Unfortunately it's for real [LWN.net]

Unfortunately it's for real

Posted Jun 10, 2008 2:09 UTC (Tue) by pr1268 (subscriber, #24648) [Link]

First, it wouldn't be a policeman ticketing the driver--instead the ~~driver~~ primary vehicle registrant is mailed a citation.

Second, even applying Cauchy's mean value theorem, it would be conceivable that a law enforcement officer could prove one was speeding: (1) the distance between two toll stations can be demonstrated to a judge and jury beyond a reasonable doubt, (2) the toll booth records Bruce mentions above could convince said judge and jury that a $YEAR $COLOR $MAKE $BODY_TYPE with tag number $LICENSE_PLATE (i.e., 2003 Blue Ford sedan with license plate "ABC-123") passed toll booth #1 at $TIME1 on $DATE and toll booth #2 at $TIME2 on (same) $DATE (with video camera footage to boot), and then (3) mathematically demonstrate that this vehicle's average speed was $DISTANCE divided by ($TIME2 - $TIME1), some value substantially greater than $SPEED_LIMIT, and thus the driver would be cited for driving at excessive speed.

Granted, I am not a lawyer, but if I understand correctly, U.S. courts of law need human witnesses to verify that the primary registrant of the vehicle was indeed driving the car. (Correct me if I'm wrong, please!) I suppose that a viable defense might be, "well, your honor, I was not driving my car at that time on that date..."

I don't mean to argue with you; perhaps the police officer would NOT want to go to that trouble. It's just that it's conceivable, that's all. It's still interesting to consider how much we're being watched and don't even realize it.

Unfortunately it's for real

Posted Jun 10, 2008 15:09 UTC (Tue) by hannada (subscriber, #4633) [Link]

In Colorado we've "solved" the problem of proving that any particular person is driving a
given car.  Automatically generated citations are now issued against the automobile instead of
against the driver (much as has always been the case with parking tickets).  The offense is
that the car was moving too quickly through traffic, not that a particular person was guilty
of driving too fast.  The main consequence is that, in Colorado, you won't get points against
your driver's license for these types of offenses, but the fine is still levied (against the
registered owner of the vehicle).

Unfortunately it's for real

Posted Jun 10, 2008 17:48 UTC (Tue) by oak (guest, #2786) [Link]

Of course they also have to prove that the clocks at both toll stations 
are in sync...

Toll booth clocks

Posted Jun 12, 2008 6:25 UTC (Thu) by man_ls (subscriber, #15091) [Link]

I believe keeping the clocks in sync is the worst practical problem. Radar detectors have to be certified by an authority that they are in working order. Certifying that toll booth clocks are accurate to a certain extent (in the order of tens of seconds) could be quite expensive. Regular use of e.g. NTP would synchronize the clock, but it would not certify it.

Toll booth clocks

Posted Jun 12, 2008 19:20 UTC (Thu) by aegl (subscriber, #37581) [Link]

My son got a ticket from a traffic cop with a radar gun. Two days later he got a letter in the mail saying that the ticket had been canceled because the radar gun used had failed its end-of-day accuracy check.

I think that it would be relatively simple to certify that the clocks are synced in a way that would satisfy a court. Just have a cop visit each machine at the beginning and end of each day, check each toll booth clock against a reference clock and then sign-off that the clocks showed the same time. If someone fights the ticket on the grounds that the clocks were not accurate, then the cop gets hauled into court to testify that (s)he personally witnessed that the clocks were working correctly on the date of the alleged offense.

Toll booth clocks

Posted Jun 12, 2008 20:29 UTC (Thu) by man_ls (subscriber, #15091) [Link]

A bit too labor-intensive, and subject to manipulation. An automatic mechanism would be more efficient... like an NTP log showing how automatic synchronizations did not drift more than a few milliseconds apart. But you probably don't want the defense to make you bring an expert witness every time someone fights a ticket -- hence some kind of certification process would be a good idea.

Too much effort, maybe. This mechanism would only catch a percentage of speedy drivers anyway: those who only speed occasionally would come out clean. Radars (and their instant speed measurement) seem to be more immediate and practical, even if their inner workings are wonderfully complex.

Unfortunately it's for real

Posted Jun 10, 2008 3:36 UTC (Tue) by AJWM (guest, #15888) [Link]

> But I'm not sure the policeman who ticketed you can prove it in front of the judge :-)

He just uses Rolle's theorem.

Believe it or not, way back in college calculus class (mid-1970s, which dates me), we actually
had a movie in class about this -- a state trooper ticketing a guy for speeding between toll
stations because of the too-high average speed, and then the cop goes on to explain Rolle's
theorum.

Hokey, yeah, but I still remember it.  (That and the one in biology class where a
football-field full of students act out DNA transcription.  Weird.)