UK government: no more smart motorways without stopped car detection

The concern is that the outside lane doesn’t get converted into a hard shoulder fast enough to provide a safe place to stop.

There's a few things that fall under the "smart motorway" umbrella. They also have systems to detect traffic flow and impose variable speed limits when approaching a queue, and usually regular speed cameras to enforce those limits.

This stopped car detection thing is indeed about the ones that can open the hard shoulder for traffic, though.

Exactly - basically there used to be 2/3 lanes and a hard shoulder, and they changed the road to 3/4 lanes and no hard shoulder.

Because branding them “motorways without a hard shoulder” sounded unsafe and bad they decided to call the proposal “smart motorways” (Because who would want to argue with that? You want regular stupid motorways?).

The political context here is: There have been four different coroners' reports where smart motorways have contributed to deaths.

Most recently in January [1] after two stopped vehicles - which would otherwise have stopped on a hard shoulder - instead had to stop in a traffic lane and were hit by a truck.

In principle, smart motorways are already supposed to be covered in cameras etc, so people in a control centre can stop the traffic in a lane if the need arises. Apparently it doesn't work all that well in practice, though.

[1] https://www.bbc.co.uk/news/uk-england-south-yorkshire-557087...

Motorways ("freeways" in en-us, I think) are traditionally 3 lanes with a 4th lane as a 'hard shoulder' that is closed to all traffic except those who have broken down and the emergency services.

Smart Motorways were a Tory government idea designed to be "car friendly" and increase apparent road capacity without actually costing much money -- by using already-existing lightable signs above the motorway to convert the hard shoulder into another lane of traffic, and adding in "refuge areas" every $[$RANDOM % 5+1] miles.

The trouble is, if you're in a motorway where every one of the now four lanes of traffic is going at the speed limit of 70 miles an hour (or, more likely, 80 miles an hour) and your tire explodes, you might not be able to coast the requisite distance to a refuge area, and you might not have the wherewithall to adequately pull into what is in effect a small siding about the size of an articulated lorry. If you break down in a "live lane" the signs are supposed to change to move traffic out behind you, so it doesn't ram into the car.

Naturally, this hasn't worked well, and a bunch of people have died. At least one coroner has specifically called out Her Majesty's Government for directly contributing to deaths and blamed the smart motorway concept in general. There's at least one harrowing 999 call I've heard somewhere of someone saying "Help, I've broken down in lane 2 of the M4 and my car has stopped and I couldn't get into lane 1 when <BANG> <sound of young child crying> <no more adult voice>".

Some of the "smarts" in smart motorways in doing lane-light switching this have been provided by humans and/or computer vision on the basis of CCTV. The trouble is, they've been ineffective. This announcement states that they're going to replace the previous approach with mmWave radar that I presume can detect changes more rapidly (quite how that works with shadows and reflections -- and cars in front of lorries, I don't know).

Fundamentally, this seems to be "papering over the cracks" of the idea -- and I unfortunately don't think it will work. Many, if not most, British people want to go back to the old system of having 3 lanes + a hard shoulder. We don't understand why the Tories really, really love smart motorways: it probably comes out of a combination of 'driver friendly' Manifesto promises combined with economic modelling about the cost of traffic jams.

Not all smart motorways have the hard shoulder thing, the ones that do have semi frequent "stopping bays". However, stopping in one without contacting the support staff(there are intercoms) or a good reason is an offence.

They all feature per lane adjustable speed limits, as well as linked speed cameras at frequent intervals. This is supposed to help ease congestion by adjusting flow in certain areas.

In practice random signs are often left on the wrong setting, so you get a speed reduction to 40mph for 250m. Or a lane is closed for 6 miles with no reason.

I've no doubt that in the scenario where there is congestion that it greatly aids safety, as approaching traffic can be slowed before they reach the affected area

I live in the Netherlands and they've been doing it here for a while, but they've called it differently (a 'spitsstrook', or a 'rush hour lane') which is opening up the hard shoulder for traffic during rush hour. There's no pretense of it being 'smart' anywhere, they had to in some places because the pressure of traffic is too high and widening the road is not an option.

> Without a hard shoulder, broken down cars have nowhere safe to stop and must stop in a lane of traffic.

Sounds absolutely insane and lethal when you write it down doesn't it?

Come to Massachusetts where "Breakdown Lane Travel" is allowed during rush-hour (only in the inbound direction in the morning and outbound in the afternoon/evening and where signposted).

We have reasonable sized pull out areas every mile or so, but the really nasty part is that the on-ramps now suddenly dump you into the moving breakdown lane traffic (which would otherwise be your acceleration / merge lane).

Thankfully, traffic is so bad around Boston, that typically it's all at a standstill anyway.

As a Uk person who has driven on them, I had to look it up as well at the time. I was shocked to find how they weren’t smart. I honestly thought the smart part was safety and breakdown detection as the hard shoulder was used for traffic. I was wrong and shocked.

I don't know if they did start showing worse results. I think the problem is that a death from a collision with a stopped car is very visible, while the drop in average accidents over a longer period isn't.

There are some schemes with automatic detection of stranded vehicles and lane closure. These work well. But some of the schemes rely on humans looking at cameras to spot people and shut lanes. This has left people sat in live lanes for 20+ minutes, eventually someone isn't paying attention and hits them.

> Think how much money can be saved by just... not allowing anyone to drive there

Completely agree that it's a dumb idea, but I think the point is that converting 3 lanes + a hard shoulder to just 4 lanes is far cheaper and logistically easier than adding a whole extra lane.

I'm not sure if Highways England have any right to the land either side of a motorway, but if not I imagine spilling over into other people's land would be a huge sink of time and money.

It's definitely not just you. Even worse than the hybrid hard shoulder is one stretch of the M5 where the road is 4 lanes at all times, and the occasional lay-by for if you break down.

Taking the outside lane out of use would reduce capacity proportionately.

It's a tradeoff though, because opening up the hard shoulder alleviates pressure on the other lanes, thus (theoretically) reducing traffic jams.

It's a patch though, it alleviates some of the symptoms but the underlying problem that is much more difficult to stop is commuting and travel in general. I'm kinda glad we've been forced into a work-from-home experiment the past year, it gives hard data on WFH's effect on traffic pressure. But it's not enough.

I think companies should be incentivized to open up smaller offices across the country, and people incentivized to live where they work / work where they live. Which will mean bringing the 'good' jobs out of the cities.

And they’ve been blocking roads with highly disruptive roadworks for literal years to “convert” them

Roads are hideously expensive - but smart motorways aren't just wifi enabled signs: there's traffic/weather detection and then the control systems behind adaptive traffic lighting/speed control - and those systems have to be reliable and maintained.

I haven't seen recent statistics/costs but remember that the cost of the converting the M42 in 2006 to a smart motorway was £9M per mile, creating a new lane would have been £16M per mile.

My guess is there are huge profits for the companies that build them.

Some of the motorways in the UK have been under construction for what feels like decades. Surely the same job could be done with machine vision these days!

None of the these devices are real-time or have any deterministic latency for communication. It's very far from a trivial issue.

The entire idea is very futuristic.

Reading that they need cameras to monitor the roads makes me think, why not have the government pay people to install data collecting sensors/cameras on their private cars.

Yeah, the answer to the "why not" is simple, because most governments can't be trusted nowadays, especially not with data. Which is a freaking sorry state of the world.

The other problem is as the road is extended either by these means or more building the traffic increases to fill that extra capacity.

Pretty sure the cameras have some leeway built in, and in addition your speedometer will probably read a few mph under at 70, unless you're using one on a satnav. I do agree though, I tend to just try and find someone else who has cruise control and keep a steady distance behind them.

I have come upon a stranded car a couple of times and it is apparent that most drivers do not notice until it is almost too late. Factor in the time it takes the brain to respond and you literally have an accident waiting to happen.

I am now at the age where I am never in a hurry so, both times, managed to avoid any contact but witnessed some hair-raising near misses.