The future of smart motorways

[ManomayLR]

This is my biggest scare of Smart Motorways. Whilst compliance to overhead individual-lane speed limit signs seems quite high, in my experience compliance to (changes of) speed limits displayed on MS4s is minimal, leading to a mix of 40 mph and 80 mph vehicles.

My suggestion would be to not allow a change in speed limit via MS4, only allowing them to be used as repeaters.

How can people not believe that a single speed limit sign does not apply to all lanes? If it was a normal fixed sign there would be a sign on the nearside and perhaps one in the central reservation, and people would know that applies to all lanes.

Gantries create visual pollution - which is one reason why lane control signals are avoided when they are not needed.

While this is covered in TSRGD S15-1-6, compliance seems to be better if there are other prescribed elements which may explain the reason for the limit.

The question is then why those other prescribed elements fail to be provided in a consistent manner.

[Big L]

…
One gantry on 20 mph and others on 40 mph = illegal signage configuration (as no division in carriageway), but some people were complying with it.
…

This morning on the N6, joined at j12 southbound. At the first gantry we had lane 1 featuring a “move to the right” arrow, lanes 2 and 3 with 30 limits and lane 4 showing 40. Same at the second gantry except it was red x, 30, 30, 40.

[tom66]

…
One gantry on 20 mph and others on 40 mph = illegal signage configuration (as no division in carriageway), but some people were complying with it.
…

This morning on the N6, joined at j12 southbound. At the first gantry we had lane 1 featuring a “move to the right” arrow, lanes 2 and 3 with 30 limits and lane 4 showing 40. Same at the second gantry except it was red x, 30, 30, 40.

What's clear to me is whatever software does set the signs doesn't have sufficient checks against invalid configurations.

It shouldn't ever be possible to set just one limit sign to 40 and leave the rest blank. Nor should it be possible your your scenario to arise.

[Conekicker]

Panorama. So many things that weren't talked about that aren't in the public domain. Just saying...

[SteelCamel]

If safety-critical equipment failed with this frequency on the railways, the whole network would implode.

That's because railway signalling fails safe. Vehicle detection (track circuit or axle counter) failures set the signals to danger until the fault is rectified. An unlit signal means danger and can't be passed without special arrangements and permission. The network will shut down rather than have trains run where it can't be proved safe.

Smart motorway signals fail unsafe. Vehicle detection failures set the signals to show the road clear. Unlit signals mean proceed at full speed. The network will keep running, even if it means sending drivers blindly into unsafe situations.

It would be easy enough in principle to change. If a radar fails, mark the lane closed. If it fails across multiple lanes, drop the limit to 40. Have the gantries display NSL signs when the road is clear, and sign that blank displays mean 40. Of course this would mean that the tech would have to get fixed or everyone would be driving at 40 for months - which is why I didn't suggest a lower limit. If the same rules as railways were followed, a failed sign would be the same as a red X. Everyone has to stop and wait for the police to arrive and escort them past it.

[ManomayLR]

If safety-critical equipment failed with this frequency on the railways, the whole network would implode.

That's because railway signalling fails safe. Vehicle detection (track circuit or axle counter) failures set the signals to danger until the fault is rectified. An unlit signal means danger and can't be passed without special arrangements and permission. The network will shut down rather than have trains run where it can't be proved safe.

Smart motorway signals fail unsafe. Vehicle detection failures set the signals to show the road clear. Unlit signals mean proceed at full speed. The network will keep running, even if it means sending drivers blindly into unsafe situations.

It would be easy enough in principle to change. If a radar fails, mark the lane closed. If it fails across multiple lanes, drop the limit to 40. Have the gantries display NSL signs when the road is clear, and sign that blank displays mean 40. Of course this would mean that the tech would have to get fixed or everyone would be driving at 40 for months - which is why I didn't suggest a lower limit. If the same rules as railways were followed, a failed sign would be the same as a red X. Everyone has to stop and wait for the police to arrive and escort them past it.

Which is why I am such a proponent of the "downward arrow" approach - every lane control signal should show a green or white arrow pointing downward, towards the lane to which it applies, when it would otherwise be blank.
And the first gantry or MS4 after every junction, should confirm the current speed limit, even if it is NSL.

[CableTracer]

But surely the power should only be isolated for the section of works, and as soon as the works finish the following gantry or VMS should be on and set to NSL. The other day I saw NSL signs with a VMS saying "roadworks end", for the first time. Ignoring the fact that all the previous signs had been set incorrectly in that instance, that should be seen more.

I would agree, but my guess is that the works areas and the section over which they can isolate power don't necessarily match up. Hypothetically (ficticious numbers), the works area might be 2 miles long, but the power supply might only be isolatable in 10 mile sections or between junctions or something like that. It would mean the roadworks would end up being longer than necessary if that was always applied.

Cabinet sites are typically up to 1km from a DNO electrical supply point, so 2km could be isolated for works. This distance is typically far less alongside junctions and interchanges where multiple power supply points are present and is favourable to the typical 2-5km closure length. A power outage is logged as an absence of connectivity to an individual asset, cabinet site or locality.

Planned outages to upgrade equipment or resolve faults along the distribution route of up to 5 cabinet sites have to be planned and authorised in advance and then further authorised by the control room once a closure is installed. i.e an incident will prevent authorisation being given. Once fixed plates are in place and the closure installed, the speed signs are switched off allowing this isolation to occur. Minimal work along the supply route is necessary once a scheme is operational - this is typically to resolve any faults or to divert power supplies for new assets and is conducted whilst permanent TM is installed.

Local isolation of a cabinet site to resolve vehicle collision damage, upgrade equipment or resolve water ingress issues is much more common, but the same authorisations are needed. The most significant contributor to wider outages are DNO outages (which affect a local area and by their nature are unplanned), cable theft / vandalism and general faults.

Off network access is available to many sites, but a combination of permanent road layouts, distance between access locations, the growth of vegetation and the level of debris and detritus along the verge prevents access to many sites without closures which typically require a 3 week booking duration. Whilst critical issues are resolved within 7 days by using other powers or sharing others closures, access to replace faulty AMI's or MS4's that cannot be resolved at the roadside (with MEWP's and Hi-ab) is often blocked by the 2.5km proximity to other pre-booked works such as HS2 closures, surfacing, VRS repairs ect which may have a conflicting closure (LBS3-4 ect). It is common that TM for critical - but quick fix issues is installed early and then removed - delaying major schemes by a few hours to allow the issue to be resolved.

Collaboration occurs between major schemes, area teams technology, surfacing and VRS teams, principal bridge / gantry inspection teams ect on a continuous basis. When one stretch is resolved, collisions, theft or a lost load will no doubt occur the next day repeating the process. The network is also busy - with closures easily being delayed until midnight or 1am due to high traffic counts and having to be vacated at 04:30 for a 6am reopening as the country wakes up.

[Helvellyn]

If safety-critical equipment failed with this frequency on the railways, the whole network would implode.

That's because railway signalling fails safe. Vehicle detection (track circuit or axle counter) failures set the signals to danger until the fault is rectified. An unlit signal means danger and can't be passed without special arrangements and permission. The network will shut down rather than have trains run where it can't be proved safe.

Smart motorway signals fail unsafe. Vehicle detection failures set the signals to show the road clear. Unlit signals mean proceed at full speed. The network will keep running, even if it means sending drivers blindly into unsafe situations.

It would be easy enough in principle to change. If a radar fails, mark the lane closed. If it fails across multiple lanes, drop the limit to 40. Have the gantries display NSL signs when the road is clear, and sign that blank displays mean 40. Of course this would mean that the tech would have to get fixed or everyone would be driving at 40 for months - which is why I didn't suggest a lower limit. If the same rules as railways were followed, a failed sign would be the same as a red X. Everyone has to stop and wait for the police to arrive and escort them past it.

An important difference is that (other than in some very low-speed situations) trains are driving to the signals, not on sight, whereas road vehicles are supposed to be driven on sight. A train driver seeing a green signal means they can proceed at the maximum allowable speed, even if they can't see far ahead. A driver on a road vehicle should never assume that any sign means they don't have to look at the road and respond to what's in it. If a driver enters an unsafe situation blindly that's entirely on them, whatever signs say. If a railway signal were to do that with a train it wouldn't be the fault of the driver.

The difference in performance between trains and road vehicles (primarily in stopping ability) makes these different approaches a necessity.

[roadtester]

I’ve just watched the latest Panorama, and once again, it seemed a bit muddled.

Two of the problem cases they complained about involved cars that had managed to get off the running lanes and onto the verge, and they also showed several cars stopped but off the carriageway, even in the absence of a shoulder. I can’t see that that’s much different to a case where a hard shoulder is present. One driver presented as a ‘victim’ was someone who had driven into the back of a stationary car in lane one. Obviously that was a situation in which risk levels were raised - there was no warning via the gantry signs - but where presumably, at the end of the day, that driver was at fault.

I’m sure our motorways would be safer if they all had hard shoulders but hard shoulders don’t solve all problems - while they are clearly less dangerous than a running lane, they are still very risky places to be, and they may not be much help if your car fails if you are in lane 2,3 or 4, as the Panorama perhaps inadvertently conceded when it suddenly switched to complaining about the failure of the smart features on a section of the M4 that does have hard shoulders.

There is a real problem here, but as I said before, this was all a bit muddled.

[jackal]

Just imagine if the same journalistic resources were put into any one of the risks decreased by ALR, and we had endless articles and programmes about crashes on unimproved motorways that could have been avoided if only they had been converted to ALR. It's a monumental failure of journalistic ethics that is distorting public policy.

[jnty]

Just imagine if the same journalistic resources were put into any one of the risks decreased by ALR, and we had endless articles and programmes about crashes on unimproved motorways that could have been avoided if only they had been converted to ALR. It's a monumental failure of journalistic ethics that is distorting public policy.

Or if they ignored motorways completely - being one of the safest roads to be on - and looked a bit more closely at the hazards still continually designed into roads often only seconds' walks from people's front doors.

[DB617]

The largest scandal in my view is the amount of money that was put into installing these assets, only for them to deteriorate prematurely - could be due to poor design or lack of interest in maintenance - and be operated incompetently and inconsistently on a daily basis. I suppose when they were installed the government of the time was committing to the installation costs but also intrinsically committing all future governments to the operational funding, else the system became diminishingly safe. Passive safety on the roads has its merits because - to a point - the system is self-governing. The proliferation of traffic lights on large roundabouts is another example of this concept falling apart - arguably if the lights were allowed to fail traffic flow would probably increase because of the usually poor designs, but that's besides the point - as detection loop and control system faults seem to be becoming more common leading to roundabouts that flowed better without lights now flow much, much worse because the control system failed back to minimum stage times. I have seen this twice within a mile of my house recently. Not to mention pinch point schemes which operate on ludicrous minimum timings or have no detection response at all...

TLDR - Because we have made the system more and more dependent on active maintenance and operational expertise and staffing, to be efficient and safe, and government and local budgets and expertise are at an all-time low, the system is failing and the enormous amount of taxpayer money spent has contributed to just making everything worse. Nice.

[ManomayLR]

The largest scandal in my view is the amount of money that was put into installing these assets, only for them to deteriorate prematurely - could be due to poor design or lack of interest in maintenance - and be operated incompetently and inconsistently on a daily basis. I suppose when they were installed the government of the time was committing to the installation costs but also intrinsically committing all future governments to the operational funding, else the system became diminishingly safe. Passive safety on the roads has its merits because - to a point - the system is self-governing. The proliferation of traffic lights on large roundabouts is another example of this concept falling apart - arguably if the lights were allowed to fail traffic flow would probably increase because of the usually poor designs, but that's besides the point - as detection loop and control system faults seem to be becoming more common leading to roundabouts that flowed better without lights now flow much, much worse because the control system failed back to minimum stage times. I have seen this twice within a mile of my house recently. Not to mention pinch point schemes which operate on ludicrous minimum timings or have no detection response at all...

TLDR - Because we have made the system more and more dependent on active maintenance and operational expertise and staffing, to be efficient and safe, and government and local budgets and expertise are at an all-time low, the system is failing and the enormous amount of taxpayer money spent has contributed to just making everything worse. Nice.

Technology is good but we need it to fail safe - or at least for it to be obvious when it has failed.

That means the VMS should be “always on” - lane control signals should always show the current speed limit, or an arrow pointing downwards, and VMS should show either the current speed limit (verge mounted MS4s on a smart motorway) or the current time (all other VMS)

The latter has been the convention in Wales for a while.

[nowster]

Capital expenditure versus operational expenditure. Different budgets.

You only have to look at the lack of simple maintenance: potholes, vegetation, etc. The same will be happening to the technology infrastructure but it won't be as visible.

[jackal]

I do think the operational point is a reasonable one. But it seems to support ALR, which is less reliant on technology, rather than DHS - and it's the latter that is the fallback position of the anti-SM crowd as well as a principal result of their campaign (i.e., DHS can't now be converted to ALR as was contracted).

[DB617]

I do think the operational point is a reasonable one. But it seems to support ALR, which is less reliant on technology, rather than DHS - and it's the latter that is the fallback position of the anti-SM crowd as well as a principal result of their campaign (i.e., DHS can't now be converted to ALR as was contracted).

Surely it's the other way around. DHS 'fails safe' in that the hard shoulder is (unless you don't understand the highway code) closed if all the overheads are off or failed. ALR 'fails flowing' in that it has the maximum capacity when all signals are off (usually 4 lanes at NSL) but with the trade off that if anyone breaks down they're on their own until they get hit.

[jackal]

I do think the operational point is a reasonable one. But it seems to support ALR, which is less reliant on technology, rather than DHS - and it's the latter that is the fallback position of the anti-SM crowd as well as a principal result of their campaign (i.e., DHS can't now be converted to ALR as was contracted).

Surely it's the other way around. DHS 'fails safe' in that the hard shoulder is (unless you don't understand the highway code) closed if all the overheads are off or failed. ALR 'fails flowing' in that it has the maximum capacity when all signals are off (usually 4 lanes at NSL) but with the trade off that if anyone breaks down they're on their own until they get hit.

ALR failure - it's just a normal D4, i.e., fails safe by any normal measure.

DHS failure - some think hard shoulder open, some don't, Highway Code or no Highway Code.

[nowster]

ALR failure - it's just a normal D4, i.e., fails safe by any normal measure.

There's no such thing as a normal D4.

DHS failure - some think hard shoulder open, some don't, Highway Code or no Highway Code.

I've never seen anyone using the hard shoulder on a DHS scheme when the signs were out. Have you?

[Bryn666]

I do think the operational point is a reasonable one. But it seems to support ALR, which is less reliant on technology, rather than DHS - and it's the latter that is the fallback position of the anti-SM crowd as well as a principal result of their campaign (i.e., DHS can't now be converted to ALR as was contracted).

Surely it's the other way around. DHS 'fails safe' in that the hard shoulder is (unless you don't understand the highway code) closed if all the overheads are off or failed. ALR 'fails flowing' in that it has the maximum capacity when all signals are off (usually 4 lanes at NSL) but with the trade off that if anyone breaks down they're on their own until they get hit.

ALR failure - it's just a normal D4, i.e., fails safe by any normal measure.

DHS failure - some think hard shoulder open, some don't, Highway Code or no Highway Code.

There are very few non-motorway D4s without hard shoulders - and where they do exist they typically are in urban areas with lower speed limits anyway! They also rarely, if ever, have the left hand verge completely hemmed in by barriers so if you do stop you're fully obstructing lane 1. People get onto a motorway, and (wrongly) assume it's a 70+ free flowing experience - and the first 5 decades of motorways conditioned people to believe that obstructions would normally be out of the live lanes. However, user error is only a fraction of the trouble - where has the rest of the problem come about?

1. Lack of coherent information campaigns. They were introduced with minimal fanfare, and unless you were specifically interested, information was hard to come by. Certainly no adverts running during Coronation Street with a "so patronising your dog can follow it" tone as per the 1970s PIFs of old.

2. Contradiction with established principles - we already had lane control signals and these would've been an easy way to operate DHS. Instead, we went for Heath Robinson signals, prism signs, and in the process devalued the point of a solid edge line by saying you can now drive over this. An unlit signal meaning the same as a non-flashing red X on DHS - how are you supposed to know this is a logical position (see point 1) - if the hard shoulder is closed it should be permanently red X, not just when a speed limit is in effect. Green arrows wouldn't go amiss either.

https://farm1.staticflickr.com/749/2061 ... 76a0_b.jpg <- Hong Kong approach to lane control and variable speed limits seems so much more obvious.

3. ALR cost cutting arrogance - "cars don't break down any more so you only need a refuge every 2 miles", yeah that's gone about as well as everyone who said it was a bad idea at the time expected.

4. The technology promised simply hasn't been robust enough.

[traffic-light-man]

...Green arrows wouldn't go amiss either.

https://farm1.staticflickr.com/749/2061 ... 76a0_b.jpg <- Hong Kong approach to lane control and variable speed limits seems so much more obvious.

I've mentioned the green arrows here numerous times before, especially give the equipment is capable, or certainly the options were there to make it so.

I do like the principle of not signing speed limits by lane, though, like in your example from HK. This sort of arrangement tends to be the case in a lot of other countries, it would seem, where overall restrictions or warnings are shown on separate signals/signs from the lane based signals.

I'm not a huge fan of the MS4 lane control method in general, I always think the graphics used for this task look like a bit of a lazy botch. I'd much rather see individual signals for each lane, however I don't think they need to be anywhere near as big or cumbersome as an AMI and thus wouldn't necessarily need the same level of support structure.