Help - Search - Members - Calendar
Full Version: A Review Of Laser Speed Meters
FightBack Forums > Queries > Technical Discussion of Enforcement Devices
Pages: 1, 2
Paul D. Lee
For ease of reference, here is the latest copy of my Laser Speed Meters Report.

Click to view attachment

Click to view attachment
Click to view attachment
Click to view attachment

(Edited 15 July 2008, with latest version of report, and example "Beam Spread" Clips)
cop a load of this



Hi paul D,

Thank you very much for supplying your information, which makes very interestin reading,
I am only half way thro your post. and will be taking some pointers when I appeal my case,

I would not bother if I had been speeding, but I am positive that I wasnt, and I really want to persue this, and your input is much appreciated.

Thank You
BetaPilot
That is an excellent and fascinating report Paul. As a fellow engineer I can see clearly that the reasons for all your findings all appear solid.

Have you lodged that report with the relevant authorities? Is there a body who can take up your case and get these speed meters re-tested for type approval?

What's being done about this?

Dan.
potkettleblack
QUOTE (BetaPilot @ Sat, 27 Jun 2009 - 16:50) *
Have you lodged that report with the relevant authorities? Is there a body who can take up your case and get these speed meters re-tested for type approval?

The official bodies that should care, have it, and don't. The Home Office Scientific Development Branch clearly didn't test it properly so they don't want to admit that. The Home Office under FOI settled on responses such as "nothing more we can say". ACPO had its flaws demonstrated infront of them (some contribution is in the report) but the action they took was to deny the knowledge, including the fact that they have their own report AAIU. CPS and SCP have no interest in proving any failings. The Transport minister doesn't understand his job, puts out spin via flawed statistics and proposes plans for more speed enforcement. And no one holds onto Secretary of State long enough to read any report. Although the Gov't ignores its citizens, it would normally repond to the BBC and the Daily Mail but that went on in 2007 and still the result was denial.

Feel free to lobby your MP.
contester
What is the situation, if you disagree with the reading on the laser equipment as above but you feel you were still above the prescribed limit? Can you plead not guilty to the specified speed?
Tancred
I believe what you are looking for is a Newton Hearing.
potkettleblack
The offence is exceeding the regulatory speed. The degree to which you exceeded it determines the sentence (fines/points). The moral conflict is that DfT figures show half the motorists on the road every day exceed the posted limit and since the UK is still within the top 10 for world road safety clearly this is safe driver behaviour. Search for Sentencing Guidelines and decide whether you want to take the fixed penalty offer or gamble on the higher limit applicable if you fight the case, which I think might only be lower than the FPN if you convince the court that you were under the speed limit in the first place ie: no penalty at all.

I advise you to complete the NIP wizard with a new thread.
Laura Bidding


It was only having gone back to the station where I was based in the camera unit to pick up some papers for court, that I happened upon a hard copy of Paul’s review. I remember at the time, it was floating about for months in various vehicles and I meant to comment then but never actually did.

I thought I’d highlight a few points linked to this quote from higher in the thread;

‘Have you lodged that report with the relevant authorities? Is there a body who can take up your case and get these speed meters re-tested for type approval?’

[/size]

[size="3"]The most crucial question that I cannot find answered by the review, is the software in use by the speedscope in use at the time of the report?


This will mean nothing to most here, but it is most apparent by the age of the report that things have moved on dramatically since the report was written. Although, the review can be updated (and has been), unless the actual model tested is at least fairly current, with the most recent software (v3.03 is the latest I think, but could be wrong on that bit), then each subsequent model of speedscope has been ‘back’ for testing before being approved for operational use. My unit waited for a whole year for the latest Ultralyte to go through type approval, so they don’t just ship ‘em over and knock ‘em out, each subsequent model is different to the previous and the whole type approval process starts again. All subsequent models aren’t automatically approved just because the original model gained approved status. Every/any enhancement resets the counter.

The point here being, that whilst I have no problem with anything that improves road safety and the concept of the report, to hold it aloft based upon an elderly speedscope and unknown software (what was it, v1042, earlier?), against greatly enhanced models with more complex metering patterns/multi-point confirmation, just isn’t going to hold the same amount of water.

Things move on so quickly technologically speaking, that it is only going to be relevant if current models and versions are actually reviewed. There might have been a brief snapshot in time when Paul’s review was relative to an unknown percentage of models in use, but now? Unquestionably not. If he continually got access to the latest model, with the latest software and ran the same tests, then it would have a lot more credence and carry a lot more weight. Unfortunately, other than to stir up the masses who base their ‘expert’ opinion on such works, it is completely outdated and subsequently pretty valueless as an example of the state of current operational equipment.

bama
From a software engineering point of view the 'latest and therefor better' stance does not hold water.
Mortimer
Quite so Bama. I work in a company that produces systems, both the hardware and the software. Hardware revisions and updates receive far more analysis and testing that software updates do. Why? Because the physical mechanics of a device are more often that not fundamental to the way something works, and the software is simply a means to gather and display the results. I would not expect the software in any speed-gun to be anything like sophisticated enough to overcome the fundamental issues of beam spread, cross-hair alignment and slip. The processing required to deal with those problems would require considerably complicated image analysis by the software, and not problems any software house would find a walk in the park.

How is a computer programmed to recognise that the beam pattern is possible overlapping two vehicles?
How is a computer programmed to recognise that the visual cross hair is incorrectly aligned with the beam?
How is a computer programmed to recognise that the beam has slipped against the target surface?

How is a computer powerful enough to do all these things and more going be crammed into a handheld device?

Nope sorry, irrespective of how good and up-to-date the software might have become, I can't see it negating all of the problems identified in the report.
Laura Bidding


Thanks for replying guys, but I am genuinely at a loss to identify your intentions?

Type approval has been granted thus far on all chronologically sequenced LTI’s and subsequent software updates, yet you voice your obvious agreement that Paul’s review somehow ‘blanket covers’ all subsequent model and software changes?

Mortimer even attempts a bit of the old credential flourishing in an attempt at endorsement enhancement!

I particularly liked these two questions which assigned the LTI with additional responsibilities;

‘How is a computer programmed to recognise that the visual cross hair is incorrectly aligned with the beam?’

It isn’t. The crosshairs are part of the video camera and nothing whatsoever to do with the speedscope. They are moved by nothing more space age than mechanical thumbwheels situated on the video camera, so;

‘How is a computer powerful enough to do all these things and more going be crammed into a handheld device?’

It doesn’t assign any processing power to ‘crosshair recognition’, so can assign everything to the job in hand.



Having achieved type approval status, the manufacturer and/or supplier doesn’t now need to prove anything. So can the review demonstrate that the latest versions/models exhibit the same characteristics claimed about the originals? The answer is clearly and blatantly no.

From the perspective where the review can do any ‘good’, influence operational use and/or convince anyone who is in a position to make those changes happen, it is unquestionably not current enough. Every week that passes, compounds the problem even more. Until that situation changes, my previous comments will stand.

Mortimer
The point of the report is to highlight some fundamental features of the measuring equipment that affect its accuracy under certain circumstances and conditions. The report suggests that the testing used to give these devices type approval is flawed, leading to the belief that the devices give infallible results, and therefore cannot be questioned in court.

I do not wish to reveal what I do, as I prefer to remain anonymous to a degree, but I have significant experience of working in the design, manufacture, programming, testing, integration, regression testing, proving and so on, of critical systems, the sort of things that if we get it wrong lots of people get hurt. So I'm familiar with extremely rigorous testing and proving regimes that would probably put the so called 'Type Approval' testing of these speed meters to shame.

The point of the report as I see it, is to plant the seed of knowledge that these tools are not infallible, and are subject to errors in use that can result in injustice. Therefore, either the testing needs to be redone, and type approval status revoke if the flaws are confirmed, or serious testing and study carried out to identify changes to the rules and guidance of there use, so that injustice cannot occur.

As to your attempt to understand the question I raised: Don't bother. They were hypothetical. Those are the sort of software improvements that would be required to negate the inherent mechanical flaws in the hardware. Unless the software updates can do those things, then ANY software update released since the report was written stands no chance in resolving the issues raised by the report.

The biggest question resulting from the report still remains:

How did these devices get type approval in the first place, with so many inherent flaws?
bargepole
I think the point here is that the original Type Approval, carried out by the Home Office scientific branch, was not sufficiently thorough in taking account of all possible operational conditions. This is borne out by the fact that in nearly half of the States of the USA, where these devices are manufactured, it is not recognised as a reliable evidential device, whilst in other States the permissible range is far shorter than in the UK.

Whether the shortcomings still apply to later models / updated software versions is, as you say, impossible to determine without retesting those models.

However, in practical terms, it is type approved for prosecution purposes, and any defendant wishing to challenge the speed reading will have to rely on establishing from the tape that the reading is wrong in their particular case, rather than trying to discredit the LTi per se.

potkettleblack
LB, two questions:

1) Are you saying that your existing equipment has had an update and so is now better (eg: LTI TS/M is now V3.03?), or that newer equipment with later software is better?

2) Since the Type Approval testing has not changed, do you have any basis for saying that newer equipment would not fail under the same circumstances?
bama
The Type Approval testing forms the "Requirements Baseline" (software techno talk, sorry. Look up the V model if you are interested) which as you say has not changed.
So there is no requirement to change the solution at all IF the original solution actually mapped correctly to the requirements. (Apart from any hardware obsolescence, which of course should have been accounted for in the original design).

Of course a 'correct' solution would mean no nice new models needed to sell to the Bib.
Laura Bidding


I will have to make this my final word on the matter as I have other commitments over my leave period;
Never have I suggested that each LTI software/hardware change is better than the last – more, the pertinent phrase would be ‘different’. I am also not championing it as a speed enforcement tool, merely saying that if someone goes to the trouble of producing and publishing any form of ‘expose’, it needs to be current to be taken seriously.


Would a report on the security shortcomings of modern operating systems produced today be relevant if the software used for the research was Windows 95?
Has my existing equipment had an update?
In the years I was involved in speed enforcement (I am no longer), my unit had four models of LTI, each with a different software version. Every model was designed to give a speed reading but each progressive model also brought with it other ‘benefits’. These included less weight due to smaller components, different power options, digital compatibility which means far less storage space needed as video tapes not used, plus reduced viewing time when searching the resulting DVD. There were also software developments which dramatically improved the system’s abilities in poor weather conditions such as rain, something not available on early models/software versions.
A quick word on the LTI in the rain. It abilities depend on five key things;
* The model of LTI in use.
* Scope to target distance.
* How heavy the rain is.
* Direction of travel of target vehicle (those traveling away chuck up the equivalent of ‘chaff’ in the form of spray from the tyres, much like aircraft eject counter-measures to confuse incoming guided missiles).
* Speed of target vehicle (faster vehicles generate more spray).

As someone who used an LTI for the purposes of detecting speeding offences, I have read everything I can find on the supposed problems associated with such equipment. Consequently, I have tried many times to duplicate ‘slip’ type effects, but every time the scope just threw an error back. I tried bouncing the laser off mirrors, signs, the roadway – just an error, no speed reading. Every time it worked exactly as it should, if you asked it do something it couldn’t deal with = error message. Error messages equate to the system working as it should, the scope is saying ‘You haven’t given me the right conditions to let me do my job’. For ‘error’, don’t read ‘malfunction’.


Out of many hundreds of thousands of trigger pulls and umpteen detected offences, I never had cause to doubt what was being displayed on the back of the speedscope or being recorded. If the system had given me cause to question its results, I would have aborted the check and reported the symptoms. I never had cause to do that. I have no reason to lie. I have no vested interest. I don’t owe anyone anything and have nothing to gain by fibbing.
So do I have any basis for saying that newer equipment would not fail under the same circumstances?


I have never seen it ‘fail’, or even a firsthand demonstration of it being encouraged to give erroneous readings.
I will finish by saying this;
So are the newer models of LTI ‘better’ than the previous models?
They definitely return a quicker speed reading, so from a (n ex-) user’s perspective I view that as being an improvement. I would expect such developments to take place and for anyone involved in software development to suggest ‘You should have got it ‘right’ in the first place’, is simply nuts given the way in which technology advances. Space shuttle ‘O’ rings, the shape of early Comet windows anyone?


Just comparing the domestic digital camera I bought in 1999 to the one I got this year being a case in point – both ultimately record an image, but the effortless nature of the more recent model, its size, its power requirements, its storage options, its near instantaneous focusing, makes the other seem virtually prehistoric.
However, none of that matters, I am not here to argue or debate the rights or wrongs of any aspect of the equipment in use, I am merely offering my view based upon the reactions I witnessed of people who were exposed to Paul’s report – everyone instantly dismissed it due to the age of the kit being reviewed and unknown software version.


Treat that information in whichever way you choose, but it will not alter the fact that the review was poo-poo’ed at a very rudimental level, so no-one at ‘approval’ level will give it a second glance due to what’s on the test bench. No doubt, this will generate all the usual, ‘What about…yadda-yadda?’, for which you will have to fill in the blanks yourself.

andy_foster
The problem highlighted by Paul's report is a direct result of a fundamental part of how the device was designed to work - it 'locks on' to the first signal received above a pre-set threshold. Unless this has been changed, or the beam divergence substantially reduced, the device would almost certainly produce the same results even with the new super-wizzy software.

As Toad Hall, who is apparently in charge of selecting devices for HOTA testing, was at the Elvington tests, one would assume that if they had ironed out this 'undocumented feature', he'd have made some noise about it.
However, he and his boss Med "90 in a 60 limit" Hughes seem to believe that the integrity of devices relied on in prosecutions can best be assured by threatening to slap anyone who dares to challenge them with substantial and disproportionate costs.

If you know how we can get borrow a current HOTA UK police spec device without having to get a Crown Court judge to order it, I'm sure that we could arrange another test.
bama
"Would a report on the security shortcomings of modern operating systems produced today be relevant if the software used for the research was Windows 95?"

OF what possible relevance is that ?
None.

I stand by post #15/ See the V Model.


"No one knows where the laser beam goes on its trip(s)"
Cargy
QUOTE (andy_foster @ Thu, 15 Oct 2009 - 06:33) *
The problem highlighted by Paul's report is a direct result of a fundamental part of how the device was designed to work - it 'locks on' to the first signal received above a pre-set threshold. Unless this has been changed, or the beam divergence substantially reduced, the device would almost certainly produce the same results even with the new super-wizzy software.


I believe there's an old saying that appears rather appropriate:

You can't make a silk purse from a sow's ear.
Hotel Oscar 87
QUOTE (Cargy @ Thu, 15 Oct 2009 - 10:44) *
QUOTE (andy_foster @ Thu, 15 Oct 2009 - 06:33) *
The problem highlighted by Paul's report is a direct result of a fundamental part of how the device was designed to work - it 'locks on' to the first signal received above a pre-set threshold. Unless this has been changed, or the beam divergence substantially reduced, the device would almost certainly produce the same results even with the new super-wizzy software.


I believe there's an old saying that appears rather appropriate:

You can't make a silk purse from a sow's ear.

Quite true. However, you can make one that for all the world looks like one - at a healthy distance. Provided you prevent anyone from looking at it too closely and effectively conceal the fact that whilst its silk-like it was still has porcine DNA.
Mortimer
Which is why an up-to-date model of speed measuring device is unlikely to be allowed to be scrutinised closely in the way the device in the report was. Because the porcine DNA will be revealed perhaps?
nomadros
QUOTE (Laura Bidding @ Thu, 15 Oct 2009 - 05:57) *
snip

In the years I was involved in speed enforcement (I am no longer), my unit had four models of LTI, each with a different software version. Every model was designed to give a speed reading but each progressive model also brought with it other ‘benefits’. These included less weight due to smaller components, different power options, digital compatibility which means far less storage space needed as video tapes not used, plus reduced viewing time when searching the resulting DVD. There were also software developments which dramatically improved the system’s abilities in poor weather conditions such as rain, something not available on early models/software versions.

snip


My FOI request to the HO asked if there had been any changes to any LTi model since their introduction and the reply was "none." There also had been no updates to the original TA for any model as there had been no changes. Given a TA covers hardware and software, I would have thought that the above suggests that there have been changes.
bama
The requirements have not changed.

'up to date' is spurious. IF the original solution actually worked.

In fact the updates themselves testify that the device was not 'fully functional' when released.

Software maintenance comes in three flavours - corrective, adaptive and perfective.

Read about them here http://books.google.co.uk/books?id=FXiRmlJ...ive&f=false

and see if there could be any reason for any software maintenance to the LTI IF it worked out of the trap.

the answer is no IMO.
nomadros
All I know (as a full time, seriously well paid, software developer) is that any update, for whatever reason, has to be tested and then achieve client approval before release.

In the case of a product whose software testing is done prior to sale, the client (purchaser) would in all cases test the product prior to purchase and the results of those tests would form part of that approval to purchase.

In this case, testing for statutory approval for use after purchase would mean that software version x.y has been proven for use on hardware version G.

However the "same" product with version Y.z even on hardware version G is a totally different can of worms and you cannot rely on the manufacturer's word that everything is still kosher, whether the previous version worked lilke a charm or was a bug ridden POS.

If those guns don't have the same software and/or hardware that was originally tested for approval, then they are not the same guns and don't have type approval.....IMVHO.

Just got to convince a court of that though!
bama
agreed.
SMURF POWER
QUOTE (Laura Bidding @ Thu, 15 Oct 2009 - 05:57) *
......my unit had four models of LTI, each with a different software version. ...... .[/size][/font]

To clarify would it be the 4 below
1.Marksman
2.TSM
3.100
4.1000
Or was it 4 TSM with diff' software, I suspect they were four LTI models which can all have different software which relate to the UK version of the device.


QUOTE (andy_foster @ Thu, 15 Oct 2009 - 06:33) *
The problem highlighted by Paul's report is a direct result of a fundamental part of how the device was designed to work - it 'locks on' to the first signal received above a pre-set threshold. Unless this has been changed, or the beam divergence substantially reduced, the device would almost certainly produce the same results even with the new super-wizzy software.

As Toad Hall, who is apparently in charge of selecting devices for HOTA testing, was at the Elvington tests, one would assume that if they had ironed out this 'undocumented feature', he'd have made some noise about it.
However, he and his boss Med "90 in a 60 limit" Hughes seem to believe that the integrity of devices relied on in prosecutions can best be assured by threatening to slap anyone who dares to challenge them with substantial and disproportionate costs.

If you know how we can get borrow a current HOTA UK police spec device without having to get a Crown Court judge to order it, I'm sure that we could arrange another test.


You are spot on with much of what you say… how ever when you do it again perhaps start at the distance the bike was caught and not at the 500 - 600 + metres the engineered test were conducted at. Engineering failure is very much different to reality …..

QUOTE
Out of many hundreds of thousands of trigger pulls and umpteen detected offences, I never had cause to doubt what was being displayed on the back of the speedscope or being recorded. If the system had given me cause to question its results, I would have aborted the check and reported the symptoms. I never had cause to do that. I have no reason to lie. I have no vested interest. I don’t owe anyone anything and have nothing to gain by fibbing.
So do I have any basis for saying that newer equipment would not fail under the same circumstances?
LTI officer
smeggy
I've given some thought to the issue of the speculated 'algorithm error' described in page 61 of the report.

This failure mode may seem plausible, but my own testing (with an actual TS/M, one that was still used 'in anger' well after those Elvington tests) indicated no such mechanism.

An alternative explanation may be a consequence of the well-known 'gaps' of the emitted laser beam pattern.
These gaps are expected; however, the actual gaps are variable. They are difficult to measure, so this usually remains unknown for individual LTIs.
The beam is emitted from three apertures; all are vertical, narrow strips, horizontally adjacent to each other. The resulting emitted beam is merely a projection of this arrangement, be it somewhat out of focus (as well as being proportionally larger with distance).

The attachment is an animated gif. It shows how the alternative explanation is possible.
This method of presentation was chosen to enable easy comparison and confirmation of appropriate placements and scalings; I've tried to be diligent with these. I don't claim pixel-perfect accuracy, but it is good enough for the simple explanation.
The images contained within are from several different runs, all extracted from the report.
All the images have the same (actual) LTI beam pattern superimposed onto them, showing what surfaces could be struck and missed by the laser. The vertical turquoise lines are to help indicate the boundaries of the lobes.
The alignments of the laser/sighting scope/video camera are reasonably assumed as fixed throughout.



As per the guidelines: operators and processing staff should not initiate or process cases where there are multiple vehicles within the active area; however, they're not perfect. Evidence showing multiple vehicles in close ‘line-of-sight’ proximity must be carefully scrutinised as there is evidently a significant risk of innocent drivers being prosecuted.

This is still one of the failings of the LTI, even if it is for a different reason than expected.
smeggy
I've had some questions posed to me by PM.

I will reproduce the relevant content in this thread so that everyone can digest them.

QUOTE
If your postulation is true then how do you account for this effect not showing up all the time, firstly at Elvington (it only happened once yet the tests were repeated multiple times),

It happened twice at Elvington.
My explanation shows both these events:
1) 12:21:57NR19
2) 12:27:33NR36

QUOTE
secondly during pre-post session alignment checks,

- Was it checked for?
- Do people know what to look out for?

Given the 'algorithm error' was first postulated subsequent to the Elvington testing, I would say the answer to both my questions are "no".

QUOTE
and thirdly during enforcement sessions on motorbikes?

I have repeatedly given my answer to this on the open forums. There are procedures and processes in place to prevent this type of error resulting with a prosecution. Operators should not target when multiple vehicles are in the active area; the processing staff also check for this.

So it shouldn't happen anyway. Even if it does, the case should be dropped due to the risk of error, regardless of whether a speed error is actually apparent.

Further to that are the circumstances needed to induce this error. Many other things have to come together:
- The optics must be set to have the gaps in the beams
- Targets must be distant enough for size of gaps to be significant enough
- The vehicle further behind must have highly reflective surfaces within the lobes; this is not a given.

QUOTE
At Elvington a Leica plate 10cm x 15cm was used vertically at 200m distance. Beam spread was measured at 4.3mR so we have a 10cm plate and an 86cm beam. If the beam has gaps that matter then the result should have been wee-waa-wee-waa-wee during alignment, but it was not.

How do you know it wasn't? Was it tested for? There is no mention of such testing in PDL’s report.

Don't forget, the gaps get larger the further the further away the target surface is; ideally, this would be directly proportional.
Then there is the further issue of focus, meaning the gaps will (and does) become disproportionately larger with distance – like I said in the open forums: the pattern becomes more defined with distance.

Further to that is the issue of implied intensity that I recently postulated and confirmed.
At 200m, gun could be getting enough of a signal from other parts of the bike, meaning a staggered (stretched) return, enough to fool the gun into deducing a strong signal, even when the return has actually become substantially weaker because a small part of the beam caught the edge of the plate. You will need to read my post on the matter to understand this.

I speak from first-hand, expert experience: moving the gun by hand to get the waa-wee-waa-wee-waa-wee-waa is difficult to do. Don’t forget we’re talking about pans of only 1/20th of a degree (from gap to lobe).
I have a motorised, star-tracking tripod that has a full manual mode via remote control, to allow very slow panning; this is the way to do it.

QUOTE
If there were such a figurative and literal hole in the operation, I'm pretty sure someone at LTI would have noticed this during development.

That literal hole exists, there is no (genuine) question of it.
It has been successfully explained, as well as confirmed by actual testing.

Do you think it would have made any different to LTI if they had or hadn't been aware of these lobes?

I have LTI documentation (1993) that unequivocally shows that Jeremy Dunne knows about how the beams are projected – yes through air! I’ve just now checked it again. He indeed has taken measurements of them, and his results are in total agreement to mine – a simple projection, resulting with thin strips of illumination on the target surface (assuming correct focus); there is no multi-modal 'smearing'.
Hoof_Hearted
QUOTE (smeggy @ Thu, 27 Jan 2011 - 23:31) *
I've had some questions posed to me by PM.

I will reproduce the relevant content in this thread so that everyone can digest them.

QUOTE
and thirdly during enforcement sessions on motorbikes?

I have repeatedly given my answer to this on the open forums. There are procedures and processes in place to prevent this type of error resulting with a prosecution. Operators should not target when multiple vehicles are in the active area; the processing staff also check for this.

So it shouldn't happen anyway. Even if it does, the case should be dropped due to the risk of error, regardless of whether a speed error is actually apparent
.


In principle, you are 100% correct. And when the alleged speed is recorded on video (by either a proper BiB, or a civvy SCP operator) - those checks and balances should indeed be carried out by the processing staff - and those cases be dropped. But this isn't always the case - there are many examples, on many differing forums, where drivers have been summonsed when two vehicles are clearly within the defined target area.

But the real crux of concern is when the laser readings are obtained by a Police officer, and where the laser device is NOT linked to any kind of video recording device - I'm 100% confident in my opinion that in these instances, these 'checks and balances' are NOT complied with.


QUOTE (smeggy @ Thu, 27 Jan 2011 - 23:31) *
Don't forget, the gaps get larger the further the further away the target surface is; ideally, this would be directly proportional.


You are correct that the gaps get larger as the distance increases - but you fail to mention that the 'edge defintion' of the beams becomes LESS defined as distance increases. See my next sentence for the explanation.


QUOTE (smeggy @ Thu, 27 Jan 2011 - 23:31) *
Then there is the further issue of focus, meaning the gaps will (and does) become disproportionately larger with distance – like I said in the open forums: the pattern becomes more defined with distance.


That is just plain wrong. Unless you are operating the laser in a VACUUM !!! In 'real life', air is contaminated with airbourne dust. And just as the number of oxygen (and nitrogen, and other 'gases' which make up the atmosphere) molecules will be exponentially different from say one cubic metre of air compared to 100 cubic metres - so does the amount of 'dust particles' increase. Each one of these dust particle contaminents can distort and/or deflect the laser beam - and is therefore fundamental 'science' which would dictate that the actually accuracy of the 'focus' will deteriorate with increasing distance.
justforthepictures
QUOTE (Hoof_Hearted @ Sat, 2 Apr 2011 - 14:03) *
But the real crux of concern is when the laser readings are obtained by a Police officer, and where the laser device is NOT linked to any kind of video recording device - I'm 100% confident in my opinion that in these instances, these 'checks and balances' are NOT complied with.

Just to confirm what you are saying?

You much prefer the actions of the operator (whoever that may be), to be recorded to some form of media (tape or DVD), so there is a way of verifying exactly what was done in the way of checks?

Can you list the ‘checks and balances’ you refer to that should be carried out by the processing staff?

You also say you are 100% confident that there are checks and balances that aren’t done by police officers. What are they and how you have arrived at such a stoic conclusion?

Incidentally, what is the relevance of your ‘That is just plain wrong. Unless you are operating the laser in a VACUUM !!!’

Isn’t the vacuum of space packed with floating particles? I was taught that the Earth collects about 40,000 tons (35,000 metric tons) of dust in one annual trip around the Sun.

Also, you state that ‘Each one of these dust particle contaminants can distort and/or deflect the laser beam - and is therefore fundamental 'science' which would dictate that the actually accuracy of the 'focus' will deteriorate with increasing distance.’

What particle density would it take to deflect the laser beam? Over what distance? Do you mean that the particles would form enough of a barrier to deflect it so it would actually no longer travel in a straight line?
smeggy
QUOTE (Hoof_Hearted @ Sat, 2 Apr 2011 - 13:03) *
QUOTE (smeggy @ Thu, 27 Jan 2011 - 23:31) *
Then there is the further issue of focus, meaning the gaps will (and does) become disproportionately larger with distance – like I said in the open forums: the pattern becomes more defined with distance.


That is just plain wrong. Unless you are operating the laser in a VACUUM !!! In 'real life', air is contaminated with airbourne dust. And just as the number of oxygen (and nitrogen, and other 'gases' which make up the atmosphere) molecules will be exponentially different from say one cubic metre of air compared to 100 cubic metres - so does the amount of 'dust particles' increase. Each one of these dust particle contaminents can distort and/or deflect the laser beam - and is therefore fundamental 'science' which would dictate that the actually accuracy of the 'focus' will deteriorate with increasing distance.

What I said was actually correct. The effect you mention is a technically valid one to raise, but it is merely additional and doesn’t invalidate what I said.

The issue you raise is more of an issue for shorter wavelengths (Rayleigh scattering: is why the sky is blue). The much longer 904nm is far less affected, although that is not to say there is no effect as that wavelength. However, for the effect at 904nm to be significant, that would inescapably result with the effect being much more obvious at wavelengths we can see - especially green colours, but there is no such effect there.

I have photos of various LTI beams at various distances. There isn’t significant edge spread (that could be detected) beyond what could be attributed to diffraction or the lack of focus.
The Rookie
QUOTE (Hoof_Hearted @ Sat, 2 Apr 2011 - 14:03) *
and thirdly during enforcement sessions on motorbikes? I have repeatedly given my answer to this on the open forums. There are procedures and processes in place to prevent this type of error resulting with a prosecution. Operators should not target when multiple vehicles are in the active area; the processing staff also check for this.

So it shouldn't happen anyway. Even if it does, the case should be dropped due to the risk of error, regardless of whether a speed error is actually apparent
.
In principle, you are 100% correct. And when the alleged speed is recorded on video (by either a proper BiB, or a civvy SCP operator) - those checks and balances should indeed be carried out by the processing staff - and those cases be dropped. But this isn't always the case - there are many examples, on many differing forums, where drivers have been summonsed when two vehicles are clearly within the defined target area.

Disagree, its the operators responsibility to be sure there is suitable targetting, and we rarely see video or still were another car is in the 'target area (by which I read as the area covered by the laser, not jsut the area in video shot.

It may be the case with unmanned radar devices, but not with laser.

Simon
justforthepictures
QUOTE (The Rookie @ Sun, 3 Apr 2011 - 21:33) *
QUOTE (Hoof_Hearted @ Sat, 2 Apr 2011 - 14:03) *
and thirdly during enforcement sessions on motorbikes? I have repeatedly given my answer to this on the open forums. There are procedures and processes in place to prevent this type of error resulting with a prosecution. Operators should not target when multiple vehicles are in the active area; the processing staff also check for this.

So it shouldn't happen anyway. Even if it does, the case should be dropped due to the risk of error, regardless of whether a speed error is actually apparent
.
In principle, you are 100% correct. And when the alleged speed is recorded on video (by either a proper BiB, or a civvy SCP operator) - those checks and balances should indeed be carried out by the processing staff - and those cases be dropped. But this isn't always the case - there are many examples, on many differing forums, where drivers have been summonsed when two vehicles are clearly within the defined target area.

Disagree, its the operators responsibility to be sure there is suitable targetting, and we rarely see video or still were another car is in the 'target area (by which I read as the area covered by the laser, not jsut the area in video shot.

It may be the case with unmanned radar devices, but not with laser.

Simon

I will add to that;

There will be instances where although the target vehicle was isolated at the point of pulling the trigger, the chronological lag before the asterix appears can often mean that another vehicle may then encroach into the 'offence shot', i.e. the moment when the speed reading has appeared.

This is relatively common with 'crossover', when detecting across opposing lines or multi-carriageways of uni-directional traffic. Whilst it is the person behind the LTI who dictates what is recorded, it isn't always possible to obtain 100% clean evidence due to unanticipated vehicular interaction.

Although processing unit viewing teams are usually very strict in what they allow through, it isn't unknown for the resulting witness statements that need your signature to include a copy of the offence shot where another vehicle has crept into the field of view.

Now of course, this would have had no bearing on the speed of target vehicle when measured, but it does encourage debate as to the uncertainty of the accuracy of the ping. Therefore. no signature would be added, a note attached saying you aren't prepared to defend this and the case would be dropped.

Obviously, statement signing is some way down the line, but very few questionable pings make it as far as court, because they are filtered out at some stage long before making it that far.
smeggy
QUOTE (justforthepictures @ Mon, 4 Apr 2011 - 16:28) *
QUOTE (The Rookie @ Sun, 3 Apr 2011 - 21:33) *

QUOTE (Hoof_Hearted @ Sat, 2 Apr 2011 - 14:03) *

... And when the alleged speed is recorded on video (by either a proper BiB, or a civvy SCP operator) - those checks and balances should indeed be carried out by the processing staff - and those cases be dropped. But this isn't always the case - there are many examples, on many differing forums, where drivers have been summonsed when two vehicles are clearly within the defined target area.

Disagree, its the operators responsibility to be sure there is suitable targetting, and we rarely see video or still were another car is in the 'target area (by which I read as the area covered by the laser, not jsut the area in video shot.

It may be the case with unmanned radar devices, but not with laser.

Simon

I will add to that;

There will be instances where although the target vehicle was isolated at the point of pulling the trigger, the chronological lag before the asterix appears can often mean that another vehicle may then encroach into the 'offence shot', i.e. the moment when the speed reading has appeared.

This is relatively common with 'crossover', when detecting across opposing lines or multi-carriageways of uni-directional traffic. Whilst it is the person behind the LTI who dictates what is recorded, it isn't always possible to obtain 100% clean evidence due to unanticipated vehicular interaction.

Although processing unit viewing teams are usually very strict in what they allow through, it isn't unknown for the resulting witness statements that need your signature to include a copy of the offence shot where another vehicle has crept into the field of view.

Now of course, this would have had no bearing on the speed of target vehicle when measured, but it does encourage debate as to the uncertainty of the accuracy of the ping. Therefore. no signature would be added, a note attached saying you aren't prepared to defend this and the case would be dropped.

Obviously, statement signing is some way down the line, but very few questionable pings make it as far as court, because they are filtered out at some stage long before making it that far.

An intelligent response.

If I may add to it:

At close ranges: there isn’t much of an issue because the size of the laser dot (at the target distance) is small, so it likely won't hit multiple vehicles (excluding secondary bounces).

When enforcing over large distances: the camera set up will be very telephoto (400mm sitting on a 6x crop sensor = 2400mm). It will be difficult to align the camera to set in perfect alignment with the laser i.e. with the centres overlapping.
Therefore, seeing two vehicles within the crosshairs doesn’t mean the laser actually hit both vehicles; that apparent error could merely be a result of the camera being only just within alignment.

The issue of camera alignment has always been a contentious one, and has previously cast reasonably grave doubts on speculations of various “false readings”.
norahl
QUOTE (justforthepictures @ Mon, 4 Apr 2011 - 17:28) *
... the chronological lag before the asterix appears can often mean that another vehicle may then encroach into the 'offence shot', i.e. the moment when the speed reading has appeared...

...and what would that lag be?
Jimzzr
QUOTE (smeggy @ Wed, 6 Apr 2011 - 16:26) *
QUOTE (justforthepictures @ Mon, 4 Apr 2011 - 16:28) *
QUOTE (The Rookie @ Sun, 3 Apr 2011 - 21:33) *

QUOTE (Hoof_Hearted @ Sat, 2 Apr 2011 - 14:03) *

... And when the alleged speed is recorded on video (by either a proper BiB, or a civvy SCP operator) - those checks and balances should indeed be carried out by the processing staff - and those cases be dropped. But this isn't always the case - there are many examples, on many differing forums, where drivers have been summonsed when two vehicles are clearly within the defined target area.

Disagree, its the operators responsibility to be sure there is suitable targetting, and we rarely see video or still were another car is in the 'target area (by which I read as the area covered by the laser, not jsut the area in video shot.

It may be the case with unmanned radar devices, but not with laser.

Simon

I will add to that;

There will be instances where although the target vehicle was isolated at the point of pulling the trigger, the chronological lag before the asterix appears can often mean that another vehicle may then encroach into the 'offence shot', i.e. the moment when the speed reading has appeared.

This is relatively common with 'crossover', when detecting across opposing lines or multi-carriageways of uni-directional traffic. Whilst it is the person behind the LTI who dictates what is recorded, it isn't always possible to obtain 100% clean evidence due to unanticipated vehicular interaction.

Although processing unit viewing teams are usually very strict in what they allow through, it isn't unknown for the resulting witness statements that need your signature to include a copy of the offence shot where another vehicle has crept into the field of view.

Now of course, this would have had no bearing on the speed of target vehicle when measured, but it does encourage debate as to the uncertainty of the accuracy of the ping. Therefore. no signature would be added, a note attached saying you aren't prepared to defend this and the case would be dropped.

Obviously, statement signing is some way down the line, but very few questionable pings make it as far as court, because they are filtered out at some stage long before making it that far.

An intelligent response.

If I may add to it:

At close ranges: there isn’t much of an issue because the size of the laser dot (at the target distance) is small, so it likely won't hit multiple vehicles (excluding secondary bounces).

When enforcing over large distances: the camera set up will be very telephoto (400mm sitting on a 6x crop sensor = 2400mm). It will be difficult to align the camera to set in perfect alignment with the laser i.e. with the centres overlapping.
Therefore, seeing two vehicles within the crosshairs doesn’t mean the laser actually hit both vehicles; that apparent error could merely be a result of the camera being only just within alignment.

The issue of camera alignment has always been a contentious one, and has previously cast reasonably grave doubts on speculations of various “false readings”.


As an off the top of my head remark has anyone given any thought to the possibility of beam stearing due to changes in refractive index of 'air' (say crossing over a heated patch of black tarmac, when remainder of road is concrete) given that optical paths can be up to 1km?
desktop_demon
In principle the 904nm infra red beam would refract differently to the visible wavelengths of light. So the resulting misalignment could introduce an error factor when targeting a given vehicle. I am not sure how one might measure such errors - the visible bit it easy - the infra red bit could be more difficult given the dynamic circumstances of the experiment.
justforthepictures
It is also worth reiterating that the crosshairs are merely an indication of the laser targeting, that depends on the camera to subject distance in relation to where along the section of carriageway the laser/crosshair alignment has been carried out.

It is very common for a picture to be posted here for opinions on the ‘ping’ to be given, but the crosshairs aren’t used as the primary targeting method, this is done via the scope’s red-dot. So although a requested photo will show the crosshairs in relation to the target vehicle, parallax error in conjunction with the distance at which the crosshairs and laser have been aligned, will play a huge part in determining what value the crosshairs actually are. Then you must also accept that the apparent relationship between crosshairs and laser will alter depending on focal length set/used.

At the cessation of every period of enforcement, the LTI/camera is removed from the tripod and put back in its foam lined case, meaning the video camera lens is retracted. So every time it is used, the procedure of laser and crosshair alignment has to be carried out (the focal length will vary from site to site).

Many people wrongly interpret that the crosshairs are used to sight the target vehicle, but the crosshairs should only ever be regarded as a guide to laser positioning.
smeggy
QUOTE (desktop_demon @ Wed, 6 Apr 2011 - 19:49) *
In principle the 904nm infra red beam would refract differently to the visible wavelengths of light. So the resulting misalignment could introduce an error factor when targeting a given vehicle. I am not sure how one might measure such errors - the visible bit it easy - the infra red bit could be more difficult given the dynamic circumstances of the experiment.

IR light refracts less (is not bent as much) than visible wavelengths. Do an Google image search for "prism".

Those experienced in photography biggrin.gif will confirm there is a 'twinkle' effect in such conditions. Such an effect would also show on the video as distant objects being blurry or jumping around (depending on the aperture of the lens). Assuming a uniform level of atmospheric turbulence between the camera and target, the amount of visible blurring/jumping of the targets vehicles is a good indication of how much the laser beam at the target is blurring/jumping (but that level will be slightly reduced for 904nm than for the visible wavelengths).


From my experience, I would say the effect is negligible (compared to the width of the laser at the target).
smeggy
QUOTE (norahl @ Wed, 6 Apr 2011 - 17:48) *
QUOTE (justforthepictures @ Mon, 4 Apr 2011 - 17:28) *
... the chronological lag before the asterix appears can often mean that another vehicle may then encroach into the 'offence shot', i.e. the moment when the speed reading has appeared...

...and what would that lag be?

I don't know what the RS232 lag is, or what it could be.
Strictly speaking, the simple answer would account for the 'timeout' delay as any of the frames up to the timeout/blocks can be used as evidence. This is usually up to around 0.16 seconds.

More complicated answers could account for the samples that can be dropped, or even timing from the end of the 'acquisition' period.
StationCat
Sorry for bringing up an old thread, and I don't know if the author still monitors, but I was curious about the parallax issue in the 'aligning the red dot' section of the report. I understand that the red dot on the laser gun appears in a head-up display and through the clever use of optics appears much further away. Clearly this would be necessary otherwise the dot would appear out of focus to the operator and impossible to view (being a few cm away) at the same time as the target vehicle (possibly hundreds of metres away). Is it therefore the case that that diagram in the report should have the 'sight' position much higher therefore producing triangles which are much more elongated? Furthermore, in the example given if the red-dot is optically a few hundred metres away then how does this affect the target error?
andy_foster
QUOTE (StationCat @ Sun, 6 Jul 2014 - 19:55) *
Sorry for bringing up an old thread, and I don't know if the author still monitors, but I was curious about the parallax issue in the 'aligning the red dot' section of the report. I understand that the red dot on the laser gun appears in a head-up display and through the clever use of optics appears much further away. Clearly this would be necessary otherwise the dot would appear out of focus to the operator and impossible to view (being a few cm away) at the same time as the target vehicle (possibly hundreds of metres away). Is it therefore the case that that diagram in the report should have the 'sight' position much higher therefore producing triangles which are much more elongated? Furthermore, in the example given if the red-dot is optically a few hundred metres away then how does this affect the target error?


No.
nomadros
this...

http://dragoneyetech.com/

...is as bad ass as it gets.
Been out for a couple of years but it's still a pant wetter.


Review complete
StationCat
QUOTE (andy_foster @ Sun, 6 Jul 2014 - 20:56) *
QUOTE (StationCat @ Sun, 6 Jul 2014 - 19:55) *
Sorry for bringing up an old thread, and I don't know if the author still monitors, but I was curious about the parallax issue in the 'aligning the red dot' section of the report. I understand that the red dot on the laser gun appears in a head-up display and through the clever use of optics appears much further away. Clearly this would be necessary otherwise the dot would appear out of focus to the operator and impossible to view (being a few cm away) at the same time as the target vehicle (possibly hundreds of metres away). Is it therefore the case that that diagram in the report should have the 'sight' position much higher therefore producing triangles which are much more elongated? Furthermore, in the example given if the red-dot is optically a few hundred metres away then how does this affect the target error?


No.


Thanks for the concise answer. Can you please expand to explain why when looking through the scope, and the red dot is in sharp focus and appears to be much further away, there is not much apparent movement in its position on the target vehicle when your eye position moves?
TonyS
If you look through a telescopic sight the cross-wires are in sharp focus, and their position relative to the target doesn't change if you move your eye away from the centre line. I don't know how that's done optically, but presumably the same may be done with this speed gun.
Mad Mick V
A couple of current threads appear to query the cosine effect. The attached LIDAR theory manual has a table which will be of interest.

I've also uploaded the old Ultralyte LRB Manual if anyone needs this.

Mick
Yvan Dutil
Hi, I am a physicist with a background in instrumentation.

I am consulting on a case using a Ultralte LTI LR100. I a trying to gather as much information as possible on this device. Unfortunately, my request are rather precise. Hence, the local distributor (I am living in Quebec) can't answer and the company refuse to answer.

Maybe someone here can answer my question.

Is the field of view is limited by the detector of the laser beam. If it is the laser beam what is its half power beamdwidth. This may look symmetrical but behaviour is different.

Does anyone know why the uncertainty is 1 MPH for a device that can measure distance to within 6 in. This look a simple limitation of the display, but the conversion in metric is 2 km/h, which eitehr say the truth or a demonstration of the incompetency of the analysis.

Third, what is the official claim about how the anti-sweep error algorithm works. I have got different version of it. IMHO the parameters are relaxed in difficult operating conditions (long distance, small target, non perpendicular target, large road reflection) leading to erroneous measurements.

I will try to build a full model of the instrument including atmospheric propagation. Off course, more I know before, the best it is.

Thanks for your help.
Jlc
As you will know the accuracy of the device will far exceed +/- 1mph (or equivalent rounded km/h) but one would guess for simplicity that the display only displays integers so with rounding (down) and the possible cosine reduction due to a non-perpendicular measurement then the manufacturer would be a hostage to fortune in listing a more accurate headline accuracy.

In the UK we have quite a generous threshold before enforcement anyway. Such accuracy makes no difference whatsoever in the real world - I don't see any lines of 'attack' here.

I don't know how Canadian law works but here the devices are 'Type Approved' and the measurement is assumed correct unless the defendant can show otherwise - which in reality is near impossible. (But of course the reading is more likely than not to be correct). Additionally, the enforcement officer should be using the device to corroborate their opinion - in most circumstances it's easy to spot a speeding motorist.

The 'anti-sweeping' (or slip as it's known here) algorithm is obviously proprietary (and thus why difficult to get hold of) but I've never heard that it was adaptive depending on external factors. I did come across a document before which, from memory, gave an overview of the algorithm - that is the multiple delta's from the pulses were effectively 'plotted' and if a preset percentage of points deviated by another preset amount then the whole measurement was voided.
justforthepictures
QUOTE (Yvan Dutil @ Wed, 20 Jan 2016 - 21:10) *
I will try to build a full model of the instrument including atmospheric propagation.

Can I ask what you plan to charge your client for such an exercise?
Yvan Dutil
QUOTE (Jlc @ Thu, 21 Jan 2016 - 10:38) *
As you will know the accuracy of the device will far exceed +/- 1mph (or equivalent rounded km/h) but one would guess for simplicity that the display only displays integers so with rounding (down) and the possible cosine reduction due to a non-perpendicular measurement then the manufacturer would be a hostage to fortune in listing a more accurate headline accuracy.

In the UK we have quite a generous threshold before enforcement anyway. Such accuracy makes no difference whatsoever in the real world - I don't see any lines of 'attack' here.

I don't know how Canadian law works but here the devices are 'Type Approved' and the measurement is assumed correct unless the defendant can show otherwise - which in reality is near impossible. (But of course the reading is more likely than not to be correct). Additionally, the enforcement officer should be using the device to corroborate their opinion - in most circumstances it's easy to spot a speeding motorist.

The 'anti-sweeping' (or slip as it's known here) algorithm is obviously proprietary (and thus why difficult to get hold of) but I've never heard that it was adaptive depending on external factors. I did come across a document before which, from memory, gave an overview of the algorithm - that is the multiple delta's from the pulses were effectively 'plotted' and if a preset percentage of points deviated by another preset amount then the whole measurement was voided.


Here, police speedometer is assumed to by a prima facie proof if it has been checked to work properly and if the policeman is qualified. In theory, you can counter that by saying that you have checked your car speedometer or that your car could not go to this speed. IN that case I could help. In practice, however, this is very dependant of the judge feeling. Nevertheless, in a previous case, I have found a non documented failure mode of the radar. In consequence, the case was closed outside the court to not create any new jurisprudence.

Anyway, I am VERY interested in any document describing how the algorithm work. The patent of LTI describe an adaptive window methodology to find the peak. This indicates that in less permissive environment the error bar of the measurement will increase. The lidar must use the first «strong» return to trig the clock. The less sharp is this return the larger are the error.

I know it must pass a NTSA test, but this is done in a lab environment. In real life, measurement are much noisier. If the anti-slip algo is statistical it will be much more prone to error that in the lab where it is qualified. I am diffing trough all previous brought to court to understand the specific failure more of lidar.

By the way, does any one know the description fo the pulse train of this radar. I am under the impression there is one pulse, 100 ms wait, then a train of pulse each 8 ms for 0.3 s. It this is right?





QUOTE (justforthepictures @ Thu, 21 Jan 2016 - 11:12) *
QUOTE (Yvan Dutil @ Wed, 20 Jan 2016 - 21:10) *
I will try to build a full model of the instrument including atmospheric propagation.

Can I ask what you plan to charge your client for such an exercise?


The answer is yes, but not enough! I do charge some time, but it is much less that the true time I am speed on the case. Hopefully, overtime I will be able to reuse my knowledge and get some payback on my effort.

Also, I need to do some technical analysis just to stay mentally in shape. I have no official job right now, and this work help me to keep my skill. Simple differential (Vis vs IR) refraction analysis is not that hard to do. I would like to do the phase screen also, but I must find the adequate scientific literature reference.

By the way, does anyone in tried to used the video as an alternate speed measure? This should by a straight forward prima facie proof.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2019 Invision Power Services, Inc.