Is ramp metering dead?

[Peter Freeman]

I think major junctions, and all junctions on the M25, should be fitted with MS4s on slip roads giving speed limit and lane info at all times as well as journey times, e.g. NSL - M40 25 mins, M4 35 mins

Speed limit: yes. It's useful to know the speed limit on the mainline before you get to it. Needn't be an MS4: just a normal VSL sign.
Journey times: no. These should be on the mainline, if at all, and not between every junction. On the mainline they are useful to a much larger proportion of users.

[traffic-light-man]

If anyone's interested, here's a link to the report paper on the M6 trials, which started in 1986.

Here's a link to the Dutch version, too, which is an interesting read. Its worth noting that they do signal multi-lane approaches independently per lane, by law.

The plus feature of these new generation sites being able to operate without MIDAS data feeds was also a feature of the M6 trial in 1986, though back then, the system did require a computer in its own building at Perry Bar.

I feel like the MIDAS data requirement is (or now, was) a bit of an oversight. Surely by the mid-2000s, the technology was there to develop the systems independent of MIDAS data (using dedicated loops and processing) and then be able to use MIDAS if and where it was available?

[Peter Freeman]

Here's a link to the Dutch version, too, which is an interesting read. Its worth noting that they do signal multi-lane approaches independently per lane, by law.

In my early observations of AU implementations I too thought "why aren't the ramp lanes signalled independently in order to avoid releasing a platoon?". This became more relevant as the maximum number of signalized lanes on an on-ramp increased from 2 to 3 and then to 4. The answer provided by the authorities in their technical documentation is that it "can confuse drivers".

Anyway, in practise when multiple vehicles are released simultaneously, their varying speed capabilities, and varying driving styles and degrees of 'aggression', cause them to stretch out as the ramp shrinks to a single lane. It's a form of the boy-racer drag-race away from the lights. Usually, by the time they merge they have effectively become single vehicles.

I have observed, rarely and only temporarily during roadworks that reduce the number of lanes, a relaxation of the rules to say "two vehicles per green each lane". The technical documentation states that, when used, this results in a practical throughput increase of only 1.7 rather than 2.0 - again due to driver confusion.

[ManomayLR]

I wonder whether AMI signals on gantries can toggle a red X with a speed limit - this doubling as ramp signals?

(With fixed sign 1 vehicle per lane each time the Red X disappears)

[Big L]

I see plenty of drivers ignoring a red X when it’s lit continuously.

[Brenley Corner]

Its seems to be very much alive, I would say

Seen installed on the M3 J4 & J3 Londonbound on-slips even though a recent smart motorway conversion. Also seen working as designed last Sunday on the M25 J11 clockwise onslip.

Tony

[ElsieSeedlow]

<snip>
I feel like the MIDAS data requirement is (or now, was) a bit of an oversight. Surely by the mid-2000s, the technology was there to develop the systems independent of MIDAS data (using dedicated loops and processing) and then be able to use MIDAS if and where it was available?

Some of the technology was 'there', but then limitations or absence of spare pairs in the longitudinal cabling for a dedicated system meant it couldn't happen. The reuse of MIDAS loop data was actually quite clever and low cost.

At the end of the day Ramp Metering is difficult to explain correctly and easily. So even though a good installation "pays" for itself quickly investment in wires and lights is, as ever, secondary to asphalt and concrete in the highway world.

[Hugo Nebula]

M60 junction 2 has them. It's been a while since I when through there during the rush hour but they are still in place. And very much still in use at jct 2 on the M56 yesterday. I'd be happy for them to be installed at the eastbound A34 on slip onto the M60 at Jct 3 where nose to tail traffic expects the vehicles in lane one to give way to them.....

If it's nose to tail traffic then what's the problem with allowing joining vehicles to merge in turn. I agree it's the last of a lousy series of junctions, but in reality at that point the left hand lane is marked as for use only being by vehicles wanting to exit at Cheadle junction 2- anyone on a through route should be using the 3 other lanes.

In my experience there is a similar number of vehicles joining at 3 to those exiting at 2, all within a less than 400m stretch - they just need to negotiate each other sensibly. (My Bold)

I do indeed exit at Jct 2 and if only the vehicles joining at 3 did act sensibly. And you could argue there's no negotiating to be had, the vehicles joining are to give way, something that often doesn't happen - exacerbated by the sight lines and that very very short give way line. it is a pigs ear of a layout.

Not helped by people not using the slip roads to accelerate to the main line speed, and also jumping across at the start of the dotted line rather than merging at the end.

Sent from my Brain using my Fingers

[Barkstar]

If it's nose to tail traffic then what's the problem with allowing joining vehicles to merge in turn. I agree it's the last of a lousy series of junctions, but in reality at that point the left hand lane is marked as for use only being by vehicles wanting to exit at Cheadle junction 2- anyone on a through route should be using the 3 other lanes.

In my experience there is a similar number of vehicles joining at 3 to those exiting at 2, all within a less than 400m stretch - they just need to negotiate each other sensibly. (My Bold)

I do indeed exit at Jct 2 and if only the vehicles joining at 3 did act sensibly. And you could argue there's no negotiating to be had, the vehicles joining are to give way, something that often doesn't happen - exacerbated by the sight lines and that very very short give way line. it is a pigs ear of a layout.

Not helped by people not using the slip roads to accelerate to the main line speed, and also jumping across at the start of the dotted line rather than merging at the end.

Indeed an unwillingness to get up to motorway speeds is often an issue at most junctions. Part of the problem at this junction is you get little opportunity to assess what's happening on the main line and accelerating up to motorway speeds can cause as much of an issue as not doing depending on traffic levels. The mess is given a further stir because the majority of traffic joining then has to cross into lane 2 to continue along the M60 as lane 1 becomes the off slip for jct 2 in a couple of hundred metres. Hence wondering if ramp metering might help.

https://www.google.co.uk/maps/place/Che ... 4115?hl=en

[Peter Freeman]

If anyone's interested, here's a link to the report paper on the M6 trials, which started in 1986.

Here's a link to the Dutch version, too, which is an interesting read.

They are interesting documents. There's a similar guide (as one chapter in) an Australian document. Find it by googling 'Austroads AGSM-16 Guide to Smart Motorways'.

There's a previous Sabre thread on ramp metering here. Perhaps these threads should be merged?

The most extensive coordinated ramp metering system is in Melbourne Australia, with approximately 110 controlled on-ramps plus 10 controlled connectors (motorway-to-motorway). The second largest is in Auckland NZ, with 91 controlled on-ramps/connectors. There are schemes in Sydney and Brisbane too. All are considered to be successful, with a >1 positive BCR.

On normal weekdays in Melbourne a majority of the signals will be active at either morning or evening traffic peak. Not now though: Covid restrictions .

Ramp metering in AU is considered to be a component of the Smart Motorway suite, which is otherwise similar to the UK's (ALR, LUMS, VSL, VMS, ERA). The UK total length of Smart Motorway is much greater than AU's: ours are urban.

[Peter Freeman]

Ramp- and connector-metering continues to roll out in Melbourne and other Australian cities.

The configuration shown here is interesting. Two independent 2-lane on-ramps converge to run side-by-side. These each flare to three lanes for signals, and then narrow to single lane, both still independently side-by-side. They then lose the barrier that separates them, and join a 3-lane carriageway at a 3+2=5 lane-add. After a few hundred metres, the left lane tapers, so the inbound M1 towards Melbourne shrinks from D5M to D4M.

The extra on-ramp arose recently from a new bridge over the railway, visible on the left of the picture. Location -
https://www.google.com.au/maps/@-37.880 ... 436,16.27z
(G Maps not yet caught up, but latest G Earth shows it)

This dual ramp metering scenario has also appeared at other points recently, including -
1. Eastbound merge of M80 into M1, a few km west of Melbourne.
2. Northbound merge of Footscray Road plus new M4 into M2, Melbourne.
3. Eastbound and westbound on-ramps from M7 onto M4 in Sydney (can be seen on recent G Earth).

[Peter Freeman]

Just a few musings arising from posts up-thread:

I believe a lot of the thinking was that the Smart Motorway speed control negated the need for ramp metering which is why some sites were removed.

Such thinking, though common, is misguided. VSL will defer flow breakdown, but, like everything, it has its limits, so ramp metering is still relevant.

There is no incompatibility between smart features and ramp metering. Indeed, in Australia ramp metering (the fully-developed single-vehicle-release drip-feed model, with inter-ramp coordination and upstream recruitment) is considered (as is a travel-time prediction display) to be simply another one of the various smart features.

I think the M1 around Sheffield is far better since the smart motorway was finished, yes it has safety issues but either way it’s got 25% extra capacity which means ramp metering isn’t really needed as you can keep the M1 and slip roads flowing rather than holding up the on slip.

Maybe, but what about when that added capacity is gobbled up by traffic growth? Or what about special (heavy traffic) events? Or abnormal downstream hold-ups that propagate upstream? If/when the metering isn't required, it won't turn on, so nothing is lost by leaving it in place.

I'm guessing that on the M42 east of Birmingham, there's effectively a lane gain on the junctions so ramp metering isn't needed.

J33 has lane gains so there doesn't seem any point in metering.

That's also also a common line of thought. I had it too when I first encountered a metered ramp that led to a lane gain. In isolation, there would be no point. However, in AU we do have metering at lane-gain ramps. They make sense in a coordinated system, because they are recruited as 'helpers' by struggling ramps further downstream. 'Struggling' means that a ramp is throttling its flow as much as allowed (and/or the 'ramp full' sensor at the top of the ramp has been triggered), yet the mainline congestion persists.

It's similar to another scenario that occurs: a skeptical driver, held on a ramp, is frustrated as he sees the associated mainline flowing freely. What he doesn't see is that the actual congestion is further downstream at the next merge, and has not been fully resolved by the local ramp meter working alone. That ramp has therefore recruited its upstream neighbour for assistance.

[Chris5156]

I think the M1 around Sheffield is far better since the smart motorway was finished, yes it has safety issues but either way it’s got 25% extra capacity which means ramp metering isn’t really needed as you can keep the M1 and slip roads flowing rather than holding up the on slip.

Maybe, but what about when that added capacity is gobbled up by traffic growth? Or what about special (heavy traffic) events? Or abnormal downstream hold-ups that propagate upstream? If/when the metering isn't required, it won't turn on, so nothing is lost by leaving it in place.

By the time that day comes, the ramp metering equipment installed years ago will be well past its service life and new equipment would have to be installed anyway. And in the meantime, why pay to maintain and supply power to equipment that isn't required over a period of 10-15 years?

[Peter Freeman]

By the time that day comes, the ramp metering equipment installed years ago will be well past its service life and new equipment would have to be installed anyway. And in the meantime, why pay to maintain and supply power to equipment that isn't required over a period of 10-15 years?

That's a fair point.

[Chris5156]

By the time that day comes, the ramp metering equipment installed years ago will be well past its service life and new equipment would have to be installed anyway. And in the meantime, why pay to maintain and supply power to equipment that isn't required over a period of 10-15 years?

That's a fair point.

No! Come back!

[Peter Freeman]

Here are two video's from Perth, Australia, that I hadn't noticed until now. This first one, four or five years ago, was an update during the smart motorway upgrade of Kwinana Freeway -
https://www.youtube.com/watch?v=MgeBMKQoJqs

And this one is a TV news item from the morning when ramp metering was first switched on, about three years ago. It shows a surprisingly good improvement - presumably the combined effect of metering plus the other smarts plus an extra lane -
https://www.youtube.com/watch?v=woqcvHXO6Jo

That first few kilometres of treatment has since been extended, and is being extended further N and S on Mitchell/Kwinana Freeway, Perth's major cross-city route. The implementation is modelled on Melbourne's, which is long-established and extensive.

BTW, note the interesting shape of the emergency stopping bay (not connected with ramp metering, but a required component of ALR, as in UK). Perth seems to be building lots of these now, even where there is a shoulder. Presumably it's future-proofing in anticipation of 'going smart', which usually includes ALR.