RDS Updates and GuardDuty Threat Detection Enhancements – RDS for MySQL now supports authentication with active directory. RDS snapshots can now...
Planning for Data Driven Smart Cities
All of this is linked to what we’re seeing today did you see in this picture is this kind of convergence between the digital and physical world. One way to say it is that the internet is entering physical space. The internet is becoming Internet of Things. And so it’s creating a new condition in our cities, what many people refer to as smart cities.
And really what is behind it is that we can collect a huge amount of data from cities. We can sense them in a different way. But then we can use this information in order to respond to it. So it’s about sensing and actuating cycles of design, where we can actually better understand the city and then transform it. I will start by looking at urban infrastructure. In particular, I want to focus on bridges, in particular is a very famous collapse of a bridge called the Tacoma Bridge.
I’ll tell you another little story about this bridge, which is called the Longfellow Bridge. It’s in between MIT campus– you see the beginning of it to the left. And to the right, there is Boston, downtown Boston. Naturally, Longfellow Bridge was working fine until one day, they realized it has very big structural flaws. And then it had to be shut down to be almost rebuilt at a huge cost with a very large investment.
This whole city was divided into two for a number of years. And most of this is because it had be monitored accurately. And the way we monitor a bridge today, as you see here, is using sensors. As you see, to the left is the Golden Gate Bridge, also with visual inspections. But both methods are very, very expensive. In particular, sensors– we couldn’t put sensors on all the infrastructure in the United States.
But they did something interesting. And the interesting thing is that all of us have sensors in our pockets. And through those sensors– smartphones, and so on– we can get a lot of information. For instance, as you see here, the information we usually get from sensors on bridges, which is vibration information– information about the frequencies of the bridge, which is what we could call a structural fingerprint of the bridge.
So what we decided was to take a number of the sensors, a number of smartphones, also a professional accelerometer– you see it there in orange– and start testing. The question was, can we use this instead of very expensive fixed sensors in order to understand the frequencies of a bridge? And we started from a very famous bridge, from the Golden Gate. And the reason is that we know everything about the Golden Gate.
In other terms, we got a very good ground truth. So we started going back and forth across the Golden Gate– around 100 times, until we were stopped by the police. They got suspicious. Here is that the type of raw data we collected from it. And then here is the analyzed data, the clean data, with all the frequencies. And it turns out that actually by looking to all this data collected from the smartphones in the car, we can reconstruct the key frequencies of vibration of the bridge.
So it becomes a very interesting way to scan the infrastructure. Now I will not tell you too much about this. But, basically, by using the accelerometers, we can actually learn a lot, first of all, about the structural health of the bridge, then also about the asphalt and about potholes. And so this information then can be the first scan.
Who are they exactly? This is an entry-level position, playing an auxiliary role in the analytics department of a company. That means they support bot...
Have you ever been in a situation where you spend a lot of your time and energy learning something, only to realize that the skills you have gained do...
Backup Explorer in Azure now in preview and Azure Key Vault private endpoints in Azure Private Link – Have you ever forgotten who you’ve g...
Azure VNet Address Translation, Lition Blockchain and Azure Stack Kubernetes support – When I think of virtual networks, I think of the most sta...
Then if we find something doesn’t work, we can do additional scans, more sophisticated scans, in order to see where we may need to intervene. And we are starting now to do the same thing at a scale of an entire country, which is very interesting. With very little money, using sensors that we all keep in our pockets, we can basically collect information about thousands of miles of road and then try to make them safer. Now this is about using data from smartphones in order to better understand the road infrastructure.
But we can also better understand traffic on the road infrastructure. For instance, look at this. This is the city of Lisbon, almost like a living organism, seen through billions of billions of data points collected from the taxi network in the city. Again, it’s digital data that’s collected, almost in real time, from the road network. Now, the same data we saw in Lisbon is the same data that you see here.
That’s again taxi pick ups and drop offs. The little yellow dot is a pick up. The blue dot is a drop off. This is JFK airport in New York. If you zoom out, you see JFK down there in the lower part of the image. You see Manhattan here and all of the borders. By the way, if you look at this picture carefully, there’s no roads there. The roads just come out because the incredible density of dots. So it’s not the road map, just the dots– the pick up and drop offs– are what define all of the network.
And then when you’ve got these data, you can then mathematically ask a number of questions. The first question we asked was, well, you know, if you look at those two points in Manhattan, in the course of the year you’ve got 100,000 of trips connecting them. Then how many of those trips could potentially be shared by two people– if people were ready to share a car with each other.
Now, if you want to answer this question, you cannot do it traditional mathematics. We had to develop a new system that we call shareability networks.
Here we see the paper. Here we see some of these images of shareability networks. And what we discovered was quite interesting– that in Manhattan we could actually take everybody to the destination when they need to be there, but with 40% less vehicles than what we have today, if people are ready to share. Two interesting things happened.
The first thing is that the first results came out before Uber launched Uber Pool. And after we published the first results, actually Uber reached out to us. We started a collaboration between MIT and Uber. And as most of you know, actually Uber Pool allows to do this– that you have two people going more or less in the same direction, to share the ride.
The interesting thing is that when the first results were published, there were a few reviews in newspapers and magazines. And I remember a review in The New York Times that went something like this– you know, well, this is interesting academic research. But you know, New Yorkers really don’t want to share anything. It turns out that that is not the case. And now that we’ve been working with Uber, using a lot of their data together with them, then Uber Pool has been a big success.
In San Francisco, I believe, it’s over 50% of all Uber trips every day. Which means two people going more or less in the same direction sharing the ride. Which means one less car on the road and also less energy consumption and less pollution in our cities. So this was a little story about how data can help us better understand the city and also inspire new mobility systems in the city.
This is a more recent piece of work from just a few months ago, where again we took a similar data set. And that’s again real data from New York. You see it visualized here. And here we ask another question. It’s not about sharing a ride. It’s actually about what’s the optimal way to do dispatching in the city. And again, today’s system has a lot of inefficiencies. You know, for instance, taxis are vacant half of the time. And then if we do better mathematics, we can actually correct a lot of that.
And so if you look at this, the minimal fleet in Manhattan is much smaller than the one we have today. Here you see how the algorithm works by doing better dispatching by matching one trip with the next one in a more optimal way. And the result of this is that you sit here– to the left and to the right– to the right, you see the situation as it is today. To the left, you see how it could be tomorrow, with much fewer vehicles actually satisfying the same mobility demand of Manhattan.
Now if you look at also a future when cars are autonomous, again this can tell us that with an autonomous car and good dispatching, we could satisfy the mobility demand of the whole city with 50% of the cars we have today. Autonomy is a very interesting thing. And I will not tell you much more about autonomy, but we’ll change a lot the way we actually run our cities.
One important thing is that today, a car is parked 95% of the time. It’s used only 5% of the time. And when it’s used, it’s used by one or two people, even if it is a five seater. So the efficiency of the car infrastructure is just 1% or 2%. And with autonomous systems with self-driving cars, we can do something different. The car can give us a lift in the morning when we go to the office, and then can give a lift to somebody else in our family or to anybody else in the city.
So we can create a new system in between public and private transportation. Now things can also be more difficult. You see some of the references– we got both positive scenarios and negative scenarios. But certainly this is going to have an impact on our cities. So now, then, when all of the cars would be autonomous, other things may change.
For instance, this– this is a well known traffic light. And traffic lights arrive on our roads when cars arrive on our roads. But if you’ve got the intelligence system where every car knows where it is and where all the other cars are, you don’t need to stop anymore. You can keep on going, just avoiding collisions like this. I would suggest you don’t try this yet.
I need to tell you a little story. I showed this video the city of Naples at some point. And then people told me, so what is new here? Now, I’m Italian. So I’m allowed to make such politically incorrect jokes. A former Italian minister said that, in Milan, traffic lights are instructions. In Rome, they’re suggestions. In Naples, they’re Christmas decorations. So, however, the future of autonomy may actually bring us back to conditions in cities which are similar to the one in Naples, which ultimately is more optimized in the way it uses space– even if it has to be done in a much safer and programmed way.
Now all what we saw so far is about using data to better understand cities, and then model cities, and then maybe propose new mobility systems in cities. I wanted to finish with some of the design work we’re doing with our design office on mobility. This is a project called Smart Road, where we have been developing a
new system for thousands of miles of highways in Europe.
And in this case, the idea was again, how do you design a new poll that can help with sensing? Not always sensing on the poll, but also sensing on drones that can help us to see what’s going on on the highway and also run there if there’s an accident? So help collect information very quickly if there is a problem. This is another project we’re working on with Google sister company called Sidewalk Lab in the city of Toronto.
It’s an experimental district. And here, for instance, part of the brief we got was, how can we rethink the road when every car is autonomous? And what you see here, you see some of the images of this road that in the morning can be programmed like this. We’ve got a lot of cars– self-driving cars– and then a few hours later can become a place for kids to play– or in the evening, a place for a party.
And it happens by both digital programming– by changing lighting and direct conditions of the road– but also by physically programming– by actually taking elements and moving them around. So this road becomes a part of the dynamic infrastructure that can be used better without that rigidity of 20th century infrastructure. And just to finish, a few additional images– that’s another design project we’re doing with the city of Paris.
It will be exhibited soon at a place called Pavillon de l’Arsenal in Paris. And it’s about rethinking how the urban highway in Paris, called the Peripherique, could evolve. And so what you see here is, again, the same idea of how lanes could be programmed dynamically in different ways, much beyond what already is done today in different parts of the world in a physical way– but here with self-driving and digital information.
Also how the asphalt could become productive. It can become something– a piece of surface you can use in order to collect energy from the sun with photovoltaics, and use the same energy to move people around. You see it here, the same energy that’s collected is then being used to charge the vehicles. Or think about the future, where because of e-mobility– electric mobility– and also autonomy– which is called EVs and
AVs, electric vehicles and autonomous vehicles. Then we’ll have most likely less pollution next to big roads. We’ll also have less noise, and everything could be safer.
So if we get to the conditioning, we can inhabit really the infrastructure in a different way. As you see here, we could have plazas much closer to the roads. We could have buildings, residences, offices that really bridge across what, in the past– especially in Paris– was a big fracture between two sides of the city. So we can close it again in a way that allow to a better mix between people and machines and infrastructure.
Now I will finish here. I wanted to share with you some references for what we’ve been discussing– some of the papers I’ve been citing– but also get to, in a nutshell, the key for today, which is really that data can help us better understand the city in general, better understand infrastructure, monitor vibrations, monitor other parts of urban infrastructure.
And also data is the beginning for a new type of design. It takes data, is driven by data, and it can help us to imagine how tomorrow’s cities could be. Thank you.
reference – Planning for Data Driven Smart Cities
Total creator. General coffe buff. Award-winning internet trailblazer. Devoted tv practitioner. Gamer. Communicator. Travel fan. AI and machine learning are everyday!