter of course, by the manufacturers.  They would normally measure the step and impulse responses of the system and establish the gain/frequency response and compare these with the design responses. They would inject noise at critical points in the system. They would want to see if the output drifted over time and if any parts of the system could go into saturation and whether there were any signs of instability. They might repeat the tests with certain faults imposed on the control module. They would also look for signs of poor electrical contact, electrical tracking, hotspots, transient overvoltages and other rather mundane causes of electrical misbehaviour. They would, in other words, be carrying out a kind of sensitivity analysis to establish how near the system was to becoming unstable. They would be able to carry out this investigatory process fairly effectively because most of the relevant design information would have already been made available to the owner of the equipment as a matter of course as a very necessary part of the extent of supply.

In the automobile industry things seem to be very different. Almost all of the design information relating to electronic systems - including details of  software and its updates  - remains privy to the manufacturer. The key information, which would allow competent independent electrical or control system engineers to critique safety-critical electronic systems, is kept securely under wraps and can only be obtained by the lengthy, inefficient and costly process of legal discovery. Effectively the automobile industry is self-certifying and customers have no independent way  of assessing whether or not safety critical products meet essential requirements for functional safety either when they purchase the vehicle or thereafter.

In my opinion, this is a thoroughly unsatisfactory situation: taking the manufacturer's word for the functional safety of what are now highly complex safety-critical electronic systems is no longer acceptable. Protocols need to be put in place, as in other industries,  that allow independent evaluation of safety critical automobile systems at the design stage so that the customer can have confidence that the systems have been designed, manufactured and installed in accordance with best current functional safety practice. 

Bookmark Section 9 Links and References before leaving this site.
Section 9.5 contains links to reports of many more sudden acceleration incidents