The future of smart motorways

The study of British and Irish roads - their construction, numbering, history, mapping, past and future official roads proposals and general roads musings.

There is a separate forum for Street Furniture (traffic lights, street lights, road signs etc).

Registered users get access to other forums including discussions about other forms of transport, driving, fantasy roads and wishlists, and roads quizzes.

Moderator: Site Management Team

Post Reply
KeithW
Member
Posts: 10309
Joined: Thu Dec 04, 2014 13:25
Location: Marton-In-Cleveland North Yorks

Re: The future of smart motorways

Post by KeithW » Mon Dec 28, 2020 08:53

jusme wrote:
Sun Dec 27, 2020 20:17
On the subject of SVD, that seems to me like an ideal application for that over-hyped pattern recognition technology, laughingly called "AI". It could be used to watch all CCTV images simultaneously and in real-time, and track the progress of every vehicle. If any unusual pattern developed a human operator could be alerted to review the incident and take action as appropriate. With this "AI" assistance a small number of operators could cover the entire network 24x7.

Of course it would require > 100% CCTV coverage (which should be a requirement anyway, IMO), and good enough CCTV images, so maybe we'd need to put in (or put back) lighting. That might also help drivers see what's happening better, given that "drive at a speed such that you can stop in the distance you can see" is never going to be a reality, especially on a motorway at night.
Or even simpler low cost radar devices that sense a vehicle stopped in a live lane, they have one rather major advantage over CCTV in that they work even in zero visibilty due to fog or spray. Such small radar transponders are cheap enough and use so little power that they are being fitted to mass market cars. Front facing systems are used in adaptive cruise control and collision avoidance systems while rear facing radars are used by blind spot warning systems.

This is a typical module
https://www.nxp.com/docs/en/fact-sheet/ ... 274_FS.pdf

Micro The Maniac
Member
Posts: 730
Joined: Thu Jun 25, 2015 13:14
Location: B3272/A325/A331

Re: The future of smart motorways

Post by Micro The Maniac » Mon Dec 28, 2020 09:45

Bryn666 wrote:
Sun Dec 27, 2020 14:51
Also how are we giving people ever more "rights" without responsibility when you yourself have complained about lower speed limits and heavier enforcement taking away these "rights"?
I don't consider implementation details (speed limits) to be remotely comparable with "rights"

I have no problem with enforcement of sensible limits... what I object to is the arbitrary cutting of limits (especially where the existing limit is wildly ignored), often coupled with the installation of a camera - perhaps they could install the camera *without* cutting the limit. Equally, I have repeatedly supported the idea of lower speed limits in RESIDENTIAL areas.


Road policing is becoming a black/white camera-driven system, with no discretion... but cameras do not and cannot detect poor driving. Middle lane-hogging, tail-gating, "break-checking" etc etc

Marginally exceeding the speed limit, or the absence of hard-shoulder do not cause the (almost daily) collisions at the M3 junction with the M25... Poor driving standards (poor anticipation, late lane changes, failure to indicate) are to blame. Not HE.
You can't have your cake and eat it.
Who do you think I am? Macron?

jusme
Member
Posts: 65
Joined: Wed Mar 09, 2016 17:51

Re: The future of smart motorways

Post by jusme » Mon Dec 28, 2020 10:29

KeithW wrote:
Mon Dec 28, 2020 08:53
jusme wrote:
Sun Dec 27, 2020 20:17
On the subject of SVD, that seems to me like an ideal application for that over-hyped pattern recognition technology, laughingly called "AI". It could be used to watch all CCTV images simultaneously and in real-time, and track the progress of every vehicle. If any unusual pattern developed a human operator could be alerted to review the incident and take action as appropriate. With this "AI" assistance a small number of operators could cover the entire network 24x7.

Of course it would require > 100% CCTV coverage (which should be a requirement anyway, IMO), and good enough CCTV images, so maybe we'd need to put in (or put back) lighting. That might also help drivers see what's happening better, given that "drive at a speed such that you can stop in the distance you can see" is never going to be a reality, especially on a motorway at night.
Or even simpler low cost radar devices that sense a vehicle stopped in a live lane, they have one rather major advantage over CCTV in that they work even in zero visibilty due to fog or spray. Such small radar transponders are cheap enough and use so little power that they are being fitted to mass market cars. Front facing systems are used in adaptive cruise control and collision avoidance systems while rear facing radars are used by blind spot warning systems.

This is a typical module
https://www.nxp.com/docs/en/fact-sheet/ ... 274_FS.pdf
Cute gadget, but these types of "radar" are rather one-dimensional - nothing like the PPI radar systems used for tracking aircraft, and I can't see how they can easily be used to monitor 4 lanes of fast-moving traffic. It seems to me that going for > 100% CCTV would be easier and more effective than adding > 100% radar coverage that only gives a limited view of what's happening, and still requires the CCTV to confirm if there is an incident or false alarm. Of course I'm not involved with development of this kind of thing, so could be well out of touch with what's actually possible these days.

In zero visibility SVD should be a non-issue. If someone hasn't slowed down because they can't see where they're going they're unlikely to take any notice of the signs anyway!

(I have a feeling that some low cost radar-like system is what we'll get though, as it'll be low cost, and mostly work in the lab. I expect to continue to pass stationary vehicles in live lanes, with nothing on the signs, for as long as I'm still driving, and will use Mk-1 eyeballs to detect them. If only other drivers could do the same...)

KeithW
Member
Posts: 10309
Joined: Thu Dec 04, 2014 13:25
Location: Marton-In-Cleveland North Yorks

Re: The future of smart motorways

Post by KeithW » Mon Dec 28, 2020 12:41

jusme wrote:
Mon Dec 28, 2020 10:29

Cute gadget, but these types of "radar" are rather one-dimensional - nothing like the PPI radar systems used for tracking aircraft, and I can't see how they can easily be used to monitor 4 lanes of fast-moving traffic. It seems to me that going for > 100% CCTV would be easier and more effective than adding > 100% radar coverage that only gives a limited view of what's happening, and still requires the CCTV to confirm if there is an incident or false alarm. Of course I'm not involved with development of this kind of thing, so could be well out of touch with what's actually possible these days.

In zero visibility SVD should be a non-issue. If someone hasn't slowed down because they can't see where they're going they're unlikely to take any notice of the signs anyway!

(I have a feeling that some low cost radar-like system is what we'll get though, as it'll be low cost, and mostly work in the lab. I expect to continue to pass stationary vehicles in live lanes, with nothing on the signs, for as long as I'm still driving, and will use Mk-1 eyeballs to detect them. If only other drivers could do the same...)
They only have too be one dimensional, their job is to detect a stationary object in a live lane not a fast moving aircraft at 40,000 ft. People should indeed slow down in poor visibility, if they did so we would not need an SVD system. Equally to the point such systems can alert the motorway control room to the problem who can set the signs accordingly, Its no consolation to be told that the truck that hit you should have slowed down. Such radar systems work now in moving vehicles with collision avoidance systems, I have one, they are not lab toys nor are they a replacement for CCTV but a supplement. A controller is likely monitoring multiple screens, we know from research that the median time between a vehicle stopping and it being detected is 17 minutes. As of today the control centres rely on information reported from traffic officers as all the existing systems report is that traffic is queueing not that single vehicle has stopped.

See
https://www.snclavalin.com/en/projects/ ... -detection
https://bidstats.uk/tenders/2020/W18/725811215

DB617
Member
Posts: 769
Joined: Mon Dec 25, 2017 00:51
Location: Bristol

Re: The future of smart motorways

Post by DB617 » Mon Dec 28, 2020 21:39

Surely if they were one dimensional, they would only detect vehicles within their linear beam. You would need a lot of them

User avatar
traffic-light-man
Member
Posts: 4187
Joined: Fri Sep 01, 2006 18:45
Location: Liverpool, UK

Re: The future of smart motorways

Post by traffic-light-man » Mon Dec 28, 2020 22:00

This page on the Navtech website gives some good information on the specs of the radar detectors that HE are using, as well as a video explaining how the data is used by the software.

Simon :driving:

Micro The Maniac
Member
Posts: 730
Joined: Thu Jun 25, 2015 13:14
Location: B3272/A325/A331

Re: The future of smart motorways

Post by Micro The Maniac » Tue Dec 29, 2020 07:50

KeithW wrote:
Mon Dec 28, 2020 12:41
People should indeed slow down in poor visibility, if they did so we would not need an SVD system.
My point exactly!

User error.

KeithW
Member
Posts: 10309
Joined: Thu Dec 04, 2014 13:25
Location: Marton-In-Cleveland North Yorks

Re: The future of smart motorways

Post by KeithW » Tue Dec 29, 2020 08:45

Micro The Maniac wrote:
Tue Dec 29, 2020 07:50
KeithW wrote:
Mon Dec 28, 2020 12:41
People should indeed slow down in poor visibility, if they did so we would not need an SVD system.
My point exactly!

User error.
Humans are not reliable - this is lesson number one in engineering for safety, this is why we have had safety systems on railways for a very long time. The early railways used manual signalling systems operated by humans, this ended up killing people so systems such as token block were introduced. If car drivers were perfect we would not need seat belts and air bags. I became painfully aware of this when at the age of 16 riding along a dry road in broad daylight I was hit by a car and subsequently spent 6 months in hospital having my right leg rebuilt. If you want to know what the results of not acting on this basis consider the road casualty rate in 1938 when with just 2.7 million vehicles on the road in the UK there were 8,500 people killed in RTC's. The ethos at the time was exactly what you are espousing , in 1931 speed limits had been abolished. This turned out to be a mistake as road casualties soared and from 1935 onwards they were reintroduced.

Consider also what happened in 1971 on the M1 when the fog warning lights were not switched on in time. Just north of the A505 junction near Luton in fog 70 vehicles were involved in a pileup that killed 7 people and injured another 42. 26 miles of the M1 motorway were closed for several hours in both directions. All it takes is one idiot driving too fast in poor visibility for some sort of herd instinct to take over and pretty soon you have a closely packed convoy of vehicles driving at speeds that guarantee they would not be able to stop if anything goes wrong.

In commercial aviation we have 2 pilots monitoring each other and we make them follow checklists to prepare everything for major changes such as takeoff and landing as well as air traffic controllers to maintain separation.

At the domestic level its why we have kettles that turn themselves off if they boil dry along with fire and smoke detectors. A high percentage of domestic fires are caused by people leaving a chip pan on the stove for so long it catches fire.

User avatar
Barkstar
Member
Posts: 1462
Joined: Sun Jul 14, 2013 16:32

Re: The future of smart motorways

Post by Barkstar » Tue Dec 29, 2020 10:20

Micro The Maniac wrote:
Tue Dec 29, 2020 07:50
KeithW wrote:
Mon Dec 28, 2020 12:41
People should indeed slow down in poor visibility, if they did so we would not need an SVD system.
My point exactly!

User error.
There is operator error to be considered here as well. Like I said such a system should be fail safe and this one doesn't.

User avatar
Bryn666
Member
Posts: 30122
Joined: Thu Jun 20, 2002 20:54

Re: The future of smart motorways

Post by Bryn666 » Tue Dec 29, 2020 12:07

Barkstar wrote:
Tue Dec 29, 2020 10:20
Micro The Maniac wrote:
Tue Dec 29, 2020 07:50
KeithW wrote:
Mon Dec 28, 2020 12:41
People should indeed slow down in poor visibility, if they did so we would not need an SVD system.
My point exactly!

User error.
There is operator error to be considered here as well. Like I said such a system should be fail safe and this one doesn't.
Pilot error contributed to the worst non-terrorism related aviation disaster in 1977. Instead of shrugging it off and saying "should have been more careful", we undertook steps to ensure that pilots couldn't act unilaterally because of their rank, and that the multi-crew approach Keith referred to became mandatory.

It's driver entitlement that says "oh well, common sense will prevail". The common man hasn't got any, that's why we have rules and regulations. There's nothing more tedious than this libertarian "do what you want, **** everyone else" approach to using the roads.

But if we are going to introduce systems that run contrary to what we've learned over 60 years of motorway operations - e.g. hard shoulders are now there for driving down - the systems and regulations in place to enable this have to be watertight.

The fact they're not is extremely alarming.

It would not have been difficult to institute a system of green arrows over every lane that is open as is done on every other country's hard shoulder running systems. This "blank means don't" is not clear enough. It should be a positive instruction by way of green arrow or negative instruction by way of a red cross.

If there's a power failure? Well that's what happens if you rely purely on electronics and don't have some kind of back up. In the event of a power failure you just have a rule that then says don't.

Or... most radical of all... we should never have introduced dynamic hard shoulder running and just done ALR from the start.
Bryn
Traffic/Road Safety Dogsbody and General Grumpy Now-a-Thirtysomething Man
She said life was like a motorway; dull, grey, and long.

Blog - https://showmeasign.online/

jusme
Member
Posts: 65
Joined: Wed Mar 09, 2016 17:51

Re: The future of smart motorways

Post by jusme » Tue Dec 29, 2020 12:11

Here's a screenshot from the Clearway demo. The red marks are the processed view:
clearway-1.jpg

This is a pattern recognition ("AI") view with vehicles identified:
ai-1.jpg
https://www.youtube.com/watch?v=_i4numqiv7Y

Night and day, in terms of what can be seen.

I really think the radar solution is not fit for purpose. Sure, it's better than nothing, but that smacks of "something must be done... this is something..." rather than actually addressing the issue. Maybe that's the point?

User avatar
traffic-light-man
Member
Posts: 4187
Joined: Fri Sep 01, 2006 18:45
Location: Liverpool, UK

Re: The future of smart motorways

Post by traffic-light-man » Tue Dec 29, 2020 12:27

Bryn666 wrote:
Tue Dec 29, 2020 12:07
It would not have been difficult to institute a system of green arrows over every lane that is open as is done on every other country's hard shoulder running systems. This "blank means don't" is not clear enough. It should be a positive instruction by way of green arrow or negative instruction by way of a red cross.

If there's a power failure? Well that's what happens if you rely purely on electronics and don't have some kind of back up. In the event of a power failure you just have a rule that then says don't.
I agree with this entirely, and it has been my main problem from the start. We've set up a system which is reliant on technology that doesn't fail safe.

Yes, regular motorways can suffer the same failure in technology, but the hard shoulder is (often) available for emergencies (live lane issues notwithstanding) and while the signals not being active is generally a sign that all is well, it isn't necessarily seen as much of a positive message as is the case on the Smart Motorway.

I think the 'marketing' behind it is partially to blame for that, too. The 'we will tell you when something's wrong using signals' approach to describing the fundamental operation of the motorway is clearly describing a safety standard that isn't always possible.

I'm not trying to absolve drivers of fault, I just think - as others have pointed out already - we've actively encouraged the removal of some of that driver responsibility and replaced it with technology that isn't (yet) what it was made out to be.

Edit to save a double-post:
jusme wrote:
Tue Dec 29, 2020 12:11
This is a pattern recognition ("AI") view with vehicles identified:
ai-1.jpg
https://www.youtube.com/watch?v=_i4numqiv7Y
Is that being used for SVD or is it just flow monitoring? The AI doesn't seem to follow the vehicles all that far, and ultimately, what happens if the camera becomes obscured or moved? At least a 360º radar will generally be less susceptible to those issues, as I understand it.

Liverpool was a big user of the FLIR video detection in the urban environment for a little while, and while they're a great idea, they're often seen facing the wrong way or otherwise obscured, which ultimately renders them completely useless. Thankfully, in most cases, they've been backed up by microwave or radar detection, too.
Simon :driving:

Bendo
Member
Posts: 1584
Joined: Sat Aug 11, 2007 02:52
Location: Liverpool

Re: The future of smart motorways

Post by Bendo » Tue Dec 29, 2020 19:21

jusme wrote:
Tue Dec 29, 2020 12:11
Here's a screenshot from the Clearway demo. The red marks are the processed view:
clearway-1.jpg


This is a pattern recognition ("AI") view with vehicles identified:
ai-1.jpg
https://www.youtube.com/watch?v=_i4numqiv7Y

Night and day, in terms of what can be seen.
Night and day, good choice of words. How does your AI based one look in the night then. Maybe with some rain throw in for good measure.

Seems a pointless debate though, HE have already successfully trialed radar and got the results they expected. It just needs to be rolled out to the many miles of ALR motorway that should never have been constructed without it in the first place.

The other thing that seems to be missing that HE should probably consider is an advertising campaign, not too dissimilar to its one about Red Xs, but instead pointing out that the lack of a red x doesn't mean that the road is clear. I suspect the vast majority of the public think that these so called smart motorways are actually smart and backed up by reliable detection systems.

jervi
Member
Posts: 999
Joined: Thu Aug 17, 2017 16:29
Location: Haywards Heath

Re: The future of smart motorways

Post by jervi » Tue Dec 29, 2020 19:48

https://www.kentonline.co.uk/malling/ne ... ar-239938/

M20 SM is getting SVD installed next month with construction starting from the 4th Jan. I'd expect installation shouldn't take too long, maybe a month or two for the whole motorway.

Micro The Maniac
Member
Posts: 730
Joined: Thu Jun 25, 2015 13:14
Location: B3272/A325/A331

Re: The future of smart motorways

Post by Micro The Maniac » Wed Dec 30, 2020 08:33

Bryn666 wrote:
Tue Dec 29, 2020 12:07
Pilot error contributed to the worst non-terrorism related aviation disaster in 1977.
And pilot error is still the primary cause of aircraft incidents.
It's driver entitlement that says "oh well, common sense will prevail". The common man hasn't got any, that's why we have rules and regulations. There's nothing more tedious than this libertarian "do what you want, **** everyone else" approach to using the roads.
You need to change this "driver entitlement" record, it's getting more than tedious.

No-one is suggesting "do what you want" - we're saying put *sensible* rules in place, and enforce them. But think about The Rules, rather than knee-jerk (or political).

Now for the typical Bryn tangential shift...
But if we are going to introduce systems that run contrary to what we've learned over 60 years of motorway operations - e.g. hard shoulders are now there for driving down
On All-Lane Running they are not hard-shoulders, they are running lanes.

However, I agree with you, regarding non-indication on Dynamic Hard Shoulder Running - the lane should always have something showing (Red X or a speed-limit).

Equally, at the end of a restriction, there should be a NSL/End - too often (and very common on the M3 ALR, a blank gantry signals the end of restrictions which is confusing (and, I believe, contrary to The Rules)
Or... most radical of all... we should never have introduced dynamic hard shoulder running and just done ALR from the start.
Hindsight is a wonderful concept

User avatar
Bryn666
Member
Posts: 30122
Joined: Thu Jun 20, 2002 20:54

Re: The future of smart motorways

Post by Bryn666 » Wed Dec 30, 2020 10:02

Hindsight? Several of us expressed reservations about dynamic hard shoulder running in 2005.

But whatever, you know best.
Bryn
Traffic/Road Safety Dogsbody and General Grumpy Now-a-Thirtysomething Man
She said life was like a motorway; dull, grey, and long.

Blog - https://showmeasign.online/

Bomag
Member
Posts: 588
Joined: Tue Jan 26, 2010 23:26

Re: The future of smart motorways

Post by Bomag » Wed Dec 30, 2020 14:49

Bryn666 wrote:
Wed Dec 30, 2020 10:02
Hindsight? Several of us expressed reservations about dynamic hard shoulder running in 2005.

But whatever, you know best.
When M42ATM was being developed, by people who could count, it was clear that the highest risk of people being wiped out at speed was where there was low flows and higher speeds. By having a dynamic hard shoulder, LBS1 use would be limited to where the amount of traffic would limit speed and limit the possibility of high speed shunts. There were negatives including that those groups with an existing low risk would not get any improvement in risk. Switching to ALR is cheaper and has a lower overall risk; the problem is the risk is concentrated on already high risk groups i.e. stopped traffic. This was not acceptable in 2005. Given the lack of consideration on safety and more on cost saving from 2010 to 2015 ALR was just what the head of the smart roads team was looking for to save money and surprise surprise the GD04 assessment said it was 'safe'.

User avatar
EpicChef
Forum Moderator
Posts: 1850
Joined: Wed Aug 03, 2016 11:47
Location: London, UK

Re: The future of smart motorways

Post by EpicChef » Sat Jan 02, 2021 13:03

I’ve been an advocate of green arrow / red X for a while now, with speed limits displayed on the MS4 above. AMI model number 460 can support green arrow but model number 450 can’t. So it might be a hardware thing.

I submitted an FOI a few years ago asking about what was in place on the M1 J10-13 - and the info they gave me made no mention of 450 or 460.

But HE have committed to (hopefully) replacing DHSR with ALR - even if there’s no concrete barrier to begin with, at least they should break the line of the hard shoulder and reconfigure the LBS1 AMI to act as a normal lane rather than dynamic hard shoulder.
Did you know there's more to SABRE than just the Forums?
Add your roads knowledge to the SABRE Wiki today!
Have you browsed SABRE Maps recently? Try getting involved!

User avatar
Chris5156
Member
Posts: 14404
Joined: Thu Sep 06, 2001 21:50
Location: Hampshire
Contact:

Re: The future of smart motorways

Post by Chris5156 » Sat Jan 02, 2021 14:18

I don’t think Smart Motorways were the result of anyone being offended by hard shoulders and demanding their removal. They are the result of traffic levels continuing to rise, the population at large very much preferring to travel by road rather than public transport, but also having a very strong NIMBY mentality that makes road building and road widening difficult and controversial. More road capacity was needed, more road capacity was demanded, but the most expedient and pragmatic way to provide it was to reallocate existing road space.
Last edited by EpicChef on Tue Feb 02, 2021 13:10, edited 1 time in total.
Reason: Quoted post had to be removed

Bomag
Member
Posts: 588
Joined: Tue Jan 26, 2010 23:26

Re: The future of smart motorways

Post by Bomag » Sat Jan 02, 2021 14:50

EpicChef wrote:
Sat Jan 02, 2021 13:03
I’ve been an advocate of green arrow / red X for a while now, with speed limits displayed on the MS4 above. AMI model number 460 can support green arrow but model number 450 can’t. So it might be a hardware thing.

I submitted an FOI a few years ago asking about what was in place on the M1 J10-13 - and the info they gave me made no mention of 450 or 460.

But HE have committed to (hopefully) replacing DHSR with ALR - even if there’s no concrete barrier to begin with, at least they should break the line of the hard shoulder and reconfigure the LBS1 AMI to act as a normal lane rather than dynamic hard shoulder.
The width of an active hard shoulder is not sufficient to give an acceptable preview time at NSL if the hard shoulder line is replaced by a lane line. You would need to run it 24/7 at 60 mph or below. To get to 70 mph you would need a 3.65m lane with the back of hard shoulder line replaced with a 200mm raised rib line.

Post Reply