What is the incentive of a manufacturer, who already has permission to operate an autonomous vehicle, as exists now, to participate in developing these models?
John Harding: That’s a great question. They’re already out there. They’re doing their thing. But what I would suggest is that what they’re doing is a little bit limited because they’re just looking at their ADS system themselves and trying to make sure that perceives the current roadway. And I would question, are they getting into any of these challenging cases like work zones or other kinds of different design features, or just traffic control device features like, say, a roundabout? So, what I would do is say, “Hey, if we collaborate together, we can think about maybe how you may be perceiving these other kinds of roadway environments.” That gives both sides some information about…the ADS developer gets information about how they’re interacting with that. But from the roadway side, what we want is to make sure a roadway provider understands how they’re presenting those different areas. And maybe there’s room for collaboration on different ways that things can be presented, and the ADS can then also include and expand from their ODD or expand their ODD to include these different types of railway features that may be something that they really haven’t addressed yet. This is going to be a collaboration, because you need to integrate—you know, it’s a system interface, and so you need more information to understand how this ADS interfaces with the roadway. I would think that both sides would win in this situation. And, hopefully, the case can be made that, based on that, they will have…the ADS developer wouldn’t have to have this extensive algorithm if they knew better how they interact. And the roadway guys could do something to say, “Hey, with this enhanced way you can interact, less of an extensive algorithm enhances the safety of the system together and allows them to address other areas.” Any of my other colleagues want to chime in on that?
Nick Hegemier: I can chime in just a little bit. I think, as you looked at the technology evolution chart on my slides, it’s pretty evident that we understand what situations and the areas in which those situations occur. What situations cause the most challenges for the transportation system? And we’ve spent thirty years trying to deploy technology in innovative ways to overcome some of those challenges and issues. So, working and collaborating with us could help you identify what situations may be prevalent that you may have not thought of yet, and incorporating those into your simulations, and then adapting your technology to meet those needs.
John Harding: Thanks, Nick. You know, we should make…you use the simulation which is also good. This framework covers simulation, test track, and roadway, so you can move along the spectrum and work it out. Get some input—like you’re saying, Nick—from the road providers on the simulation, so they can accurately understand what are the different situations that are portrayed in the simulation and work it through the system. Maybe you can move the test track next and then the roadway. So yeah, good perspective.
How has FHWA seen collaborations be successful? How can they be improved?
John Harding: I can’t remember the examples off the top of my head, but there’s other examples in the report of collaboration that have been successful, that have provided both sides with the information they needed to move forward. So, you can work through that in the report. And there’s multiple incidents in different areas and that’s what we pulled from. Those were successes and we tried to pull their experiences into the report.
Martha Eddy: Nick gave us a link that maybe we could pull up some more. We wrote this a couple of years ago now, so some of the more recent ones may be through what Nick provided
Nick Hegemier: The link that I shared is going to become live mid-June. That’s where we can share a lot of lessons learned, ask a lot of questions, get those clarifications and collaborations that are needed.
A common response to that perspective is “good-state-of-repairs.” Trying to drill down into that one is difficult. Sometimes the ADS needs are different than the human needs, and vice-versa. This is where we could make improvements, right? What else is there that could be helpful too?
John Harding: Sometimes they say, well, what’s good for the human is good for the ADS. But I think that still has to be worked out a little bit more. And I think this collaboration will help to make sure that that assumption is valid or not, and it may be valid in some cases and may not be valid in others. There’s instances where, I’ve seen in research, that ADS really have a problem with roundabouts and it depends on the configuration of the roundabouts. How does it perceive the roundabout, and how should it navigate around the roundabout? So, yeah, there’s a lot to think about as we move forward and check some of the things that maybe people are assuming that will move us forward. It’s always—like you said, Nick—good to have information, so we can move forward together. I look for a future with automated driving systems integrated into the roadway system. I think a lot of us also do. I like to say continuity and investment could be needed from the industry. If we know what we’re going to invest in, in helps them understand how they’re going to invest into ADS developments and their ADS products in the future. So, I think, everything works out well.
United States Department of Transportation
