The on-going development of driverless cars is seldom out of the headlines. There are questions about the reliability of the software which will drive the cars and who will be liable should there be an accident.

Our own family car displays the current speed limit on the dashboard, having read this from the last speed limit sign that it passed, so it appears that the developers of driverless car are working on using AI (artificial intelligence) to interpret the input from the cameras on the cars. This make sense, after all, if one were to rely on maps etc, one could not react to changes in speed limit displayed on intelligent motorways.

Once the software to control driverless cars has matured, it will be necessary for various countries to “marry” up systems from various countries so that the owner of a driverless car can use it to visit other countries. This “marrying up” involves not only the technical aspects of the software associated with the cars, but also the legislation surrounding the use of the vehicle in countries other than the country where it was registered. Three of the most prominent issues facing Britain when Continental drivers bringing their driverless cars to the UK (and when British drivers taking their driverless cars to the Continent) are:

• Driving on the left or the right.
• Rules pertaining to priority at junctions.
• Possible conflict caused by switching between metric and imperial units.

Although the first two will probably pose bigger problems than those caused by switching units, it is the latter that is of concern to the editors of Metric Views.

A Look at the Railways

Before looking at the problems associated with changes between mph and km/h on the roads, it might be instructive to recap what is happening on the railways. Over the years, the various railway companies in Europe developed a multitude of signalling systems which hindered international railway travel as locomotives often had to be changed at the borders. In December 1989, European Transport Ministers decided that a standardised Train Control System for Europe’s railways should be developed to reduce the way in which the various state railways’ disparate signalling systems hindered cross-border traffic. The resultant system is known as the European Rail Traffic Management System (ERTMS), with the European Train Control System (ETCS) as its signalling component.

ETCS is now being rolled out across much of Europe. Unsurprisingly ETCS uses metric units. At the time of writing, it is operational in the UK on the Cambrian Line (where it was first trialled), parts of Crossrail and the Thameslink line and is being rolled out on the East Coast Mainline and will also be used on HS2. By the end of next decade, the bulk of the British railway network should be using ETCS.

What is ETCS? ETCS is a digital based system that does away with line-side signalling equipment. As a result, the train driver relies on the instruments in front of him rather than those on the ground. In this way he is more like an airline pilot than the driver of a car. Trains pick up information telling them about the state of signals that are ahead of them, how fast they are allowed to travel etc from boxes (called Eurobalises) that are mounted on the track. In order to allow for its roll-out, ETCS can operate at one of three levels – at Level 1, it merely replaces existing controls in the cab often replicating lineside equipment, while at Levels 2 and 3 it calculates the train’s exact position and shows all information inside the cab without reference to lineside equipment.

The use of mph was of particular concern to British users, especially when a train would be operating at Level 1 and the driver had to follow lineside speed restrictions given in mph and then the train moved into a Level 2 area where the driver would receive instruction in km/h via his display. The solution was that ETCS would use metric units throughout, but the British variant would display target and actual speeds in mph when operating a Level 1. As the trains enter a Level 2 area, the display will change to a km/h display, with the speedometer graduations changing such that the indicator needle was in the same position regardless of units displayed.

Lessons from the Railways

The biggest difference between the development of the driverless car and the development of ETCS is that on a railway system, railway management has complete control over the system including the vehicles using the track. This is not the case for driverless cars. It is commonplace for drivers to take their cars to foreign countries – I am sure that many MV readers have taken their cars to France, Spain or Belgium on a ferry or on Eurotunnel.  Furthermore, one does not need to stop when motoring between Dublin (speed limits in km/h) and Belfast (speed limits in mph) except to pay a toll.  If one is using a driverless car, it follows that one should be able to use one’s car anywhere in Europe where driverless cars are permitted.

The first lesson that can be learnt from ETCS is that in order to minimise the risk of problems when changing units of measure, the UK agreed to phase in the use of metric units as and when needed.

Another interesting aspect that could be learnt from the railways is that if one uses an analogue speedometer with a digital backdrop (whether the needle is genuine or digital) the positions of the mph or km/h graduations can be set such that if the backdrop is flipped between mph and km/h (or vice-versa) the needle will not change position unless the speed actually changes.

What problems could happen?

One of the sayings in software engineering is “Bugs congregate on the borders and breed in the corners.” In the case of the driverless car, one such border is the metric/imperial dividing line.

In theory, it should be possible to program the AI module of the driverless car to operate in any country – on start-up the car would use GPS to check which country it is in and then use the instructions for that country.

Should a country decide to change its regulations there could well be a problem – for example, the maximum speed limit might be increased or decreased, which, in the United Kingdom could be done without changing any road signs. This would necessitate an update to the configuration tables in every car in Europe. Such changes would obviously have to be kept to a minimum and if the UK decided to change over to the metric system once driverless cars were commonplace, the logistics of managing the change are such that while the change-over is taking place, the use of driverless cars could be banned. In the Republic of Ireland, such a change was done over one weekend.

It is impracticable for speed limits to be transmitted to driverless cars in real time – they change quite frequently on “intelligent motorways” and again communicating these changes to drivers would poses problems. The obvious way is for the cars to rely on their on-board cameras to read the speed limits in the same way that human drivers read the road signs.

In practice, one could expect AI upgrades to be installed in driverless cars at regular intervals. In theory, each new release should be thoroughly tested before release, but in practice, corners are often cut and this is where bugs are able to enter the system. The more complicated the “rules” pertaining to the software changes, the more likely it is that subtle bugs could enter the system. One such scenario is where a Continental-based car manufacturer regards mph as an “add-on” and does not adequately test the associated software. Such an error might only be discovered when the owner decides to visit the UK.  The obvious way around this situation is to remove the relevant restriction altogether and for the UK to adopt km/h in anticipation of the introduction of driverless cars.

Similar situations could arise with misinterpretation of height and width limit signs.

What should be done?

What should be done to forestall the problems that I have highlighted above? The obvious solution (and one obviously favoured by MV) is to speed up the conversion of road signs to metric units. However, the editors of the Daily Express and the Daily Mail are likely to have fits should a wholesale change be announced without a sound technical reason behind it. There will be calls for “Betrayal of BREXIT” and the like. The solution is therefore for the Government to clear the path for a change-over in a low-profile manner so that should a change-over be necessary, it can be done once it becomes obvious and with a minimum of upheaval.

It is suggested that the Government takes steps to ensure that all height, width and length signs are upgraded as soon as possible to show both metric and imperial units as per the 2016 version of the TSRGD which only mandates dual units once signs reach the end of their useful lives. Thereafter it should become mandatory for the display of vehicle heights in the driver’s cab to be in metric units. With all this in place, it should be possible to safely align vehicle height, width and weight signs with those on the Continent at short notice.

As regards speed limit signs, the Government can make a few small changes in legislation to mitigate the impact of a conversion to km/h should it be technically necessary to do so.  The one change that could be made is a requirement that the user of any new motor vehicle that is sold in the United Kingdom should be able to change all the units on the dashboard from imperial units to metric units and vice-versa without outside assistance. Moreover, the metric dashboard should be no different to its continental equivalent.

I believe that if these changes are done to satisfy a technical need rather than a political need, there will be little resistance from the general public. It should be noted in passing that I have made no reference to the display of distances on road signs.

References

https://www.modernrailways.com/article/digital-signalling-and-control

https://railengineer.co.uk/etcs-more-deliberations

https://www.thalesgroup.com/en/markets/transport/signalling/signalling-solutions-main-line-rail/european-train-control-system-etcs

https://en.wikipedia.org/wiki/European_Train_Control_System

https://tsrgd.co.uk/pdf/tsrgd/tsrgd2016.pdf

International Vocabulary of Metrology (VIM) is a “System of Units”, not a “System of Measurement”. The VIM explains the terminology used in publications related to measurement, including the official SI brochure.

Background

Communication between people relies on an agreement as to what various words/gestures mean. The Oxford English Dictionary (OED) provides a reference point for words used in everyday English. Its model is to catalogue words that have become commonplace in the English language once they have been widely accepted. However, this does not prevent differences in meaning emerging on each side of the Atlantic Ocean. The motor industry is a prime example of how two different vocabularies have emerged despite the same technology.

Many specialised disciplines have produced dictionaries of jargon that has become commonplace in that particular discipline. Often the dictionaries concerned use definitions that originate in academia to forestall multiple concepts being attached to the same word or one concept having multiple names. In the case of metrology, eight international organisations including BIPM and ISO have joined forces to create an organisation called the Joint Committee for Guides in Metrology (JCGM). Their brief is to produce documents that ensure consistency in language across the work of all the member organisations. [Ref 1] To date two documents have appeared – International Vocabulary of Metrology (VIM) [Ref 2] and Guide to the Expression of Uncertainty in Measurement (GUM). The documents themselves are published by the constituent organisations using their own covers and are available free of charge on the Internet.

International Vocabulary of Metrology

The first of these two documents is of principal interest to MV readers as it defines the terminology used in the SI Brochure and elsewhere. Moreover, both VIM and GUM appears as bilingual documents with the aim of ensuring a one-to-one correspondence between the English and the French languages. Countries that use languages other than English or French are encouraged to translate these documents into their own languages. To date the author has identified translations in Spanish [Ref 3], Italian [Ref 4] and German [ref 5].

The fourth edition of VIM is currently undergoing review, so this article will discuss the third edition (published 2008, updated 2012). The document contains five chapters:

1. Quantities and units
2. Measurement
3. Devices for measurement
4. Properties of measuring devices
5. Measurement standards (Etalons)

Readers of MV are probably most interested in Chapter 1 of the guide with Chapters 2 and 5 putting the contents of Chapter 1 into context.

The first section has formal descriptions of the terms used to describe a system of units starting with the description of a “quantity” which can either be continuous in nature such as “length”, “mass” etc or discrete in nature such as “Rockwell hardness”.  It formally defines a hierarchy of concepts, many of which are dependent on concepts that have been previously defined. For example, the formal definition of the “International System of Units” (VIM definition 1.16) is:

“system of units, based on the International System of Quantities, their names and symbols, including a series of prefixes and their names and symbols, together with rules for their use, adopted by the General Conference on Weights and Measures (CGPM)”.

This definition is dependent on the definitions “system of units” (VIM definition 1.13) and “International System of Quantities” (VIM definition 1.6), both of which are defined within the document. [Note 3]

The way in which systems of units are defined is done in a way that is context neutral – the definitions hold equally well for the imperial system as they do for SI with the obvious exception that the imperial system does not comply with the definition of a “coherent system of units”.

The rest of the document is oriented towards making measurements and how one should express uncertainties associated with such operations.  The concluding chapter provides a vocabulary for the creation of physical standards such as the UK national prototype metre and kilogram which are held by the NPL and copies of those standards that are held by trading standards officers. I will not describe those sections in detail –VIM can easily be downloaded from the internet.

“System of Measurement” or “System of Units”?

The document’s introduction has a paragraph on the use of the French words “mesure” and “mesurage” and in particular why the word “mesure” can be ambiguous unless it has a qualifier such as “unité de mesure”. One of the consequences of this is that the English equivalent of the expression “système de mesure” is “measuring system” (VIM definition 3.2) while VIM makes no mention of the English phrase “system of measurement”.  [Note 1].

VIM describes the term “system of units” (VIM definition 1.13) [Notes 2 & 3] as being equivalent to the French “système d’unités”. The Spanish equivalent is “sistema de unidades” while the Italian and German translations likewise have equivalents of the word “units” rather than the word “measurements”.  This suggests that the use of the expression “System of measurement” should be abandoned for of the expression “System of units” (which can be taken to be a shorthand for “system of units of measurement”).

SI is, of course, an abbreviation for “Système international d’unités” or, in English “International System of Units”.

Finally …

The author recommends that anybody who is writing about the metric system reads the International Vocabulary of Metrology, if only to ensure that they are aware of the correct use of vocabulary.

Notes

1. “Measuring System” (VIM definition 3.2) is defined as “set of one or more measuring instruments and often other devices …”
2. “System of Units” is defined “set of base units and derived units, together with their multiples and submultiples, defined in accordance with given rules, for a given system of quantities”.
3. Items in bold are defined within VIM.

 References

1. https://en.wikipedia.org/wiki/Joint_Committee_for_Guides_in_Metrology
2. https://www.bipm.org/documents/20126/2071204/JCGM_200_2012.pdf/f0e1ad45-d337-bbeb-53a6-15fe649d0ff1  
3. https://www.cem.es/sites/default/files/vim-cem-2012web.pdf
4. https://www.eurachem.org/images/stories/Guides/pdf/TAM_2011_IT.pdf
5. https://www.libristo.eu/en/book/internationales-worterbuch-der-metrologie-international-vocabulary-of-basic-and-general-terms-in-metrology_01480812

The International Organisation for Legal Metrology (OIML from the French “Organisation Internationale de Métrologie Légale”) is housed in a non-descript office in the Rue Turgot in the 9th arrondissement of Paris and is about 700 metres the from the Gare du Nord. Like the BIPM, the OIML is an inter-governmental organisation which gives its staff quasi-diplomatic status, though, like the BIPM, French nationals pay French income tax. The OIML works closely with BIPM, the International Standards Organization (ISO) and other international bodies in coordinating metrology around the globe, with each organisation having its own specific role.

Logo of the OIML

MV readers are no doubt aware of the role of BIPM is to define units of measure and the means to realise those definitions to the highest degree possible. This might be good for science, but what about the practical world? If I buy a kilogram of sugar, am I really bothered whether it is a few grains short or whether it is a few grains over the specified weight? The cost of getting that packet correct to the nearest microgram will far exceed the cost of the sugar. This is where the OIML come in. The role of the OIML is best explained by an example. In the years leading up to 1976, the OIML and the EEC worked together to produce a specification (Directive CEE 76/211) which resulted in the “e” (for “estimated”) symbol that now appears on packets of sugar (and many other goods).  In the case of 1 kg packets of sugar, the EU directive requires that the average weight per packet in a batch shall be at least 1 kg and that the maximum short weight allowed for any one packet in that batch is 7.5 grams.  A number of countries outside the EU including Australia and South Africa have included the EEC specification into their domestic legislation. The UK, on leaving the EU has retained the legislation that was produced under the directive.

OIML Establishment

The OIML is an inter-governmental organisation established in 1955 to facilitate and coordinate legal metrology on an international scale. “Legal Metrology” is the discipline of making and recording measurements that have a legal implication. This includes but is not restricted situations where the amount of money that changes hands is dependent on a measurement, where a party might have a liability for their product which is measurement-dependant or where a measurement will determine whether an offence (such as speeding, drink-drive etc) has been committed. Part of the OIML’s brief is to assist in the international harmonisation of legislation in such matters.

The OIML has a three-tier structure – the “Conference” which consists of representatives from the member states, and which meets once every four years, the “International Committee” which oversees technical work, and which drafts formal output from such work and the “Bureau”, a secretariat based in Paris. Much of the CIPM’s work is done through specialist committees whose members are employed by metrology-oriented organisations in participating member states.

The three principal activities of the OIML are:

• Overseeing of and issuing of internationally recognised (“Mutual recognition”) certificates to laboratories around the world certifying that their work is in accordance with international standards.
• Publication of recommended standards pertaining to legal metrology and production of expert examinations from a metrology point of view of proposed legislation.
• Joint publications with other organisations such as ISO and BIPM to ensure that there are no contradictions between outputs from the various organisations.

OIML Outputs

The principal outputs from the OIML are sets of publications. Those of interest to readers of MV are:

Joint activities with other organisations which include:

• The International Vocabulary of Metrology (VIM) which is a joint production between eight organisations.
• Liaison with the BIPM’s Consultative Committee on Units, the committee which, amongst other things, produces the SI Brochure.

Recommended standards that member states are advised to incorporate into their own laws. To date 150 recommendations have been published. Published recommendations are revisited once every few years at which time some are revised while others might be deprecated. At the current time there are 104 current recommendations. Recommendations that have been published that are of interest to MV readers include:

• R 52-en – “Hexagonal weights – Metrological and technical requirements”. Hexagonal weights have been used for metric units for over a century. This recommendation describes standard sizes so that the user can mix-and-match weights from different manufacturers.               
• R 76-en – “non-automatic weighing instruments” – This recommendation describes the specifications for four different classes of weighing devices – defined as Class I, Class II, Class III and Class IIII respectively. Supermarkets and similar establishments are usually required to have Class III devices (ability to weigh objects up to 1 part in 10 000) whereas Class IIII devices (up to 1 part in 1500) are designed for use in mobile medical health clinics.
• R 111-en – “Weights of classes E1, E2, F1, F2, M1, M1-2, M2, M2-3 and M3”. This recommendation defines eight different classes of weights, each class having different accuracy requirements. If goes without saying that cheaper weights have less accuracy but are often more robust.
• R 138-en – “Vessels for commercial transactions”. This recommendation describes, amongst other things, the allowances made when calibrating glasses for use in pubs.

Other recommendations include taxi meters, speech audiometry equipment, automatic rail weighbridges, vehicle exhaust measuring instruments, speedometers and so on.

Finally, the OIML is sometimes asked to produce expert reports of certain topics. Seven such reports have been produced including the report that led to the EEC (as it was) introducing the “e” mark for an “estimated quantity”.

Conclusions

Rather than summarising the work done by the OIML, let me give two examples from their various outputs which I believe will give MV readers food for thought:

• Imperial pint glasses are variously described as being equivalent to 568 ml or 570 ml. OIML recommendation R 138-en allows a maximum error of (10 ml + 2.5%V) for glasses where the volume shown by a gauge mark (V is the nominal volume of the glass). This is far greater than the 2 ml variation in describing the size of the glasses. Thus, if a manufacturer exercises reasonable care in manufacturing such glasses, their products can fit either description of the pint.
• The SI is often described as a “System of measurement” which is a shorthand for “System of units of measurement”. VIM describes the “International System of Units” as a “system of units based on the …” and it describes a “Measurement system” as “set of one or more measuring instruments …”.  VIM also draws to attention that in French, the translation of “measurement system” and “system of measurement” are identical – “système de mesure”. This is the underlying reason why VIM describes the SI as a “system of units” rather than a “system of measurement” – after all, “SI” translates to “International System of Units”, not “International System of Measurement”.

References

1. OIML home page: www.oiml.org
2. OIML analysis of EEC directive CEE 76/211:  https://www.oiml.org/en/files/pdf_e/e004-e04.pdf/view
3. SI Brochure (9th edition): https://www.bipm.org/documents/20126/41483022/SI-Brochure-9-EN.pdf (Para 1.3)
4. List of OIML Recommendations: https://www.oiml.org/en/publications/recommendations
5. Text of the International Vocabulary of Metrology: https://www.oiml.org/en/files/pdf_v/v002-200-e07.pdf