Traffic light programming, and the tale of the Magic Laptop

The Magic Laptop

One evening many years ago some PTUA bods and I were meeting with a Vicroads bloke about traffic light priority and other related issues.

He had a laptop with him, and it displayed a diagram of a major intersection; I think it was somewhere out on Burwood Highway.

While pondering topics such as tram priority, he talked us through how the traffic light sequences worked, and how the traffic flows, showing us on the laptop.

And he showed us what would happen if the sequence was tweaked; part of the sequence runs for longer, causing some vehicles to pass through more quickly, some to be delayed a few seconds. Really interesting.

Someone asked: “So that’s a simulation?”

The response: “No, that’s real. It’s happening right now.”

So he’d been fiddling with the traffic lights in realtime, and local motorists were probably wondering why they were zipping through or being slightly delayed.

That wasn’t just a laptop, that was a Magic Laptop.

Spencer Street and Collins Street intersection

Programming traffic lights

Anyway, via this and other discussions with people who seem to know what they’re talking about, I get the sense that Melbourne’s traffic lights are reasonably flexible in terms of their configuration, and can be controlled remotely.

But there’s a limit. They can’t handle all scenarios automatically, so for instance when trials of absolute tram priority were done in Nicholson Street, it needed someone to manually control the lights to give a green for the tram.

There are also apparently limited resources, so opportunities to re-program traffic lights don’t come up as often as they’d like.

Why is it so?

Everywhere in government (as well as in the corporate world), if you go digging, you’ll find there’s usually a reason for something.

Sometimes it’s a reason which doesn’t quite make sense, or is outdated in the face of changing circumstances, but a reason nonetheless.

A couple of weeks ago I noticed that the traffic lights at Spencer and Little Collins Streets had an extraordinarily short green man, only about 18 seconds. Then the red man would flash for about another 10 seconds, and then there’d be a solid red man for a full 40 seconds before the parallel traffic light turned yellow.

This is utterly ridiculous in the central city, next to a major railway station, where pedestrians should be the priority.

Setting it like this is just goading people to cross against the lights.

I made enquiries with City of Melbourne, and discovered it is a road managed by Vicroads. So I approached them about it, and eventually I got a response.

Why was it like this? Because Spencer Street is closed for sewer works south of Collins, and they wanted to allow vehicles to detour into Little Collins easily.

But — as shown by the video — there wasn’t much traffic coming down Spencer that actually needs to detour.

Once they realised this, they set it back. Just like that. Someone probably clicked some buttons on a Magic Laptop, and it was done.

A good outcome, with some delicious technical tidbits in the email trail which I won’t publish, other than to say yes, they really do use the reference numbers on traffic control boxes.

(The few cars, and the number of people crossing Spencer Street against the lights would appear to indicate more needs to be done at this intersection to accommodate pedestrians. Note also that this is just metres from where the old pedestrian subway under the road from the station used to emerge.)

The bigger picture

I’ve also had a discussion about that super-annoying crossing at Centre Road/Eskay Road in South Oakleigh. I haven’t seen it yet, but apparently it’s been tweaked too.

But the bigger picture issue is that traffic lights (even in the CBD) are being programmed with poor outcomes for pedestrians. Sometimes as above there’s a reason — sometimes, apparently, it’s just an error.

Much the same issue occurred at Elizabeth/Little Collins a couple of years ago.

And more recently, City of Melbourne has put in brand new installations that failed to auto activate the green man, despite it being policy within the Hoddle Grid.

(And after they fixed that one, the timing was wrong, with — again — too little green man time.)

These things do make a difference. It’s not just about compliance and safety. The travel mode you want to thrive is the one you should encourage. Make it easier for people to walk, and more people will walk.

What I have learned is that Vicroads is now consulting on some of these issues with groups such as Victoria Walks. This is definitely progress.

Be polite, but firm

Individuals shouldn’t really have to get these things fixed. But in the real world, everybody (including Vicroads and City of Melbourne) is stretched for time, and clearly some things simply aren’t being spotted and fixed otherwise.

So…

Put in a report. Twitter may not be sufficient, so do it via their feedback web site. Include a photo if it’s at all useful.

Be polite. Scrupulously polite. You won’t get anywhere by shouting.

Explain your case. Present the evidence, the logic.

Keep a copy of your query text, and the reference number, because some web sites (such as Vicroads) don’t email you a copy back, and it may be useful at the next step.

If you get a pro forma reply which doesn’t make sense or doesn’t address the issue, query it. Be polite, but firm.

And with a bit of luck, and if your point is convincing, you might just get it fixed.

Of course, what I really want is a Magic Laptop.

Some thoughts on CBD street design and safety

Thursday’s awful incident brings back memories of January’s attack, and authorities are rightly saying that it will be investigated from all angles.

One issue is road infrastructure that allows a vehicle to reach high speed in a constrained, pedestrian-dense area like Flinders Street.

The speed limit is 40. But apart from the placement of tram superstops (one of which finally stopped the vehicle) there’s been little change to this section of road in decades. It’s wide and straight, with no form of traffic calming; no speed humps or other treatments that you might see in a suburban school zone, for instance.

Not that the road was clear. The traffic was apparently congested, but the driver apparently veered out onto the tram tracks to speed towards the intersection.

The Prime Minister commented:

“[Melbourne] has big wide streets, wide footpaths and of course, it has trams, and the tramways enable a [car] driver, as this driver did to pull out of stopped traffic, get into the tramway and then make an attack. This is an issue for protecting, for example, the Bourke Street Mall.”

In Flinders Street, it shouldn’t be too difficult to make it harder for motor vehicles to get onto the tracks. A small separation, preferably with the tracks raised slightly, would do it — such as this design I spotted in Brussels in July:

Separated tram track on Avenue Fonsny, Brussels

The beauty of this design is that it helps stop all sorts of unauthorised vehicles getting onto the tracks, or turning across them. This would improve tram safety and cut delays to trams.

It does come at the cost of convenience for some motorists (frankly, not a priority in the city centre) and making it trickier for emergency services to use the tram tracks, as well as limiting where people with wheelchairs or other mobility aids can cross. These issues would need consideration.

Apparently none of this is impossible to resolve, because you can already find this design in central Melbourne — in Spencer Street:

Separated tram track in Spencer Street, Melbourne

Depending on the precise design, it doesn’t make it completely impossible for a vehicle to get onto the tracks, but it could make it a lot more difficult.

The PM is right to note the Bourke Street Mall presents some challenges. But it would be wrong to assume it can’t be solved.

Permanent bollards have recently been placed either side of the tramway, at both ends.

Bollards in Bourke Street, Melbourne

But the tramway itself remains open — to trams, obviously, but also to other vehicles, including unauthorised and/or clueless motorists.

What do other cities do? Some of them use retractable bollards, to let through only authorised vehicles.

Here’s a quick video of them operating in Cardiff. This is the main bus route between the city centre and Cardiff Bay.

This has potential, though in this case, you can see it’s a little problematic:

It’s far too slow to respond. I assume it’s controlled remotely from an operator somewhere watching on CCTV. To cut delays, the bollards should retract as the bus approaches, not wait for it to stop and wait.

If bollards like this were placed at either end of Bourke Street Mall, hopefully engineers could come up with a system that sees them retract automatically as trams approached or waited in the tram stop.

Another issue might be the risks of putting them in areas of heavy pedestrian traffic. Careful placement (between tram stop platforms, or away from the footpath, on the edge of the intersecting roadway) might resolve this.

Trams on Bourke Street

Getting this right brings numerous benefits, not just for safety from vehicle attacks, but also keeping pedestrian areas free of unauthorised vehicles, and also preventing them disrupting trams.

Authorities shouldn’t give up. Other cities have solved similar problems. I’m sure we can too.

Brussels has zebra crossings. Lots and lots of zebra crossings. Could we have more too?

One of the things I found fascinating about Brussels on our recent holiday was – in contrast to Cardiff – how they’ve gone out of their way to make life easy for pedestrians.

Most striking was that there were zebra crossings. Lots and lots of zebra crossings.

Zebra crossings in Brussels

When I first spotted how many there were, I wasn’t totally sure what I was seeing, and actually warned my fellow travellers to watch and observe the locals, just in case the road markings didn’t mean what they mean in Australia. Did vehicles really have to stop for pedestrians at all these locations?

Yes. The stripes mean the same thing (except if there are traffic lights). There are just lots of zebra crossings.

Walking to the Metro in the rain, Brussels

Zebra crossings on main streets, zebra crossings on minor streets, zebra crossings on divided roads with trams in the middle, zebra crossings at intersections and mid-block.

Lots and lots of zebra crossings, and drivers observed them – perhaps because they’re so used to them.

Roundabouts? Not a problem. In Australia, these are virtually the only locations where vehicles in any direction don’t have to give way. Exceptions are rare. The Belgian roundabouts I saw had zebra crossings on all sides:

Brussels street

Two T-junctions so close together that putting zebra crossings on every side would mean three in row? Sure, go ahead. The motorists will survive:

Zebra crossings in Brussels

Generally, motor vehicles had to give way to pedestrians, but pedestrians had to give way to trams.

Zebra and tram crossing in Brussels

How much does having lots of zebra crossings affect traffic? It’s hard to say, but the cars driving around didn’t seem to be unduly held up. When I saw peak hour traffic set in, it was clear that – as anywhere else – the main thing delaying cars was other cars.

Some wider streets had traffic lights with pedestrian crossings. At many of these, you didn’t have to press a “beg” button – there was no button. The green man triggered automatically:

Pedestrian crossing in Brussels

This not only tells pedestrians approaching that they don’t have to press a button to cross. It also indicates the authorities have no intention of changing it (and necessitating having a button) any time soon.

This of course is how it should be. If you’re giving the green to vehicles, why wouldn’t you also give the green to pedestrians? (More about this in another rant post soon.)

Note the signalised crossings have the same on-road markings as zebra crossings. I wonder if that helps with compliance? They’re much more obvious than the Australian dashed line markings.

Pedestrian crossing in Brussels

At a few spots I saw, buttons were necessary to trigger the green man. These seemed to be reasonably responsive, not making you wait too long:

Pedestrian crossing button in Brussels

In some locations, presumably those that get very busy at times, the crossings were very wide.

Pedestrian crossing in Brussels

Along with mostly wide footpaths (at least, wide enough to cope with pedestrian traffic), the design of the crossings left one with the impression that Belgian authorities would prefer you walked than drive.

It’s the sort of thing that some might not even notice, but it left an impression on me. If only Australian authorities were so inspired.

Could we do this in Australia?

Sure. But while some new zebra crossings have popped up over the last few years, they don’t seem to be routinely installed.

This spot outside Gardiner Station clearly should have been a zebra crossing:

This was almost a zebra crossing, but someone messed up. (I shouldn’t have opened my big mouth. It’s now entirely a signalised crossing… which thanks to the beg button, many people ignore):

This is the newish tram stop on Collins Street at William Street. It could have had zebra crossings at the non-intersection end. But someone decided a signalised crossing was a better idea. It’s maddeningly slow to wait for if you’re crossing, and many people just cross whenever there’s a gap in the traffic:

Collins St near William St

I would think there’s also scope to place zebra crossings on side streets at intersections, particularly in suburban shopping centres.

Main road/side street intersection, Bentleigh

The law says a vehicle turning into the street gives way, but convention is often the opposite, with vehicles exiting the street often giving way instead.

And pedestrians sometimes wave stopped motorists on, when the motorist is doing the right thing and giving way. (Do me a favour: if you’re crossing and other people are too, don’t wave the car on. You might not be in any great hurry to get where you’re going, but you don’t speak for everybody else.)

Painting zebra crossings right across the side street would not just encourage walking, it would also help reduce the confusion over who’s meant to give way to whom, in what are typically high traffic (pedestrian and vehicle) areas.

Ditto car park entrances, where motorists entering and exiting are meant to give way to pedestrians.

More zebra crossings are perfectly possible. Here’s what they’ve done in Footscray. It was quiet when I took this photo, but often there are lots of pedestrians around. Somehow, the traffic still gets through:

Zebra crossings in Footscray

Potentially two-lane main roads like Centre Road and McKinnon Road could have zebra crossings too. That would be bringing it up to Belgian standards, and would be in line with the Vicroads Smartroads strategy which says it’s meant to prioritise pedestrians and buses. What would be the effect on traffic? It would be interesting to see it modelled.

Ultimately, if we prefer people walk where possible, more needs to be done to encourage it.

Congestion is not the enemy

This in the Herald Sun a few weeks ago: Melbourne traffic congestion on par with world’s biggest cities like London, Rome and New York (paywall):

TRAFFIC congestion in Melbourne is on par with New York and could rival the worldโ€™s worst cities if nothing is done to combat the problem.

Figures supplied by Tom Tom show congestion levels in Melbourne are at 33 per cent compared to its population.

This means motorists are sitting in peak hour congestion a third longer than if the traffic was free-flowing.

And then this in The Age: Melbourne now as clogged as Sydney, and the city’s north-east has worst traffic:

The Grattan Institute research is based on an analysis of Google map data for more than 300 routes in and out of Sydney and Melbourne. It was collected 25 times a day over 12 weeks between March and June 2017 and found:

An average morning commute to the Melbourne CBD by car takes almost 70 per cent longer than in the middle of the night.

These are based on very similar surveys: the Herald Sun used the much-criticised Tom Tom congestion survey. The Age used a Grattan Institute survey.

Both use the same flawed methodology. They compare a city’s traffic speed at quiet times with the traffic speed at peak hour.

Apart from assuming that getting around by car as fast as possible is automatically the most important thing, you get wacky conclusions because a city with 24/7 congestion (only slightly worse at peak hour) is deemed to be less congested than a city where most of the time there’s free-flowing traffic but for a couple of hours a day it’s proportionately worse.

(The Grattan Institute says there’s more coming. I hope it’s more well thought out than just looking at motor vehicle commuting.)

In any case, if this is our big conclusion that drives transport policy, I think we’re asking the wrong questions.

CBD traffic, Lonsdale and William Streets

The debate shouldn’t be about congestion

I don’t think the debate should be about congestion. It shouldn’t even be about mobility. It should be about access to opportunity: jobs, education, amenity.

It’s not about whether people who choose to drive* are delayed by others who choose to drive. It’s about whether everybody (including those who don’t drive) can get to the places they need to get to.

*Or are forced to do so for lack of viable alternatives.

(To contradict myself for a moment: congestion that gets in the way of efficient transport modes absolutely is the enemy. Ways need to be found to get pedestrians, bikes, trams, buses and trains around it, or at least through it quickly.)

One of the major benefits of a big city, if there are lots of opportunities well-serviced by public transport (and walking and cycling), is that it makes it easier for everyone, of every age, and every income level, to access them… provided they don’t insist on bringing their 2 tonne private vehicle, of course.

Bourke Street Mall, lunchtime

What sort of city do we want?

There’s also a lesson in the headlines. In the Herald Sun story, New York, London and Rome are cited, and compared to Melbourne. In other words, the most prosperous, vibrant, successful cities on Earth have congestion. And we’re becoming more like them.

Is that actually a bad thing?

As Samuel Schwartz says in his excellent book which I just finished reading:

…a study from Texas A&M’s Transportation Institute found a powerful correlation between per capita traffic delay and per capita GDP; and the correlation wasn’t negative, but the opposite. For every 10 percent increase in traffic delay, the study found a 3.4 percent increase in per capita GDP. It’s not that congestion itself increases economic productivity, but that places with a lot of congestion are economically vibrant; those without, not so much.

Should we really be trying to stamp out congestion, or should we look at how other cities deal with it?

The big world cities don’t deal with congestion by eliminating it – which basically isn’t possible; building more roads just grows more traffic.

Rather, they provide lots of ways of avoiding traffic congestion, by making sure more people can get around without driving in it and adding to it: by providing viable non-car modes for most trips, including non-work, non-CBD trips.

Decentralisation and liveability

Meanwhile, the state Coalition is calling for more decentralisation to maintain Melbourne’s liveability. By “liveability” I suspect they actually mean crowding and congestion.

Again, look around the world at the cities we might aspire to be.

What makes Melbourne’s congested city centre successful in this age of the Information Economy is lots of people in a relatively small space. The majority, who come in by train, are simply never affected by traffic congestion. (They are affected by rail disruptions such as last week’s major outage, but that’s not an everyday thing.)

Decentralisation plays against one of our key strengths.

Which is not to say we shouldn’t increase the number of viable business districts, if it’s possible.

But just moving lots of people to car-dominated regional towns doesn’t really help. As Alan Davies notes, decentralisation is just another name for regional sprawl. And replacing urban sprawl with regional sprawl isn’t actually a positive.

Okay I’ve rambled a bit again.

But my key point is: congestion isn’t our enemy. Lack, and inequity of access is what we should be talking about, and seeking to fix.

Surprise surprise, expanding the Dingley Bypass caused more traffic

Remember the Dingley Bypass?

The western end was built as the “South Road Extension” as two lanes (one each way) just last decade. It mostly wasn’t dual carriageway, but it was otherwise suspiciously freeway-like. Having appeared in the 1969 freeway plan, it wasn’t difficult to see that it would be expanded.

Dingley Bypass in 2015 (Google Streetview)
(Dingley Arterial 2015 – Google StreetView)

Sure enough in 2010, it was announced it would be expanded to four lanes (two each way).

Fast forward to March 2016 when it opened… and somehow it’s become six lanes (three each way). Classic salami tactics!

Sow roads, grow traffic

So wouldn’t making a road three times as big be likely to lead to problems on the roads it connects to? Why yes.

What a surprise — it seems South Road now has congestion problems:

Vicroads: South Road traffic study

Vicroads: South Road traffic study

Having fed lots more traffic into South Road, what can they do? Currently through Moorabbin it’s six lanes (three each way, but with one taken by parking). They could try and fix this bottleneck by imposing clear ways — effectively expanding traffic capacity by 50% and leaving a long mostly residential stretch of the road as a huge six lane traffic sewer. Eugh.

Making the environment even more hostile to pedestrians, bus/train and bike users sounds like a great way to produce even more traffic. (The intersection of South Road and Warrigal Road is horrible.) And it would probably just move the bottleneck elsewhere. Nepean Highway next for widening, perhaps?

The proposed freeway to remove the bottleneck created by the previous proposed freeway...

Even with no more road expansion, more traffic lights might be needed to help pedestrians and cross traffic now that the road is more congested — especially at the eastern end around the Sandbelt Hotel. (Somewhere I saw an online petition for a pedestrian overpass; I can’t find it just at the moment.)

Meanwhile, the main bus route along there is only every 20 minutes on weekdays (including peak hours, when it no doubt gets stuck in traffic), and every hour at night and on weekends. Most other suburban routes run even less frequently — generally only every 30 minutes on weekdays. No wonder so many people are driving.

I don’t know quite what the solution is. But I do know that free-flowing urban traffic is a myth. At least if there’s a bottleneck now, it’s preventing traffic levels growing even more.

Ways have to be found to move more people around the suburbs them each bringing a two tonne chunk of metal. Expanding road capacity further is the last thing we should be doing.