journal
A conversation about delivery and transportation in an uncertain future with Costa Samaras, Director of the Center for Engineering and Resilience for Climate Adaptation. Costa is also an Associate Professor at Carnegie Mellon University and an Adjunct Senior Researcher at the RAND Corporation.
From understanding how to make decisions under deep uncertainty to researching climate resilience, automated vehicles, and low-carbon energy systems, Costa Samaras is anticipating Urban Air Mobility problems with a holistic, integrated approach. We spoke to him about scanning extractive capitalism, building life cycle models, and why you shouldn’t discount vanilla technologies.
What I do is try to solve hard problems. I'm on the faculty of the engineering department here at Carnegie Mellon University, but I also do a lot of work for cities and governments who want to try to understand technology evolution and technology integration and how that would affect the world. Much of the tech world is based on hitting performance indicators or metrics, but most of those folks don't really have any training or interest in the policies and social or environmental impacts of their work, and don’t always consider the concerns local stakeholders might have.
The work that we do centres around three trends. One is climate resilience, the next is the future of energy systems, and our third focus is future transportation. So that really started off with our deep understanding of electric vehicles, which has now transitioned into automated vehicles and other non-vehicle modes of transit.
The work that we do centres around three trends. One is climate resilience, the next is the future of energy systems, and our third focus is future transportation. So that really started off with our deep understanding of electric vehicles, which has now transitioned into automated vehicles and other non-vehicle modes of transit.
Sure — this current project that I'm doing now is around the future of delivery. We're driving sidewalk robots and flying drones and driverless cars to try to understand both the energy use, the noise, and other factors that these might have when they are put into an urban delivery network.
So we’ll watch the data around traffic in a certain area and simulate injecting a bunch of bike delivery, drone delivery and sidewalk robot delivery to see how those things change and how might we design the system efficiently. Because before all this tech gets integrated, we need to make sure that the outcomes that people want are enforced.
We had one paper last year that looked at drones alongside regular delivery vehicles, electric vehicles, personal vehicles and different fuels. In the case of drones, we found that for small packages they emit the lowest CO2 per package delivered than many of the other modes.
But the challenge is that if you need additional warehouse space to house additional urban products — as you would for a drone-based delivery network — those are going to have second-order upstream lifecycle effects. If I need more square feet to deliver the same amount of products, then I'm going to have to heat or cool that warehouse area as well. So when we build models, we need to always build the model according to the life cycle of what we’re looking into.
You regularly see in the news some consultancy or professor saying something like, ‘100% of vehicle travel in 2030 is going to be autonomous,’ and I find that ridiculous. It’s not going to happen that way.
We're definitely evaluating how drone deliveries would work in the urban context, both on the sidewalk and in the air. It's not guaranteed because there are key factors such as noise, but if society decided that they want these, then there’s also a problem when they take up sidewalk space, it's also bad for people who use wheelchairs and other mobility devices. There are several other pros and cons, of course.
But if society decides that they want these, we're thinking about a number of factors. Firstly, how would one design what might be thought of as a highway system, both on the ground and in the air — a virtual highway system? What are the rules? How high does it go? What happens on the sidewalks? What are the hours of operation? If you imagine that you're going to have to redesign the whole urban space to accommodate different types of vehicles, what is the virtual architecture that would allow this to work? How do the delivery vehicles make the final handoff — whether they’re in the ground or in the air?
Architects have now been prototyping what are called rota ports, which are hubs that people will use to take these vehicles and travel themselves. I don't know if you've seen Uber Elevate or other services that one might call a flying car, but it's not really a flying car, it’s an electric drone that can carry passengers; as ridiculous as that seems.
For example, a person can get from Schiphol Airport to downtown Amsterdam pretty fast on the train. But there is definitely a market for people who are willing to pay 200 euros to take an electric drone to and from the airport to the city in just 15 minutes or less. How might we regulate that? How would we integrate them into buildings — both structurally and energy-wise? That might seem like some kind of dystopia but at the same time we should not be ignoring it, we should be thinking about these issues already.
Urban Air Mobility is not just about package delivery it’s also about people delivery. We all have to think through how we move people and goods through a city in a world with increasing automation.
I'm on a national committee that’s charged with defining the future of Urban Air Mobility which is not just package delivery, it’s also about people delivery, and there's a whole bunch of people who want to do this. My job is to think through some of the environmental infrastructure and policy implications of that and make recommendations to keep the public interest in mind.
But a designer, a city designer, an urban designer, a building designer, an architect, they will all have to think through how we move people and goods through a city in a world with increasing automation. That’s all connected into a lot of the work that we do around sensing and resilience as well as automation.
You’re right. The heart of our methods in the lab are based on what’s called in the field ‘deep uncertainty’, and decision making under deep uncertainty. So we apply the methods of that discipline to these questions. When will autonomous vehicles enter the urban space? How fast will they be adopted and how many? What will be the electric use of these? What will vehicle travel look like?
It’s a similar approach to climate change. How much will it rain in the future? How can I make a decision now to be adaptive? What's the cost? What are the options? What are my consequences of choosing one thing and then getting locked into that pathway? So that method is called decision making under deep uncertainty, and we apply that to all the stuff that we do. I'm a part of the Society for Deep Uncertainty and they use this method.
I think that one of my biggest complaints is that you regularly see in the news some consultancy or professor will say something like ‘100% of vehicle travel in 2030 is going to be autonomous,’ and I find that ridiculous. We tried to physically model how that would work and the pathways that it would take to get there, and I can tell you — it’s not going to happen at 100% by 2030. Right now I'm trying to do a paper for the United States looking at how fast could we turn the vehicle fleet over into electric and what would be autonomous due to that.
We will run out of a livable climate way before we're going to run a place to store full batteries.
Those issues are important and we for sure deal with them in the kind of end-of-life and extractive industries, but I want to stress that we will run out of a livable climate way before we're going to run a place to store full batteries. The scale of the climate problem is going to require a transition to a more ethical system. However, there's still going to be ethical trade-offs that we're going to have to think through and think through together and not dismiss.
Transportation has provided incredibly amazing benefits, not just in the US but around the world. However, it's also created some really bad social outcomes, including inequality, racism and pollution. People knew those things before and they didn't care, but now we care and so if we are to transition to the next stage, we have to incorporate all those concerns from the very beginning.
E-bikes are the plainest vanilla technology that we have right now which could change a lot, and really fast.
I really like the idea of finding ways to reduce the unnecessary single occupancy car trips, and we have a lot of those here in the states. The most boring technology that is available right now are e-bikes and I would love some kind of policy program to emerge that vastly scales up e-bike tax credits, subsidies and infrastructure, because I feel that that is an enabling technology, not in the future but today — this afternoon!
Of course, we still have to worry about accessibility for folks who can't ride bikes, but I think e-bikes are the plainest vanilla technology that we have right now which could change a lot, and really fast. In the future I am excited about automated brake delivery on the highway with electric and autonomous vehicles and automated air logistics with larger electric vehicles. Those are all important and exciting but really, we have to get CO2 output to zero and how we do that is by going after the big wedges which are personal transport and electricity.
If we can get those close to zero as possible in the next 30 years - which is by no means easy as it's a huge undertaking - then that will enable a whole bunch of other exciting things to happen that we don't even know about yet. I'm kind of lucky that I'm working at the edge of all this.