Driver who killed woman denies mistaking accelerator for brake
2 February 2017
Ann Diggles told Preston Crown Court that her car would not stop before hitting a pedestrian
'Lengthy driving career'
Brake 'not working'
If you take a pedal sensor apart you'll see that is exactly how they are designed, there are multiple resistance tracks rather than just one. Also designed so that a break in a track or failure of the wiping contact results in a safe failure mode. The angle sensor at the throttle end uses a similar scheme.
He didnt fully discount that it was possible
The problem is that if the software has crashed - how do you know whats going to be recorded?
My gut instinct is that the dear old lady has probably by the law of averages got the wrong pedal but firmly believes she
Having said that, how many times have I had to restart a pic or iPhone.
In my previous job I wrote software for a specific air traffic control system. The software had been in use worldwide for
some 20 years. One day in a simulator running regression testing on a new modification there was a major safety error in
the software found, nothing to do with the new change but in the original baseline software. Admittedly the scenario and
user/operator switch and key actions requires to trigger the error were very specific and the chances of it happening
millions to one but a student controller during a free play session happened to trigger it. He argued that it had happened
we all said no you must have imagined it. In the end we replayed a recording of the test session. He was right we were
wrong and all resources were dropped on everything to fix this problem.
My long winded point is that even when someone says it's impossible, there's always that one in a million sequence of
things that can cause unexpected errors.