Thoughts on politics and life from a liberal perspective

Tuesday, 15 June 2010

Are motorists being wrongly penalised by speed guns?

This is a guest post from Dr Kalvis Jansons, an applied mathematician and motorcyclist. Kalvis hit the headlines last year when he launched the "Gordon Brown resign" No 10 website petition. He eventually got his wish!

Shortcomings of the LTI 20-20 laser speed gun

By Dr Kalvis M. Jansons

The LTI 20-20 is a laser speed gun in common use by the police throughout the UK. It works by sending a stream of short infrared pulses from the gun to the target, and measuring the return times to determine distance. The rate of change of this distance is then interpreted as speed. Since not all parts of the target vehicle will be at the same distance, small movements of the gun can lead to erroneous speed measurements, as the rate of change of distance measurements is not entirely due to the motion of the vehicle. This is called the slip effect, and has been the subject of many reports, many of which feature Dr Michael Clark. The manufacturer and UK distributor of the LTI 20-20 boldly claim that the closed-source error trapping software will trap slip and other errors, provided the LTI 20-20 is correctly used by a trained police officer. However, Dr Clark's observations make this claim look somewhat suspect, and the closed-source nature of the code means that professional scientists like myself cannot check the algorithms used. Furthermore, many scientists with whom I have discussed the LTI 20-20 agree with Dr Clark that the design concept of this device is fundamentally flawed.

Based on the LTI 20-20 UtralLyte 100 user manual, the acquisition time is less than 0.4 seconds, the range is 23 to 610 metres, and it can determine speeds from 0 to 299 mph with error less than 1 mph. It can measure the distance of the target within 0.1 metres and the beam divergence is 3 milliradians.

Many speed convictions are obtained by police officers not following the LTI 20-20 instruction manual. To avoid slip errors, the LTI 20-20 manual says that the target should be the number plate as this will be a good reflector of the infrared pulses and has the correct orientation. However, the LTI 20-20 is also used on approaching motorcycles, which do not have front number plates.

There are many other issues with its use on motorcycles that concern me. Motorcycles are generally much more irregular in shape than other vehicles, particularly from the front, and even more so if they are without fairing. The scope for slip in motorcycle speed measurements seems intuitively to be much greater than that for cars, but I am unaware of any research which quantifies this effect. In fact, I am unaware of any research that determines the rate of false positive errors for different vehicle classes. Furthermore, these rates of false positive errors should be determined for different distances and speeds, and the data made available to the public for use in court.

The manufacturer of the LTI 20-20 produced a report on the concerns of Dr Clark entitled:

Review, Analysis and Criticism of the Report by Michael A.G. Clark, Ph.D.: "Technical note: Unreliability of Laser Speed Meters" (Issue 1, Copyright 2005) prepared by Daniel Y. Gezari, Ph.D. and Chevy Chase, MD

This report is, in my opinion, a very unprofessional and unscientific attack on Dr Clark, and is filled with unsupported assumptions regarding the LTI 20-20, and what could be expected of a police officer using the LTI 20-20. It contains no new science to address Dr Clark's concerns, and suggests that Dr Clark does not know how to do science, but that the author of the report does. If the report had attempted to address the issues in a more professional manner, I would have taken those points one by one and replied to them here. However, this report is not a useful basis for discussion.

Having read the report I was left wondering about the scientific background of the author. Most of the research of Dr Gezari is not, in fact, in a field directly applicable to the theory behind the LTI 20-20. However, some of his recent papers are, at least, concerned with laser measurements of the Earth-Moon distance. In his article "Lunar Laser Ranging Test of the Invariance of c", he believes he has found "a first-order violation of local Lorentz invariance; the speed of light seems to depend on the motion of the observer after all, as in classical wave theory, which implies that a preferred reference frame exists for the propagation of light." Any scientist reading this would have alarm bells ringing and maybe they would be hearing the music of "The Twilight Zone". Roughly speaking, based on a very simple but erroneous calculation, Dr Gezari believed that he had overturned most of modern physics from Maxwell, Einstein, Hawking, etc. Most of us would probably think hard before publishing such a paper, and most likely assume that there was an error of some kind. What was his error? He forgot to correctly take account of the motion of the Earth and the Moon.

It is clearly true, as pointed out by Dr Gezari, that the working of the LTI 20-20 does not depend on special relativity, but it is noteworthy how Dr Gezari appears to deal with obvious errors in his scientific work. He seems to totally ignore them, and assume everyone else is wrong!

In my opinion, based on the report itself, and what I now know about the work of Dr Gezari, this report has no credibility, and almost no content. Furthermore, I also find it disturbing that Dr Gezari has served as an expert witness in a dozen court cases involving laser traffic speed measurements, and his report is used in UK courts.

Assuming that Dr Clark's observations are correct, and they should be independently checked, I believe that the approval of the LTI 20-20 for use in the UK was initially a cock-up but the defence of it risks degenerating into a conspiracy. So much time and money has been invested in the use of the LTI 20-20 in the UK that in the short-term it is easier for all involved with its use to defend it rather than to review it properly in an open and scientific manner, and the report of Dr Gezari is just one part of this tangled web.

The fixed penalty fine system does not help either. Many people who know they have been wrongly charged for speeding using the LTI 20-20 will just pay up and take the points on their licence as the penalties can be so much larger if you go to court and fail to win your case. Most cases are won on the basis of incorrect use of the LTI 20-20, rather than on its flawed design concept and the erroneous false positive results that are likely to follow. A device like this should have never been approved without much more thorough testing with the results of those tests published in full and posted on the web for all to see.

Dr Clark appears to have demonstrated that the LTI 20-20 can be unreliable, so the use of the LTI 20-20, and similar devices based on laser distance measurements, needs to be urgently and thoroughly reviewed. Remember, that an erroneous speed measurement by such a device can put you in prison, or get you banned from driving, and several people have been clocked by the LTI 20-20 at speeds beyond the capability of their vehicles according to the vehicle manufacturers.

Many of the documents on which this comment is based can be found here:

1 comment:

Rob Fisher said...

Thanks for this. I ride a motorcycle and I'll be keeping a link to this article handy, just in case!