Trams that never stop at traffic lights

Trams that never stop at traffic lights could be part of Melbourne’s people-moving future

By Loretta Florance

ABC News

Trams that never have to stop for traffic lights could be the norm in Melbourne in the future, under plans being developed to deal with population growth in the city.

For the past seven years, VicRoads has been using mathematical modelling to figure out how changing conditions – such as traffic light frequency and clearway times – affected traffic flows.

VicRoads director of network policy and standards Andrew Wall said public transport was the focus of future planning because of its ability to carry far more people.

“We’ve used the model to assess what happens if we for example give absolute priority to the tram, which means that when a tram gets to a set of traffic signals, it never has to stop,” Mr Wall said.

“The model lets us look at those sort of situations, so it’s giving us some insights into how hard we can push the traffic signal system to give priority to trams.”

With Melbourne’s population projected to hit just under 8 million by 2053, Mr Wall said the roads authority had to look at “clever ways to move people around the network”.

“The focus on moving people is probably the critical thing for us,” he said.

“We want to set the road network up and the traffic signals up so we’re maximising how many people we move, not necessarily how many vehicles.

“Over time we have given more and more priority to trams, and we’re likely to give even more in the future.”

The mathematics of being stuck in traffic

Professor Jan De Gier and Dr Tim Garoni began working with VicRoads in 2008, as part of an internship with the Australian Mathematical Sciences Institute and the ARC Centre of Excellence for Mathematics and Statistics of Complex Systems.

Since then the University of Melbourne researcher has been applying the basic principles of mathematical physics and statistical mechanics to roads in a unique way.

“We studied simple particle models in mathematical physics and noticed that some of these ideas were actually applicable to traffic flow, so we could use very efficient models and intuitive ideas that we learned over the years, in the modelling of traffic,” Professor De Gier said.

“You can view cars and buses and trams on a traffic network as simple particles, so you ignore a lot of the details that are inessential for traffic.

“If you model particles flowing through on a graph … and you set up the rules properly, you’ll see that they behave very much like traffic, so they spontaneously form traffic jams, and other things.”

“So they’ve provided the tool, it’s up to us to use the tool to help inform decisions on how we operate the road network,” he said.

He said mathematics was being used in similar ways in European countries like Germany and Switzerland because it was a cheap and effective alternative to tradition modelling.

Mr Wall said using mathematical modelling was actually more informative than conducting real world traffic tests.

“We did a [real world] trial around Kew, a network of around 50 intersections, over four weeks, where we ran the traffic signal system differently on a day-on, day-off basis,” he said.

“What we found with that is that the variability in traffic from day to day made it really difficult to assess whether we were having an impact.

“So that sort of pointed us to that there’s got to be another way to analyse and assess the changes, which led us to a discussion with some mathematicians on a different approach to how we might do this.”

To develop a sophisticated traffic model of each individual road, intersection and vehicle in detail would cost between $500,000 and $1 million, Mr Wall said.

“This model can be set up for that same network for a fraction of that cost and lots of different options tested.”

Professor De Gier said the method meant that they could see how changing parking policy, tram frequency and traffic light intervals affected the network several kilometres away.

“VicRoads didn’t have tools to model traffic on a network, so all they could do was model in a few intersections or very global origin-destination modelling,” he said.

“What we can do with them is simulate suburbs or collections of traffic intersections of about 100, and see how traffic evolves on such a network if we change certain things.”

Street parking benefits trams

One of the surprises that has come from his research came from when his team modelled the effect of clearways on tram times, Professor De Gier said.

“The traditional thought was that if you have clearways all day then that would be good for cars and trams,” he said.

“It turns out that [with clearways] cars overtake trams and can obstruct the tram at the next intersection.

“So having parking provides a gating effect for trams who can then move through the network better.”

For the first time this year, Professor De Gier’s team will not be looking at generic traffic models, but will be looking at a specific area – Melbourne’s inner north.

Much of this research will be done at Monash University by research fellow Joyce Zhang of the new ARC Centre of Excellence for Mathematical and Statistical Frontiers.

“They’re extending the use of this model to look at the road network in the inner north,” Mr Wall said.

“So while we’re using this as a generic tool to look at networks, it’s actually being applied to look at the whole road network in the inner north and how that might work in the next 10 or 20 years.

“There’s a lot of north-south tram routes in that area, there’s a lot of population growth happening on those roads and to able to cater for that population growth, we need the north-south tram routes to work really well in terms of moving people.”