The future of smart motorways

[DB617]

I think it's hard to argue at this point that the people who designed both DHS and later (and moreso) ALR with or without reduced ERAs, had their heads screwed on properly. My theory is that there was significant pressure from above to make the system appear both viable and safe. Sprinkle in a dash of over-optimism, and engineers' tendency to think that humans behave in a predictable fashion, and you have the current mess.

[Conekicker]

I think it's hard to argue at this point that the people who designed both DHS and later (and moreso) ALR with or without reduced ERAs, had their heads screwed on properly. My theory is that there was significant pressure from above to make the system appear both viable, cheap and safe. Sprinkle in a dash of over-optimism, and engineers' tendency to think that humans behave in a predictable fashion, and you have the current mess.

FTFY.

For a given value of cheap. Obviously.

[SteelCamel]

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.

Supposed to be, certainly. In which case there's no problem with there being a stopped vehicle in a live lane, and ALR is perfectly safe with no "smart" needed - drivers will see the stopped vehicle and stop before hitting it. In practice that isn't the case - so motorway traffic is clearly not driving completely "on sight" in practice. Which is why my suggestion was that, in case of failure of the "smart" systems, the speed limit should default to 40mph - at that speed, there's much more chance that drivers really can stop short of an unexpected obstruction.

[jackal]

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?

"If I haven't personally observed a high risk, low frequency event, it's impossible".

[Big L]

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

Quite often on the M6 north of junction 10 up to 10A.

[Bomag]

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.

Supposed to be, certainly. In which case there's no problem with there being a stopped vehicle in a live lane, and ALR is perfectly safe with no "smart" needed - drivers will see the stopped vehicle and stop before hitting it. In practice that isn't the case - so motorway traffic is clearly not driving completely "on sight" in practice. Which is why my suggestion was that, in case of failure of the "smart" systems, the speed limit should default to 40mph - at that speed, there's much more chance that drivers really can stop short of an unexpected obstruction.

On a normal D3 with hard shoulder there should be nearly always the full 295M SSD (120kph/70mph). HSR sections I am aware of varies, as a minimum, between 160m (80kph/50mph) and 215m (100kph/60mph) SSD. With ALR there are many relaxations on SSD in LBS1 (and occasionally LBS2).

M42 ATM etc was designed to be safe with everything off. This is not 'partially working'. Given the reliably of the technology, going all the way back to 2002, did not allow for 24/7 use, there was never the chance that the system could put out a default 40mph, nor drivers assume a 40 mph limit if every thing was off. In 2005 MTBY for a AMI in use up to 18 hours a day was supposed to be 5 years; on 24/7 it was calculated to be barely 6 months.

It has to be remembered that ATM/MM/SM were never safety schemes. Also ALR was driven by the technology specialists and not highway safety standards specialists. Without being smug, they were told.

[SouthWest Philip]

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

Quite often on the M6 north of junction 10 up to 10A.

Yes, I have often seen cars using the the hard shoulder when closed when using the M6 and that's just on the dozen of so times a year I might find myself that way. It isn't even unheard of to now to see cars using the hard shoulder on sections of motorway that aren't DHS-running.

I am firmly of the opinion that the solid line needs to return to having an unambiguous 'do not cross' status. Either have a separate marking for DHS schemes of standardise ALR.

[Chris5156]

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?

Yes, multiple times. I’ve seen people using the hard shoulder on non-DHS motorways close to DHS sections in heavy congestion too. I’ve also seen how much less used the hard shoulder is when open compared to the other lanes. I would conclude that a significant minority of drivers don’t know, or aren’t paying enough attention, to understand whether the hard shoulder is open or closed.

[domcoop]

The one thing the anti-smart-motorway brigade like Panorama and Clare Mercer, etc., don't seem to say is what they think the alternative is.

I assume they, and the general public who listen along and say "there should be a hard shoulder" think and expect that an entirely new hard shoulder lane will be built alongside the running lanes? Would there be as much support to "end Smart motorways" if people were told the result would be just one existing lane taken out and dedicated to hard shoulder use? (Obviously the DHS sections are like this much of the time, but the ALR sections you'd effectively knock 25 to 33% of the capacity out overnight).

[DB617]

The one thing the anti-smart-motorway brigade like Panorama and Clare Mercer, etc., don't seem to say is what they think the alternative is.

I assume they, and the general public who listen along and say "there should be a hard shoulder" think and expect that an entirely new hard shoulder lane will be built alongside the running lanes? Would there be as much support to "end Smart motorways" if people were told the result would be just one existing lane taken out and dedicated to hard shoulder use? (Obviously the DHS sections are like this much of the time, but the ALR sections you'd effectively knock 25 to 33% of the capacity out overnight).

This is the big issue really. The economic and social impact of just removing them and permitting congestion to massively outstrip pre-implementation levels (induced demand on these sections is very real, and demand evaporation won't be cost-free) would be enormous. The cost of putting things right, by standardising on one design implementing a lot of the suggestions here and securing long term funding for properly staffed and maintained systems, surely is smaller than the cost of restoring intermittent (by necessity) hard shoulders if that's even possible, or the aforementioned indirect costs. The problem is, I'm thinking like a long term oriented director, not a fourth-rate collection of populist politicians.

[Bomag]

The one thing the anti-smart-motorway brigade like Panorama and Clare Mercer, etc., don't seem to say is what they think the alternative is.

I assume they, and the general public who listen along and say "there should be a hard shoulder" think and expect that an entirely new hard shoulder lane will be built alongside the running lanes? Would there be as much support to "end Smart motorways" if people were told the result would be just one existing lane taken out and dedicated to hard shoulder use? (Obviously the DHS sections are like this much of the time, but the ALR sections you'd effectively knock 25 to 33% of the capacity out overnight).

Going from 3 to 4 lanes does not increase capacity by 33%. A three lane controlled motorway, with the benefit of maintaining capacity longer as traffic increases, is much closer to D4M than D3M and is more resilient to stopped traffic (if it used the hard shoulder). The near works will be interesting as to residual capacity.

[Peter Freeman]

... you'd effectively knock 25 to 33% of the capacity out overnight).

Going from 3 to 4 lanes does not increase capacity by 33%.

Whether that statement is true or not requires clarity on what "going from" means. In the simplest sense, disregarding the SM issue, it certainly does increase capacity by at least 33%.

We've had this discussion before, where, for some perverse reason, a 'law of diminishing returns' is postulated for added lanes. In other words, saturated capacity of a carriageway is less than proportional to its number of lanes. My claim in that previous discussion was/is that the relationship is actually more than proportional. So a D3's capacity is more than 3/2 times that of a D2's. A D4's capacity is more than 4/3 x a D3's. D4 is more than double D2. Etc.

I won't go through the whole argument again here, but, briefly, left-most lanes operate in an inefficient 'lane-changing' state because they're influenced by exits and entrances, whereas right-most lanes are insulated and flow un-interruptedly.

A three lane controlled motorway, with the benefit of maintaining capacity longer as traffic increases, is much closer to D4M than D3M and is more resilient to stopped traffic (if it used the hard shoulder). The near works will be interesting as to residual capacity.

There you're into more complicated issues. However, Domcoop is right that if you de-ALR by changing Lane 1 back to being a hard shoulder, leaving 3 running lanes, you will lose at least 25% capacity.

[Conekicker]

The one thing the anti-smart-motorway brigade like Panorama and Clare Mercer, etc., don't seem to say is what they think the alternative is.

I assume they, and the general public who listen along and say "there should be a hard shoulder" think and expect that an entirely new hard shoulder lane will be built alongside the running lanes? Would there be as much support to "end Smart motorways" if people were told the result would be just one existing lane taken out and dedicated to hard shoulder use? (Obviously the DHS sections are like this much of the time, but the ALR sections you'd effectively knock 25 to 33% of the capacity out overnight).

And yet on the ALR sections where extra ERAs are being installed, a lane is taken out for the duration of the construction period. In the case of the M1 in South Yorkshire, that's around 18 months, up to "Winter 2024". Might one suggest that for that length of time one might have been able to construct a hard shoulder or two?

[Micro The Maniac]

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.

I've seen enough vehicles driving up hard-shoulders on "normal" motorways, after the end of ALR/DHS, to suggest that DHS was the worst idea ever.

Drivers have got into the habit of driving on the hard-shoulder...

[Micro The Maniac]

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.

Not sure what you mean by that - there are plenty of stretches of non-ALR/DHS D4... the M1 and M25 being two obvious motorways - in fact the M25 had D5 and even D6 on its western stretch.

[Chris5156]

And yet on the ALR sections where extra ERAs are being installed, a lane is taken out for the duration of the construction period. In the case of the M1 in South Yorkshire, that's around 18 months, up to "Winter 2024". Might one suggest that for that length of time one might have been able to construct a hard shoulder or two?

An objective of “pave every place where there’s enough room to put a hard shoulder in” would have resulted in more opportunities to pull over than the present work to build a small number of extra gold-plated ERAs with extravagant earthworks and retaining structures. But that would have been logical, and the last remaining traces of logic departed the government’s Smart Motorway policy a while ago now.

[jnty]

And yet on the ALR sections where extra ERAs are being installed, a lane is taken out for the duration of the construction period. In the case of the M1 in South Yorkshire, that's around 18 months, up to "Winter 2024". Might one suggest that for that length of time one might have been able to construct a hard shoulder or two?

An objective of “pave every place where there’s enough room to put a hard shoulder in” would have resulted in more opportunities to pull over than the present work to build a small number of extra gold-plated ERAs with extravagant earthworks and retaining structures. But that would have been logical, and the last remaining traces of logic departed the government’s Smart Motorway policy a while ago now.

I have to wonder whether extremely discontinuous hard shoulders do more harm than good, though. At what point do the benefits to motorists experiencing sudden breakdowns get outweighed by the risks of a distracted motorist nipping into the hard shoulder to go to the toilet only to slam into a bridge parapet at 60?

[ManomayLR]

And yet on the ALR sections where extra ERAs are being installed, a lane is taken out for the duration of the construction period. In the case of the M1 in South Yorkshire, that's around 18 months, up to "Winter 2024". Might one suggest that for that length of time one might have been able to construct a hard shoulder or two?

An objective of “pave every place where there’s enough room to put a hard shoulder in” would have resulted in more opportunities to pull over than the present work to build a small number of extra gold-plated ERAs with extravagant earthworks and retaining structures. But that would have been logical, and the last remaining traces of logic departed the government’s Smart Motorway policy a while ago now.

I have to wonder whether extremely discontinuous hard shoulders do more harm than good, though. At what point do the benefits to motorists experiencing sudden breakdowns get outweighed by the risks of a distracted motorist nipping into the hard shoulder to go to the toilet only to slam into a bridge parapet at 60?

But then the argument is people (ideally) should not be nipping into the hard shoulder to go to the toilet!
I don’t know exactly where I read it, but apparently misuse of ERAs is much lower than misuse of the hard shoulder.

Part of the reason why, will be that with a hard shoulder you just build up speed and merge into traffic smoothly - no attention drawn to you. With ERAs that’s not possible and you would need to contact the Regional Control Centre, and tell them that they now have to close a lane because you needed a wee.

[DB617]

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.

I've seen enough vehicles driving up hard-shoulders on "normal" motorways, after the end of ALR/DHS, to suggest that DHS was the worst idea ever.

Drivers have got into the habit of driving on the hard-shoulder...

Well this raises the question of what level of driver ignorance you are willing to design for. For example - should we avoid designing urban roads that are unsafe at 30mph because we know drivers will probably do 28 if we make the speed limit 20?

A much more robust red X enforcement program would have been very beneficial IMO. The fact there basically wasn't one in the beginning and now there is but only at very specific gantries, is a wasted opportunity to educate the less attentive or get them off the road with points.

[jnty]

An objective of “pave every place where there’s enough room to put a hard shoulder in” would have resulted in more opportunities to pull over than the present work to build a small number of extra gold-plated ERAs with extravagant earthworks and retaining structures. But that would have been logical, and the last remaining traces of logic departed the government’s Smart Motorway policy a while ago now.

I have to wonder whether extremely discontinuous hard shoulders do more harm than good, though. At what point do the benefits to motorists experiencing sudden breakdowns get outweighed by the risks of a distracted motorist nipping into the hard shoulder to go to the toilet only to slam into a bridge parapet at 60?

But then the argument is people (ideally) should not be nipping into the hard shoulder to go to the toilet!
I don’t know exactly where I read it, but apparently misuse of ERAs is much lower than misuse of the hard shoulder.

Well, quite. Motorists have very different preconceived ideas around how hard shoulders operate versus how ERAs operate, and will therefore, as you say, use them differently. If we make yet another change to motorway norms, and turn motorway hard shoulders into something only slightly better than the wavy shoulder lines that you get on S2s which vary wildly according to the terrain, the safety calculus is so finely balanced that it could easily cause more accidents through confusion and misuse. We do have discontinuous hard shoulder on the motorway network already, but it's either a clear and long-lasting absence, usually with laybys, or continuous-with-occasional-interruptions like bridges. It feels as though people are implicitly arguing for a move to a system of indeterminately long and irregularly placed shallow laybys, which I think would 'look' better but perform worse.