Majid Sarvi: AIMES for the future

Majid Sarvi is the chair in Transport Engineering, the professor in Transport for Smart Cities and the program director of Transport Technologies at the University of Melbourne. He is the founder and the director of the world-leading transport testbed, the Australian Integrated Multimodal EcoSystem (AIMES).

How did you arrive in the area of smart mobility, Majid? What path did you take in study and was it an interest pre-study?

My undergraduate degree is in civil engineering. Transportation traditionally started in civil engineering, simply because the majority of it involved a relationship with infrastructure.

Next, I was at Tokyo University for almost three years, doing my post-grad and PhD in intelligent transport systems. Then I worked in industry for a while, including NEC and some start-ups in intelligent transport systems. The more I progressed, the more the topics in and around smart cities were at the top of my agenda.

When was it that you were doing your post-grad and work with NEC?

I finished my PhD around 2001, and then I worked for NEC and a few other companies in Japan, through to probably 2005. Since then I’ve held various positions at various universities in transportation systems.

Where did you obtain your undergraduate degree?

I studied at Tehran University, which is the oldest university in Iran, I think it’s about 120 years old. It was a civil engineering degree, with some basic subjects in transportation. It was a style of program that I think still runs in most universities around the world.

When you graduate from undergraduate study in civil engineering, probably, you know about a couple of subjects in transportation, but in my PhD I was totally focussed on intelligent transport systems.

So you didn’t, during your undergraduate degree, decide that ITS was the way to go, that was later?

Yes, I was very interested in transportation systems but my idea of that area was not very clear. But then I moved to Japan and started my postgrad at Tokyo University, where and when there was a lot of discussion about intelligent transport systems, it was a very hot topic then, as it is today. Japan was the perfect place, because you could see a lot of intelligent transportation systems in operation, even back then.

If it wasn’t Intelligent Transport Systems for you what might it have been?

I was also very interested in earthquake engineering. But probably …. maybe I would have become a structural engineer, or something along those lines.

I ask this question of people, because some people’s university bios contain a list of areas of interest that incredibly long and varied. What would you list as your various general and specific areas of interest in ITS?

From my PhD all the way through my career in industry, and then back to university, I’ve worked in only two areas. The first one is connected multi-modal transport systems, and the second one is pedestrian crowd dynamics. That’s the two areas I work on and that’s it.

Okay, hypothetical time. If you had an unlimited budget a not unreasonable time to carry out the work, what do you think would have the most impact on a problem we have in transport?

This won’t surprise people, as I always advocate this. To me, the key technology, and the key sort of measure that can pretty much tackle the majority of the issues we have in transport systems, as well as basically the availability of that technology today, is connectivity. I would make the transport system, I’m talking about the whole transport system, connected and integrated, and act as one system in a multi-modal way.

That would be, to me, the most important aspect of transportation systems that is missing, and have the most desired results, with appreciable impact.

To run such a connected, integrated smart city transport system, is there available today a management interface that could handle all of that work?

I don’t think it exists, but I also don’t think it would be difficult to build it. The problem is that to build such a system you need to have the right data. To get that kind of data you need to start connecting and integrating transport systems.

Then once data starts coming, I don’t think that building the methodology to allow you to run multimodal management is that difficult. There is some good research happening in my group, and indeed other groups around the world, that if they have the real raw data at hand they can make just such a management tool happen.

Okay, similar hypothetical, but this time except you’re under pressure. You’re on a tight budget and you need to act reasonably quickly, what transport problem would you attack?

I think if you look at Australian cities, they’re struggling with congestion and safety. Let us talk about safety, I think, including everybody in transport systems, so I will probably start with this too.

Again I start with connectivity! Unlimited budget, limited budget, it’s all the same to me. Connectivity is the only way you can fix these difficult issues.

Would you phase this work, rather than fix it all in one fell swoop, with a possibly large cost and large inconvenience?

Yes, I think you could do it in a staged way, but the beauty of this is that it doesn’t really require a lot of funding. If you look at what we’re doing for example now in Melbourne or in Sydney, we’re spending over $50-60 billion dollars.

To make a connected transport system, the cost would probably be less than $500-600 million dollars. If you look at the scale of that amount of money, compared to what they’re spending in expanding current systems, it is paying a lot less for a system that would be much better.

Is that figure per city? Or overall?

I think this is the figure that I can see for Melbourne, but probably the same amount for Sydney. I think I’m even overestimating what is needed, but the value of that system is going to be very important.

Alright, and I think I know the answer to this, but I’m going to ask it anyway, and see where you go with it. What work have you done so far that you are most proud of?

I think academically and professionally, I’m very proud of what we created in Melbourne, with the AIMES test bed. Because it has allowed us to tap into areas that would have been otherwise difficult to work on. Also, I am very excited about all the collaboration that’s happening across the world between industry and government. I think that’s very important.

As we all know, there are a lot of verticals in transport systems, and unless you’re able to connect them together, basically being able to work with companies that provide communication, that provide services such as buses, trams, and so on, you cannot really obtain an optimum important outcome for the end users.

I think AIMES is an example of the cooperation we need in transport. Even though AIMES is new, and with a lot more development needed, it is a good step forward.

I know AIMES is a long-term project, and it is early days for it, but at what point do you think you’ll start to make – for want of a better word – a meaningful assessment and be able to use it to optimise transport?

I think the good thing about the AIMES project is we set it up in gradual way. From the first day it started to give us insight and provide useful information to universities, to partners, and to government.

For example, when we started to look at the installation of devices in the system we very quickly realised that, for example, it is vital to understand what is the structure of the data structure that you need to underpin the system. There is a massive number of off-the-shelf applications out there, they weren’t suitable for transport, and that was a crucial thing to learn.

Initially we thought we could go to this company, or that company, and very quickly buy the service we needed, but very quickly we realised that no, that’s not going to happen. We need to have a very good understanding of specific data that comes from different parts of the transport system, and to manage it the way it is needed.

Of course this shouldn’t be seen as negative. And to be fair to us, at the start of a massive project such as AIMES, these are things you come to learn, rather than foresee.

Also I know, for example, when we’ve had government work with us, it very quickly becomes aware of gaps in the system, both in policy and in terms of getting its taskforce ready for future implementation. So there has been a lot of learning, even in the very short two-year life of AIMES. We’ve already had some great outcomes, and that’s one of the reasons we have so much interest so quickly from our partners.

And as AIMES gets better, and becomes more operational in different types of projects, we will see more and more good outcomes.

Do you think projects like AIMES are helping us to (again) think in the longer term?

Yes because I think that’s a very important aspect of smart cities. Just giving you one example, Melbourne city has free wi-fi, right? But it was predominantly designed for tourism, for people to browse the internet on their smartphone, or check their emails. But we never really thought, or perhaps even now think, ‘This could be a fantastic platform for IoT. If we have sensors that would like to communicate, we could use that wi-fi system!’ But the answer to the question is ‘Is that wi-fi system designed for that?’, is most probably, no.

This is the kind of long-term thinking that you need to have when you’re designing applications for smart cities and again, I think projects like AIMES allow us to start thinking, and also to highlight the issues. I think that is hugely important.

Installing and analysing data from sensors … you cannot fix a problem unless you know what it is, and its extent, repercussions, and so on.

Yes, absolutely. Some of the problems are so multidisciplinary and it doesn’t matter how many resources you have at your disposal, or how big your company is, you cannot tackle them unless you collaborate across the board. Or unless you really understand the issue first-hand.

I think we all know many transport systems have congestion problems. At the same time we have a lot of people living in cities, which is great, but we have more people living around transit stops, and at the same time a lot of people are using bicycles and walking. We’ve created a lot of conflicts that could result in accidents, or people get hurt. We need to look at all of the usage. We cannot really tackle or have a solution for part of the transport system, we must have a holistic view.

Alright, if we were to tear you away from AIMES, what would you like to get your teeth into next? Or given that you’re so busy, should there be two of you? What else would you like to do?

That’s a very good question. I think what I would like to see happening, what I am very keen to do, is basically to scale-up AIMES. Right now we’re looking at a big chunk of the city and we’re able to provide particular outcomes. That’s great, but we’re also going to be putting pressure on the system that we’re developing.

But if you really have a larger scale city operating in a way that you would like it to do so, would it be able to survive? Can you handle thousands of sensors and get the data moving around and make decisions at a fraction of a second? Can you really do it? Understanding that first-hand I think is going to be a very valuable. I would like to get to that stage.

Basically AIMES plus.

Yes, absolutely because at the moment we’re talking about certain applications, with a limited scope, but if we could ramp-up AIMES, then we could really tackle, for example, Mobility as a Service (MaaS).

A lot of people talk about it but I really find it very funny when people say we have a MaaS application ready to use now. But those apps in their current form, that’s the easy, easy part, the top layer. But there are a hundred steps before you get to what that application does.

I’m very interested to see how this is going to work. Can you come and build up a MaaS system for me that provides values and people will start using? Building a complete system is going to put a lot of pressure on companies, government, and also entrepreneurs. Positive pressure, to be honest, to allow them to show the technology they are developing is useful, and could work in the real world.

So you think that MaaS is pretty much still conceptual?

Yes, I think it’s still conceptual, and neither is it a new concept. It’s a fancy name we put around it, but again, if we don’t have data you cannot provide that service and if you have the data building that application I think that’s the easiest part. That’s my understanding of MaaS.

It comes back to how you’re going to integrate it, because we have private entities such as bus operators, private operators of freeways, and all of them they need to somehow come together to provide that background you need for MaaS.

It’s not just about, “I have an app, you can pay for it!”, that’s the easiest part. I think to get all of these operators and operations happening seamlessly, and also being able to create performance and efficiency through the system, is important because then the government will come. That’s crucial, because at the moment basically public agencies are running the transport systems.

This is all going to have an impact, because when people travel from A to B they’re going to go through all these steps, so how do we ensure provision of a good system, good efficiency, for people? And again, are we going to link all of this together and have a transport-integrated smart city? We can talk now about autonomous, on-demand mobility, and more, which would definitely help. But before we get there, there is going to be a lot of hype, so we need to wait and see.

I gather that you wouldn’t put a date on when a city could truly be considered ‘smart’ ?

I think in a transportation sense we could move very quickly. We are getting close to that critical moment, and also that critical mass. We need to really go and revolutionise our cities from their existing transport systems to a smart transport system, and again if we go back to connectivity it’s already here. If we can deploy properly you’re going to massively enhance the transport system, and fix current congestion and safety issues.

But we need to, and we can do this, make decisions. Now. A lot of people are basically distracting this very important possibility by talking about ,’Okay, maybe this technology is not right yet, maybe we should wait for 5G, or 6G, or 7G.’ I think all of this is a distraction, because those technologies are simply the last layer.

You need to have a policy and start with a good understanding of processes that you need to put in place. From a technical level, all the way up to an operational level. If you delay that, you’re going to delay the benefit that comes to the people. So we need to start that yesterday, and move toward it, and whenever we have a new technology we can just throw it in.

This is, again, the core attractiveness and value of AIMES. From the beginning we said we’re going to be technology-agnostic. Whatever is available today we are going to use it, and when the new technology comes in we’re going to throw it in. But we don’t wait, we don’t wait for that second, third, fourth, fifth, better technology that will come along eventually, because there is no ending to that. Technology comes and goes. The problems need attention now.

That makes sense, treating technology as evolutionary rather than revolutionary.

Absolutely. What is the value of saving people’s lives? I mean we’re spending money by the way already, it’s not that we’re doing nothing. We’re spending money but it’s all about ‘Can you spend that money in a better way? Is there a better way to try it?’ There’s a lot of evidence now that shows the existing technology can save lives. It could improve the traffic efficiency, it could create that multi-modal system that everyone talks about. So why are waiting? We need to invest in it, we need to be brave enough, as I said, if we’re already spending $60-70 billion dollars in Victoria or New South Wales.

We can afford to have a fraction of that going to the efficiency of systems, to smart roads or intelligent transport systems. For example Melbourne has put work into ramp metering, and obtained a significant benefit from it. That’s a very good example of success, which should not have stopped, we need to get going with that again.

I think VicRoads is talking about local road efficiency because Melbourne’s system is similar to Sydney. It doesn’t have lots of freeways, instead it has lots of arterials, that we see carrying a significant amount of traffic. How we could get them to become intelligent and smart? I think, again, this connectivity can help a lot because it’s not just for basic traffic efficiency, it can work for safety, and it can work for releasing the capacity that system has in terms of handling different types of information. You can provide information to end users that everybody is asking for these days, so it could work in different ways.

As is sometimes said, the future is now. Let’s not wait, let’s do it.

Yes! I think with transport we are in a very exciting time. I think autonomous vehicles and automated systems are going to help a lot, but I don’t have a crystal ball for when it’s all going to be here.

But we need to create a system ready for those technologies. And when they come, if we have a good understanding of what we need, we can use those technologies to complement, basically, the performance of a system. Because at the end of the day a transport system is not about autonomous things moving, it’s about people and goods getting from A to B in the most efficient, safe way. That’s the outcome we need to keep front-of-mind.