Streetsblog Interview : UCLA Researchers Who Examined the Systemic Impact of Autonomous Vehicles
Last month, UCLA released “Connected and Automated Vehicle Impacts in Southern California: Travel Behavior, Demand, and Transportation System Perspectives,” the first research study that examined the impact that autonomous vehicles would have on the transportation network in Southern California. The policy brief and full research report are online.
It is anticipated that switching to autonomous vehicles would benefit safety and efficiency, but the report warns there are also drawbacks. In the study, a survey showed that increased car trips and vehicle miles traveled would be a side-effect of the transition. The authors assert that the region needs to start planning on how to mitigate these increases.
SBCA had a chance to sit down with two of the report’s authors, Jiaqi Ma and Yueshuai He, to discuss the report, the methodology and the findings. The transcript is lightly edited for brevity and clarity.
Damien Newton (DN) :
On December 12 of last year, your team published, “Connected and Automated Vehicle Impacts in Southern California: Travel Behavior, Demand, and Transportation System Perspectives” a policy brief with UCLA. To start with, what interested you in studying connected and automated vehicles (CAVs)? And what were sort of the big findings that came up as you were going through it?
Jiaqi Ma (JM) :
When we first proposed this research, we realized that there’s a lot of praise about these new technologies, either automation or connectivity. Everybody’s talking about the benefits: how it will save lives, how it will improve roadway capacity. And I want to make a distinction, the connectivity, the CAV, you had three letters connected and automated, it doesn’t have to be together. C is connected with vehicles that have communication capabilities that will improve traffic. That’s the idea, improved safety, but automation, as a consumer surveys, helps you drive the car automatically. It also potentially improves safety by getting humans out of the loop.
There’s a lot of reports about how well this technology will improve safety, mobility, efficiency, and environmental aspects of the transplant systems. A lot of these are looking at various operational level technologies. They are looking at how vehicles interact with each other locally, and so on. But then we think about them, there has to be some system level impacts to transform our transportation systems perspective. People’s travel behavior might get impacted. Because of the ease of driving, people may travel more – which then in turn increases congestion from some perspectives. Are there these benefits? Or, are there additional benefits? So we just want to understand from a bigger system level, we look at the literature, nobody’s ever done that. And we will want to see if it will actually be good from a system perspective or induce a lot of additional trouble actually making traffic worse. That’s the very simple logic behind this study.
If I understand the brief correctly, do you believe that there will be some induced demand and possibly increased congestion or decreased congestion as a result of a switch over to CAVs?
Our approach is using activity-based models and system level simulation. We collected data on people’s travel tendencies and their behavior, their preferences. Through a complex modeling process, building upon the model that has been developed by the Southern California Association of Governments (SCAG), we developed additional models building upon it to incorporate these CAV/AV components. And we’ve developed additional models building upon it to incorporate these CAV/AV components. And we do find that people travel more. And they’re willing to travel longer. And this will in turn cause congestion…more congestion and potentially more emissions as well.
CAVs may increase road capacity because they just work closely. They’re not driving like humans, some humans drive crazily, which will cause congestion. CAVs by driving smoothly, they will increase roadway performance. But this additional demand will induce travel. It will actually, sometimes, cancel all the benefits, in some cases even making the traffic condition worse.
That’s an interesting point because a lot of people are assuming a move to CAVS means it’s going to take less time in a car to move from point A to point B. But because there’s going to be more cars on the road because there’ll be more travel and longer trips, you’re saying that we won’t actually see a time savings for people that are driving now versus riding in a CAV at some point in the future.
It’s highly dependent on the actual network, people’s behavior, their willingness to use this technology and their behavior when they have access to this technology. We collected data in Southern California. So a lot of the data we have done is for the analysis we have done for Southern California using the Southern California residents data. So that’s the potential impact here. On a different network, on a different behavior set, there might be different results.
One of the things I thought was interesting was that you talked that not just about that would be increasing trips, but they might also be increasing the length of the trips. The number of work trips probably isn’t going to change by very much, but the number of trips that weren’t work trips would be increasing more of I think you said almost 10 percent. That just struck me as something interesting. And I was wondering if you had an idea, why it is – just because people would perceive it as easier? Or is there some other reason?
Yueshuai He (YH) :
Yeah. Basically, in our survey, we give an assumption. We talk to people, “Here is the condition: you don’t need to drive anymore.” We explain what a CAV is and then let them make the decision. How many trips? Whatever purposes did you take trips for before? And what if you have these new services, what will you do next?
Based on the data, we found out there’s a trend of increase for both the trip numbers and trip lengths. That’s what we identified from the data. We don’t ask questions such as, “Why did you choose that?” One possible reason is that the ease of travel by CAVs will induce people to travel more, especially for non-work trips and trips at nighttime.
A lot of people when they look at the sort of global warming crisis that’s here and worsening, they look at technology as the probable solution. But at least as far as switching over to CAVs, we would actually see an increase in greenhouse gas emissions over what would happen if people continued their driving patterns they have now.
In some networks…
In this network. In Southern California.
Yes. And the key metric is vehicle miles traveled. That’s what the VMT nine percent increase is that we talked about in our report. That’s a key performance metric that the Department of Transportation, or Caltrans is very interested in because their whole goal is reduced VMT, reduced vehicles on the road and potentially reduced emissions.
But we see VMT increase in the L.A. network based on our modeling. And then it is this problem that should be addressed. There are a lot of benefits that CAVs bring to us, but with these benefits we also propose some kind of transportation demand management strategies to curb that increase in VMT.
The things that are listed in the brief: parking, pricing, remote work, or auto trade and things like that; those all seem like things that would be a good idea whether or not we move on to CAVs. Were there any of those that really stuck out that was something that is almost like a would be a requirement as we’re looking at possibly transforming our fleet into more of a CAV one?
The policy or strategy we proposed is from the SCAG operations. They have a travel demand management manual. And we have listed a couple of their existing strategies they use to manage the travel demand of an existing driving system. Our objective here is to propose some policies that are ready to implement, to curb the negative impact of the CAVs for the near future. That’s why we use existing strategies adopted by the agencies.
I assumed that switching to a CAV fleet would be something that would increase the disparity in equitable access to our transportation system. But as you were breaking down the research and the information, you saw some pros and cons… not just a concern that the more expensive vehicles would reduce access. Could you discuss what the pros and cons are as far as more equitable access to the system?
In the survey we asked the question, assuming people don’t have to consider the cost to purchase a CAV. From the government side, sometimes they have a subsidized system to support low income communities.
In our survey, our assumption doesn’t consider expenses, to purchase a vehicle, then how you (people responding to the survey) are going to change your behavior. That’s a pre-assumption we made. And based on that, we still find that both high income and low income communities benefit from the accessibility or like the efficiency. They both benefit from it. But there are disparities, like high income people may gain more than lower income people (even with the cost of purchasing a new car removed.) That’s our findings.
In terms of number of trips, and travel distance, and also travel benefits, including travel time savings and things such as that. We found that this technology may benefit both high income and low income communities, but the disparity between them will increase if governments do nothing.
I always like to, especially with these interviews that are about research just bluntly, ask at the end, “is there something I missed in here that really struck out as interesting or important to you that you’d want to make sure that we highlight for our audience?” And it’s okay to say “no, you did a great job.”
You did a great job.
But I do want to mention one more point. When this report came out a lot of people thinking, “Oh, you know, you’re against CAV technologies.” But it’s completely different.
Personally, I’m definitely pro-CVA as well, I’m definitely pro this technology. But the key thing we want to communicate to the whole world is about more system level analysis of these technologies.
And it’s not just CVA, we talk about EVs. Our team is using a similar approach, looking at EVs as well. So when this technology comes into play, it will… I wouldn’t say disruptive, but it’s actually a non-trivial impact on your system. Induced demand perspective.
So these effects have to be accounted for. When we started this report, we did a very comprehensive review of all these modeling prediction efforts. We were very disappointed that we saw people talking about the CAVs for so many years, and nobody actually takes this induced demand analysis seriously in the literature. So we actually took on this task and did his work. Again, technology is good, but there was study, in fact, maybe to predict (VMT) well, and using in this case, transportation demand management strategies to control this.
Our study is based on a near future. It’s not 50 years from now when we have a fully automated system. Our assumption as seen in the report, only half of the people who might like to use it. Not everyone would like to use it. So that’s one insight we need to some like AV adaptation to persuade people to accept this new technology because sometimes they don’t trust it. They don’t understand that so they don’t trust it. From the system parts, our assumption is that CAVs are running with human driving vehicles. It is not a fully automated transportation system.
That’s the reason why the benefits are not fully leveraged because we consider a mixed traffic condition. That’s why we have induced demand. We have improvements in the system but not on a very large scale. So that’s why in the near future, the benefits are not as large as imagined.