Interview with Paul Green, Part One

Paul Green is a professor at the University of Michigan Transportation Institute (UMTRI). He conducts research in Human Factors and Ergonomics, one of the fields that Human-Computer Interaction (HCI) draws upon. Over the years, he has partnered with the local chapters of the Association for Computing Machinery (ACM) Special Interest Group on Computer-Human Interaction (SIGCHI) and the User Experience Professionals’ Association (UPA; now known as the User Experience Professionals’ Association) to help organize meetings and act as a bridge to the student chapter of the Human Factors and Ergonomics Society (HFES).

In this first half of the interview, we discussed his work at UMTRI and the Detroit auto industry.

Driving Safety Research at UMTRI

What kind of research do you do at UMTRI?
We’re doing research on driver interfaces, on topics related to distraction, workload, warnings, menu systems, touch screens, how people deal with the stuff inside the car, how they drive, and how outside the car affects them. We do studies in simulators and on the road.

How long have you been doing this?
I’ve been here since 1979, and I’ve been doing the same stuff the whole time. I mean, the names have changed and the emphasis has changed, but it’s still been this kind of work.

How would you characterize the kinds of things that have changed?
Well, technology has changed and the simulators are better, but the same basic questions that we’ve been trying to answer haven’t been answered very well. We’re doing our best, but people have been thinking about driver workload for years, and then distractions became important. People wrote design guidelines. And now people say, “We should have automated vehicles.” All right, but what’s the workload when it’s being manually driven, and what’s the workload when it’s automatically driven? So the same questions are there, it’s just the way we ask for money is different.

What do you mean by simulators?
We have driving simulators for researcher to use—we have one downstairs. The one downstairs is a full size car, in a big room with all kinds of stuff, and then we have some lower fidelity simulators over in Engineering which are cheaper to run, so a lot of times we’ll use that as opposed to the big fancy one. Also, there are fewer issues about scheduling; I can change things and I don’t have to ask anybody for permission.

Then, finally, we have some simulators that students can borrow. This is unique to University of Michigan. We found some simulator software on the web called OpenDS (Open Driving Simulator) built by the German Research Center for Artificial Intelligence. OpenDS is okay, but it’s supported. It’s not some student’s dissertation, where once they’re finished, it dies. The code is open source. I also went out and bought several Logitec G27 game controllers.

The equipment’s at the library, so when students in my class in Automotive Manufacturing want to do a driving simulation study, they go to the library and borrow the equipment. The library treats it like a reserved book—you can borrow it for a few hours. You download the software and the documentation, figure out what to do, and run your own driving simulator study. It’s been remarkably successful.

Students have struggled with the documentation, but at any other university, it’s like “Oh, we have to go to the lab and go find a graduate student, and then you can’t use a thumb drive because it’ll poison the hardware, and what if you break something? And who’s going to come in to oversee them?” It becomes an administrative nightmare. This way it’s “Here’s the software, you do what you want with it. Go find your own computer—you’ve got one that’s probably fast enough. Here’s the hardware. We’ve made it so it’s going to be hard to break it. Plug it in. Do what you want to do.”

We’re probably one of the few universities in the world that use this library mechanism, and it kind of opens things up for students because up until now the only universities that could do driving research were the ones with a lot of money—that could spend 50 to 100 thousand dollars to buy a driving simulator, or where the students spent the first year of their dissertation writing some code, and they came out with something that was really clunky and wasn’t very good. This equipment overcomes that problem.

Detroit’s Automotive Industry

Did you have a previous act before 1979?
No, I got my PhD from here and looked around for jobs, and came to UMTRI. I had very minimal contact with UMTRI—then called HSRI (Highway Safety Research Institute)—when I was a graduate student, both in Engineering and Psychology. After I was here for a while, they asked, “Oh, do you want to teach?” and I said okay. It’s unusual—usually people go away after graduating. I didn’t, but it’s enough of a shift that it was an intellectual broadening.

But if you want to do motor vehicle work, it’s really good to be in the Detroit area.

How much have you interacted with the auto companies over the years?
Most of my money comes from the auto companies, so I interact with them a lot. It depends on the company, but recently we’ve gotten funding from Hyundai and from Nissan, so I’ve interacted the most with them, and from the others, it’s been a much lower level of interaction because with them there’s no funding. Professionally, I’m on a bunch of SAE (Society of Automotive Engineers) committees. I participate in all those activities. There are technical committees, writing standards—it takes a lot of time.

Are you aware of Human Factors professionals working at the auto companies?
I know many of them through SAE, and they know me because of what I’ve done.

Have auto companies always had these people in house since you’ve been on the scene?
Yes, and I’d say, over the last 10 years, those groups have grown considerably. After Chrysler came out of those bad times, the human interface group at Chrysler grew tremendously, and there’s been growth at Ford and at GM, and now we’re seeing it at Hyundai and Nissan. All the groups have grown.