The future of smart motorways

[mikehindsonevans]

Perhaps motorway limits when the road is unlit should be dropped to 30 at night then just in case there is someone walking along?

Hear, hear!

[mikehindsonevans]

Updated that for you!

You are supposed to leave a 2-second gap. 2 metres is barely half a car length.

No, I know that, I was making a joke about social distancing.

More specifically, to an instance where in a queue at a store, I left a two-meter gap with the person in front of me, and sure enough, someone jumped the queue. and stood in front of me by slotting into the gap.

"EpicChef" : at least *I* appreciated the pun!

[Enceladus]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

[ManomayLR]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

I severely doubt that. Various safety improvements have been committed to be made, though on an awfully slow timescale.

But ditching capacity upgrades or opting for full carriageway widening has social, economic, and environmental impacts too, and the Government will likely only ditch smart motorways as a last resort.

Full carriageway widening takes more space and is way costlier and lengthier. It causes more disruption during works, and will almost inevitably mean that fewer road stretches would benefit from capacity upgrades.

This, in turn, means more congestion, pollution, and frustration for drivers, on stretches that ended up losing out on extra capacity and technology.

It's a balancing act, and the Government isn't likely to hugely favor one side over the other.

[Bendo]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

No, with properly working SVD there is no reason it can't be done safely. If they had just done it from the start, it's unlikely the debate would be happening now.

[Micro The Maniac]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

Despite all the furore, the numbers do not stack up... sure there have been a small number of (high-profile) fatal stopped-vehicle collisions and fatalities, but these are (probably) matched by a similar number of hard-shoulder collisions and fatalities prevented.

On the other hand, given that the smartified A14 remained just that - not becoming A14(M) - and the LTC is now planned as a smartified three-lane A221 rather than as the Mxx, the concept clearly lives on, albeit under a green sign, rather than blue.

[KeithW]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

No, with properly working SVD there is no reason it can't be done safely. If they had just done it from the start, it's unlikely the debate would be happening now.

They were not ready from the start, work on systems that could be deployed in large numbers only really started in about 2008, earlier attempts with systems based on CCTV and induction loop systems had real problems. The number of induction loops needed ruled that out and optical systems dont work well in poor visibility. So the first thing that was needed was small high discrimination radars that could be produced at an affordable cost, the work done on such radars for automotive collision avoidance systems and adaptive cruise control helped there.

You are not done though when you have a suitable radar. The return from any such system mounted alongside or above a multilane road will show a large number of stationary objects that should be there, Armco, CCTV masts, SOS phones, roadside cabinets etc. So the next trick is to come up with heuristic software that can learn which stationary objects are part of the infrastructure and those which are not. Trying to manually add them to a memory map was just not practical. To complicate things farther it has to avoid learning that the stalled vehicle is part of the furniture as it were. Some of the early systems tried involved sensors designed for selfdriving cars but the combination of lidar and radar systems were just too complex. Once a system evolved that was deployable trials were carried out on a number of roads including freeways in California and the M25. This identified problems that had to solved before the system could be rolled out for real life use.

Its not just Smart Motorways that are the problem of course, ordinary motorways see a large number of collisions with vehicles stopped on the hard shoulder being struck every year. Then we have all the major A roads such as the A1,A11,A12,A13 etc.

In fact the development of affordable self driving cars has also progressed as a result of this work. Early systems such as that pioneered by Google required onboard memory maps of the road that had to be manually surveyed and loaded. With the new techniques the system could update stationary object maps dynamically. After motorway maintenance a vehicle can drive down a stretch of road and the operator can accept a 'new' return and it may be added to the known static obstacle map.

I was not involved in this project but a number of my colleagues were.

[c2R]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

I severely doubt that. Various safety improvements have been committed to be made, though on an awfully slow timescale.

But ditching capacity upgrades or opting for full carriageway widening has social, economic, and environmental impacts too, and the Government will likely only ditch smart motorways as a last resort.

Full carriageway widening takes more space and is way costlier and lengthier. It causes more disruption during works, and will almost inevitably mean that fewer road stretches would benefit from capacity upgrades.

But, if full widening is provided the hard shoulder can then be used for future maintenance activities, to facilitate contraflows, and to assist the emergency services in getting to accidents.

Of course, there's a scale of these things; intermittent hard shoulders are cheaper but lose some of the "works" advantages; soft shoulders are cheaper again.

Unfortunately, what we've had with these ALR upgrades is an enormous amount of penny pinching, so that infrastructure is placed right up to the edge of the road, along with crash barriers, meaning that there is no space to even do a bit of earthworks to create soft hard shoulders and move the barriers back.

[ManomayLR]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

I severely doubt that. Various safety improvements have been committed to be made, though on an awfully slow timescale.

But ditching capacity upgrades or opting for full carriageway widening has social, economic, and environmental impacts too, and the Government will likely only ditch smart motorways as a last resort.

Full carriageway widening takes more space and is way costlier and lengthier. It causes more disruption during works, and will almost inevitably mean that fewer road stretches would benefit from capacity upgrades.

But, if full widening is provided the hard shoulder can then be used for future maintenance activities, to facilitate contraflows, and to assist the emergency services in getting to accidents.

Of course, there's a scale of these things; intermittent hard shoulders are cheaper but lose some of the "works" advantages; soft shoulders are cheaper again.

Unfortunately, what we've had with these ALR upgrades is an enormous amount of penny pinching, so that infrastructure is placed right up to the edge of the road, along with crash barriers, meaning that there is no space to even do a bit of earthworks to create soft hard shoulders and move the barriers back.

Interestingly, when they widened M1 J6A-10, they did full reconstruction of bridges etc. and a continuous hard shoulder.

That means if needed in the future, the route can be converted to D5ALR too.

[mikehindsonevans]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

No, with properly working SVD there is no reason it can't be done safely. If they had just done it from the start, it's unlikely the debate would be happening now.

They were not ready from the start, work on systems that could be deployed in large numbers only really started in about 2008, earlier attempts with systems based on CCTV and induction loop systems had real problems. The number of induction loops needed ruled that out and optical systems dont work well in poor visibility. So the first thing that was needed was small high discrimination radars that could be produced at an affordable cost, the work done on such radars for automotive collision avoidance systems and adaptive cruise control helped there.

You are not done though when you have a suitable radar. The return from any such system mounted alongside or above a multilane road will show a large number of stationary objects that should be there, Armco, CCTV masts, SOS phones, roadside cabinets etc. So the next trick is to come up with heuristic software that can learn which stationary objects are part of the infrastructure and those which are not. Trying to manually add them to a memory map was just not practical. To complicate things farther it has to avoid learning that the stalled vehicle is part of the furniture as it were. Some of the early systems tried involved sensors designed for selfdriving cars but the combination of lidar and radar systems were just too complex. Once a system evolved that was deployable trials were carried out on a number of roads including freeways in California and the M25. This identified problems that had to solved before the system could be rolled out for real life use.

Its not just Smart Motorways that are the problem of course, ordinary motorways see a large number of collisions with vehicles stopped on the hard shoulder being struck every year. Then we have all the major A roads such as the A1,A11,A12,A13 etc.

In fact the development of affordable self driving cars has also progressed as a result of this work. Early systems such as that pioneered by Google required onboard memory maps of the road that had to be manually surveyed and loaded. With the new techniques the system could update stationary object maps dynamically. After motorway maintenance a vehicle can drive down a stretch of road and the operator can accept a 'new' return and it may be added to the known static obstacle map.

I was not involved in this project but a number of my colleagues were.

Anyone remember the Nimrod AEW aircraft programme? At one point, the RADAR systems were flooded with lots of returns at zero altitude, "flying" at zero altitude.

So I can appreciate your point that SVD should be calibrated. But, surely, *before* you start building a new type of motorway which kills people if they stop?

And yes, I would prefer real widened motorways - the land take is bog-all as a percentage of the land-mass of this country. Additionally the building work is, naturally, a lane further away from the traffic flows. This protects the construction crews.

Just my opinion, natch.

[FosseWay]

And yes, I would prefer real widened motorways - the land take is bog-all as a percentage of the land-mass of this country.

Except possibly in urban/semi-urban contexts where it may involve compulsory purchase of things that people tend to get irritated by the loss of - their own homes of course, but also other community assets. In these cases perhaps a compromise would be to use the smart system but either reduce the maximum speed ever permitted to 50 or make greater use of lower posted limits at times of high density traffic. The total mileage affected wouldn't be huge, and the annoyance that would be predictably caused by the speed limit cut would be offset by the fact that outside these areas, the speed limit would remain at 70 more of the time with a full hard shoulder available.

[KeithW]

Anyone remember the Nimrod AEW aircraft programme? At one point, the RADAR systems were flooded with lots of returns at zero altitude, "flying" at zero altitude.

So I can appreciate your point that SVD should be calibrated. But, surely, *before* you start building a new type of motorway which kills people if they stop?

And yes, I would prefer real widened motorways - the land take is bog-all as a percentage of the land-mass of this country. Additionally the building work is, naturally, a lane further away from the traffic flows. This protects the construction crews.

Just my opinion, natch.

The elephant in the room is cost. there simply is not the money around to widen all the motorways we would like. A secondary point is with many older motorways widening is going to be difficult and the end result may well have to be a rather expensive off line upgrade as happened with most of the A1 north of Doncaster, this sort of problem is why the Doncaster bypass and Redhouse to Darrington upgrades were pushed back into RIS 3. There is also the issue of the disruption during widening works.

The motorway does NOT kill people, inattentive drivers do that and they also do that on motorways with hard shoulders and fast D2 roads with no hard shoulders so even without Smart Motorways these are systems that could save lives, be affordable and installation would be relatively simple. Hard. shoulders are in fact so dangerous that the advice given to drivers stuck on them is to get out of the car and get behind the barrier.

Motorways, smart or dumb are already the safest roads in the country. From a safety point of view the roads that frighten the c**p out me are the fast busy A roads without hard shoulders or refuges such as the A1, A12, A14, A14, A19, A34, A50 etc

This is the road that tops my list, carries huge amounts of traffic, few if any smart signs or management and there isnt even an Armco barrier to get behind if you are stuck. Its no wonder its regarded as one of the most dangerous roads in the country.
https://www.google.co.uk/maps/@51.80953 ... authuser=0

If we look at the situation in England we have 1,941 miles of motorway, 2,418 miles of Trunk A roads with another 12,101 miles of major A Roads. Motorways are the stars but the other roads are just as important and considerably less safe.

[Debaser]

Interestingly, when they widened M1 J6A-10, they did full reconstruction of bridges etc. and a continuous hard shoulder.

Which is possibly one reason why the design of the next section north, J10 to J13, was changed at virtually the last minute from full traditional widening to the managed motorway layout it has now. That sort of work costs significantly more.

[c2R]

6A to 10 is also a different kettle of fish, particularly at the southern end with the merge with the M25 and the 7/8 collector distributor lanes.

Many of the structures on this section also carry the motorway over (rather than being a bridge over the motorway). These were typically altered as widening of the existing structure
https://www.google.com/maps/@51.7878111 ... 384!8i8192 OR tubes were inserted through a structure and it backfilled: https://www.google.com/maps/@51.7866204 ... 760!8i2880

... rather than demolition and building a new replacement. Some of the structures, such as the M10 and A4147 overbridges had already been demolished and re-built as part of earlier widening schemes.

[Bendo]

So, given the less than stellar safety issues associated with the so-called “smart” motorways, are we to take it that the concept is now a dead duck?

No, with properly working SVD there is no reason it can't be done safely. If they had just done it from the start, it's unlikely the debate would be happening now.

They were not ready from the start, work on systems that could be deployed in large numbers only really started in about 2008, earlier attempts with systems based on CCTV and induction loop systems had real problems. The number of induction loops needed ruled that out and optical systems dont work well in poor visibility. So the first thing that was needed was small high discrimination radars that could be produced at an affordable cost, the work done on such radars for automotive collision avoidance systems and adaptive cruise control helped there.

If the technology didn't exist, then the inquiries are going to be interesting. How did they model the time it would take to detect a stopped vehicle and was it accurate. If the projections were accurate, how was it deemed acceptable for 17 minutes to detect and close a lane acceptable?

If the projections were inaccurate and shown to be so on the initial schemes, their reasoning for continuing to implement ALR will be ineresting.

I suspect those in power at the time that signed off on it now will be feeling a bit uncomfortable for ploughing on regardless. Hopefully, and it does seem to be the case, the roll out of SVD can be done fairly quickly, ideally before a scabby old minibus full of kids suffers a full electrical failure at night on an unlit section.

[Stevie D]

Which is possibly one reason why the design of the next section north, J10 to J13, was changed at virtually the last minute from full traditional widening to the managed motorway layout it has now. That sort of work costs significantly more.

Traffic levels are about 10–15% higher south of J10 than north of it, so it doesn't seem unreasonable that that might have been the threshold at which it was worthwhile going for the more expensive option of full widening up to J10 while the slightly lower traffic volumes beyond didn't justify it.

[thatapanydude]

I wonder if any consideration has been taken to ALR motorways in closing lane 1 in the evening/early morning off-peak 10pm to 5am - similar as a dynamic hard shoulder but less clunky as the layout is standardised. After 10pm on all ALR smart motorways the demand is not there to have all 4 lanes open so why not close lane 1 or lane 4 or both so it acts as a hard shoulder but should only be used if you cannot reach a refuge area.

[Bendo]

I wonder if any consideration has been taken to ALR motorways in closing lane 1 in the evening/early morning off-peak 10pm to 5am - similar as a dynamic hard shoulder but less clunky as the layout is standardised. After 10pm on all ALR smart motorways the demand is not there to have all 4 lanes open so why not close lane 1 or lane 4 or both so it acts as a hard shoulder but should only be used if you cannot reach a refuge area.

Lets hope not, DHSR is hard enough for some people to understand but doing it on a lane marked out as a standard running lane is just asking for trouble.

[KeithW]

I wonder if any consideration has been taken to ALR motorways in closing lane 1 in the evening/early morning off-peak 10pm to 5am - similar as a dynamic hard shoulder but less clunky as the layout is standardised. After 10pm on all ALR smart motorways the demand is not there to have all 4 lanes open so why not close lane 1 or lane 4 or both so it acts as a hard shoulder but should only be used if you cannot reach a refuge area.

Varying the operations like this is more likely to cause accidents than improve things in my opinion, drivers are creatures of habit. This is why you sometimes see signs saying things like 'New Road Layout Ahead'

[ManomayLR]

I’ve always said that DHSR would have worked if they’d used a proper system of green arrows and red Xs.