“Question About Accuracy Of Aerial Surveillance In Court”

Hey Guys,

As a result of a recent ticket I got over the weekend, I've been reading up on Vascar / Aerial Surveillance and have been quite surprised as to the accuracy at the results by this measurement method. There seems to be a lot of room for error seeing as how the start / stop is human triggered .

From what I've heard, the OPP evaluates the time to travel a 500 m distance, from which a velocity is calculated; Please correct me if im wrong! If this is the case,

at 145 km/h it would take 12.4 seconds

at 155 km/h it would take 11.6 seconds

That's a difference of 0.8 seconds between velocities

Considering the average human reaction time is 0.3 seconds, and that the error can be repeated twice during the measurement, both at the start and stop of the measurement, wouldn't this make for a good case in court for a reduction of an infraction? After all, the average human reaction could be 0.6 seconds for the duration of the measurement, which could, in many cases, mean the difference between getting a 172 or 128 infraction.

At the above mentioned speed ranges a 0.6 s total reaction time error would equate to an approximate error of 8 km/h.

Does anyone know the success rates of negotiation with repect to Aerial tickets reduced to lesser infractions? How much can I expect to have removed if negotiate?

Thanks

Also, to calculate speed over 500m, simply divide 1800 by the time (500m x 3600 seconds/hr divided by 1km/1000m). So say for example, you time a car going 12.4 seconds between hash marks. That car is travelling at 145km/h.

Now, some will say if the average human reaction time is .3 seconds, then that could be as little as 12.1 sec. (148.7km/hr) or as high as 12.7 sec (141.7km/hr) Meaning a 7 km/hr spread.

But I don't think that's an accurate way to look at it either. (I may be wrong) but a person's reaction time (braking for a vehicle that suddenly stops in front) varies immensely compared for an actual known reaction time (known stop approaching) itself.

check your reaction time here: http://www.humanbenchmark.com/tests/...time/index.php. I averaged 220 milliseconds.

On the reaction time test above, I averaged .220, (190 to 230) a spread of about .04 seconds. So with that the 12.4 seconds can mean a spread of 12.36 sec (145.63km/hr) to 12.44 sec, (144.69km/hr)spread of less than 1km/h (for the sake of argument). That doesn't seem that inaccurate at all

Having said all that, the above internet test does not factor in a huge missing piece of this calculation – one can not predict when the green box will "pop-up", therefore there is a delay. The ability to see the target "hash mark" would enable the reaction time to be that much faster.

Now the benefit to the driver is:

1) that every speed is always rounded down and the speed is an average over a complete 500m.

2) the speed is an average over 500m, therefore the driver was going AT LEAST the speed shown for 50% of the time/distance.

On the reaction time test above, I averaged .220, (190 to 230) a spread of about .04 seconds. So with that the 12.4 seconds can mean a spread of 12.36 sec (145.63km/hr) to 12.44 sec, (144.69km/hr)spread of less than 1km/h (for the sake of argument). That doesn't seem that inaccurate at all

The spread was .04 but the actual was still .2 you have to factor that in too.

Statistical meanings:

Repeatability: same person same task same outcome

Reproducability: different person same task same outcome

I would say even with what da 'bear gives us is there should be some margin of error given on these tickets, more so then radar/lidar/pacing.

Reflections wrote:I would say even with what da 'bear gives us is there should be some margin of error given on these tickets, more so then radar/lidar/pacing.

there is already margin given....

Now the benefit to the driver is:

1) that every speed is always rounded down and the speed is an average over a complete 500m.

2) the speed is an average over 500m, therefore the driver was going AT LEAST the speed shown for 50% of the time/distance.

Where as radar/lidar are a precise moment in time, not an average over 500m

I'm confused about the reaction time. Specifically the 'spread' (.3 seconds on both ends).

{hash} (.3 sec) [START] -------- 500m -------- {hash} (.3 sec) [STOP]

If the officer presses the start button .3 seconds AFTER the start hash marks (or whatever those lines are called) - then he would have .3 seconds after the stop has mark as well. so 500m = 500m.

The only thing would be human error... he started the timer after you passed the hash mark and then anticipated you reaching the hash mark and stopped before you actually passed it. Therefore Actual distance is less then 500m = faster speed traveled.

Your assessment is correct. Now make the officer say it on the stand.

the 0.3 is inaccurate as this is human reaction time to a "unknown".

As I have posted a prior link to a website that will give you a reaction time in "milliseconds", not 10ths of a second (ie: 0.3). However the web link is also flawed somewhat. There is a "known" change of screen coming, but uncertain when that known is coming.

Now take the same test with 2 known items.....you know the hash mark is coming and you know exactly when. I'm 100% postive I would be under 100 milliseconds with both knowns. So double it for both ends of a hash mark area and still 200 milliseconds.

It is still an average over a complete 500m, with the average being a speed of "X", which is the mid-point of speed travelled during that 500m. Which means the person has been over the speed of "X" and under during that 500m, so is in fact giving a lower speed than the highest reached during the 500m. Incidently the speed is rounded down, giving the motorist more benefit.

Here's MY experience with the plane and the plea I took. I really wanted to challenge the plane-guy... even had a full defense planned out, but the plea was really reasonable.