Could you get flashed from a standing start?
Discussion
I don't know why I thought of this! But can anyone think how fast a car (bike, whatever) you'd need to:
stop right next to a 30mph speed camera, then accelerate from a standstill and get flashed by it?
Silly question I know, but I'm sure one of you maths/physics types can come up with the answer!
Thanks in advance,
Shentodj
stop right next to a 30mph speed camera, then accelerate from a standstill and get flashed by it?
Silly question I know, but I'm sure one of you maths/physics types can come up with the answer!
Thanks in advance,
Shentodj
BO55 VXR said:
Aren't gatso's triggered from an embedded coil ... the "ladder" is just used as a reference to check the speed
The scamera that uses an embedded coil is a Truvelo (front-facing). GATSO's use radar and take you from behind
Yes the ladder is for corroboration. Not sure what it would mean if one of the photos showed the car out of range of the ladder, though...
If you take a car that can do 0-60mph in 5.0 secs, then assuming the acceleration is linear (which it almost certainly isn't as the car will accelerate much quicker 0-30 than 30-60 but for this exercise let's assume it is linear) then the speed will be
v = 7.32 x squareroot (s)
where v is the speed in mph and s is the distance travelled in metres
So after
10m speed = 23.14mph
20m speed = 32.73mph
30m speed = 40.09mph
Just to highlight the linear bit if the Supra above kept going at that rate it would hit 60mph in 3.77 secs.
v = 7.32 x squareroot (s)
where v is the speed in mph and s is the distance travelled in metres
So after
10m speed = 23.14mph
20m speed = 32.73mph
30m speed = 40.09mph
Just to highlight the linear bit if the Supra above kept going at that rate it would hit 60mph in 3.77 secs.
Edited by over_the_hill on Friday 21st July 15:22
If a car were to accelerate as such a stupendous rate past a gatso camera (uses doppler shift of radar), there is a good chance the algorithm processing the speed data would throw out the measurement because there might be too much variance of speed. Also, the speed derived from the corroborating photos would not match the speed from the sensors; this may or may not save you.
Truvelos use pressure strips as the primary sensor, not coils. In this case, the car might not be within the 3 lines in the photo, therefore the prosecution cannot proceed.
Truvelos use pressure strips as the primary sensor, not coils. In this case, the car might not be within the 3 lines in the photo, therefore the prosecution cannot proceed.
Smeggy, I assume that the Doppler shift measurement does not use anything sophisticated in that way. A vibrating tailgate on a van can reportedly record a higher speed than the van. I expect that the high frequency vibration reflection signal will superimpose a higher peak shift onto a signal that must be dominated by the principal shift due to vehicle speed.
Richard C said:
Smeggy, I assume that the Doppler shift measurement does not use anything sophisticated in that way. A vibrating tailgate on a van can reportedly record a higher speed than the van. I expect that the high frequency vibration reflection signal will superimpose a higher peak shift onto a signal that must be dominated by the principal shift due to vehicle speed.
I know gatso take hundreds of readings a second so they must do something fairly smartish.
The signal bouncing back to the gatso will have lots of shifts due to: other vehicles, vibrating surfaces, rotating objects, cosine effects, etc. If vibrating surfaces are indeed fooling the sensor then I can only assume the peak shift is checked for consistency and used for the speed measurement (don’t quote me on that).
I don't know what frequency gatso work at, but it must be fairly high to get the required range accuracy. Therefore any doppler shift imparted by a moving car will be fairly small. High speed vibrations and whatnot can easily be filtered out with velocity gates. Amplitude variations caused by vibration can be averaged out with integrators over a number of return pulses, whereas angular variations do not have a significant effect at such short ranges.
geoff82 said:
High speed vibrations and whatnot can easily be filtered out with velocity gates.
Indeed so - but where to set the thresholds ?!?
geoff82 said:
Amplitude variations caused by vibration can be averaged out with integrators over a number of return pulses,
Again yes, but are they so? If the ‘shutter effect’ is fact then I would tend toward a negative answer.
Masking out (by averaging) a characteristic which would actually indicate a problem is phenomenally bad practice (even the LTI2020 doesn’t do that).
geoff82 said:
whereas angular variations do not have a significant effect at such short ranges.
Short ranges would exacerbate the problem as the beam would be more likely to fire at a steep angle (less redshift from a given target where the velocity approaching perpendicular). Given the inevitable spread of shift, which reflection would you accept?
Stubby Pete said:
Interesting reading chaps.
If it was possible to isolate the frequency used by the Gatso, surely it would be possible to produce something similar from a car to negate its effect?
If it was possible to isolate the frequency used by the Gatso, surely it would be possible to produce something similar from a car to negate its effect?
I don’t see why that wouldn’t be possible, assuming the incident beam amplitude isn’t cunningly modulated.
smeggy said:
geoff82 said:
High speed vibrations and whatnot can easily be filtered out with velocity gates.
Indeed so - but where to set the thresholds ?!?
geoff82 said:
Amplitude variations caused by vibration can be averaged out with integrators over a number of return pulses,
Again yes, but are they so? If the ‘shutter effect’ is fact then I would tend toward a negative answer.
Masking out (by averaging) a characteristic which would actually indicate a problem is phenomenally bad practice (even the LTI2020 doesn’t do that).
geoff82 said:
whereas angular variations do not have a significant effect at such short ranges.
Short ranges would exacerbate the problem as the beam would be more likely to fire at a steep angle (less redshift from a given target where the velocity approaching perpendicular). Given the inevitable spread of shift, which reflection would you accept?
Where to set the thresholds? I believe gatsos only work up to 167 mph. Even if vibrating parts of the car were causing large doppler returns, the very large return of the car itself would mean that the "fake" doppler would be ignored, or even completely masked.
Averaging out for amplitude variations is common practice in many radars. All you are doing is using an m of n count. I.e. If you receive 7 out of 12 pulses in any sector (for example), then the target is considered genuine. This is actually considered a pretty accurate method of accounting for amplitude variations in echo returns. I didn't mean an averaging out of the doppler frequency: i meant in a noise cancellation sense of the term.
I'm not sure what you are referring to with the angular variation? I meant a phase shift of the return wavefront caused by vibration, which does not modify frequency, but causes an apparent bearing error. (I.e. the car looks like it is on a different part of the road).
Edited by geoff82 on Monday 24th July 17:37
geoff82 said:
If i knew what mechanism the gatso used (I.e. pulse doppler or CW) i would be able to give more definitive answers.
Unfortunately I can’t help you there.
geoff82 said:
Where to set the thresholds? I believe gatsos only work up to 167 mph.
That’s a wide window
Simple mathematics indicate that vibrations won’t account for a significant amount of that. geoff82 said:
Even if vibrating parts of the car were causing large doppler returns, the very large return of the car itself would mean that the "fake" doppler would be ignored, or even completely masked.
I would agree except that a phenomenon called ‘shutter effect’ has been accepted to exist (Richard C’s prior post)
geoff82 said:
Averaging out for amplitude variations is common practice in many radars. All you are doing is using an m of n count. I.e. If you receive 7 out of 12 pulses in any sector (for example), then the target is considered genuine.
That sounds more like variance checking rather than averaging. Besides, I’m wasn’t discussing amplitude variations.
geoff82 said:
I'm not sure what you are referring to with the angular variation? I meant a phase shift of the return wavefront caused by vibration, which does not modify frequency, but causes an apparent bearing error. (I.e. the car looks like it is on a different part of the road).
OK, crossed wires! My original comment was regarding the fact that the radar beam is not parallel/perpendicular to the various vehicle surfaces, hence an unpredictable spread of shift could result.
I fear we are turning this into an Überthread
Gassing Station | Speed, Plod & the Law | Top of Page | What's New | My Stuff