A Day at the University of Michigan


Going for a run, exploring research labs, and learning how much you can communicate without saying a word.

I’ve been working with Professor Sile O’Modhrain at the University of Michigan as an advisor on my research project, Run Like a Fish. Most of our conversations have been over Zoom, and until recently, we’d only met in person once, when we had breakfast together in July.

So I was really excited to have the chance to visit her in Michigan and spend more time together.

This visit was different. Instead of meeting for just a meal, I got to see a little bit of her everyday life. We rode the bus together between her home and her office on campus, went for a run with her guide, Mikel, and had breakfast afterward.

I also got to meet researchers, learn about some pretty unusual experiments, and help set up a demonstration involving accessible musical instruments.

Somehow, the visit involved self-driving cars, inflatable bladders, and two people controlling the same joystick.

Not exactly what I expected.

Starting with a run

One of my favorite parts of the visit was going for a run with Professor O’Modhrain and Mikel.

I’ve spent more than 30 hours volunteering as a guide runner with Achilles International, so I’ve had quite a bit of experience running alongside blind and low-vision athletes. But this run was special because it was with someone I’ve gotten to know through my research.

Usually when Professor O’Modhrain and I talk, we’re discussing papers, engineering ideas, or what I should investigate next. This time, we were just out running together.

Afterward, we had breakfast, and we also spent time riding the bus between her home and the university.

I really enjoyed those moments. Our first meeting had also been over breakfast, but this time we had much more time together. I got to know Professor O’Modhrain better as a person, not just as the professor helping me with my project.

It was one of the things that made the visit feel different from a normal research meeting.

Self-driving cars and inflatable bladders

One of the people I learned from was Professor Gillespie, who explained a research project involving automated driving and a surprisingly clever way of communicating through touch.

The researchers had developed a driving simulation where a car could navigate around obstacles by itself. There were also gold stars along an ideal path, and the goal was to collect as many as possible.

The automated car was good at avoiding obstacles, but it didn’t necessarily follow the path with the most stars. So the person using the simulation sometimes wanted to change where the car was going.

That’s where the inflatable bladders came in.

The researchers used bladders that could inflate to tell the person what the automated car was planning to do next.

For example, if the car was approaching an obstacle and planning to go right, the person could feel that intention before the car actually turned. If they wanted to go left instead, they could squeeze a bladder to change the car’s direction.

I thought that was such an interesting idea.

Instead of waiting for the car to make a decision and then reacting to it, the person could feel what the car was planning and influence the decision before it happened.

The bladders were essentially creating another form of communication between the person and the computer.

What happens when a machine loses your trust?

The simulation had another interesting part.

At some point, the automated driving system would start making mistakes. Instead of avoiding obstacles, it would sometimes head toward them. The person had to recognize that something was wrong and take over.

Then, later in the simulation, the system would start working correctly again.

But there was a problem: even though the computer was working again, the person didn’t necessarily trust it anymore.

I had never really thought about trust as something you could measure in an engineering experiment.

The researchers were investigating whether the bladders could help people regain trust in the automation more quickly. The results Professor Gillespie described suggested that people with that extra form of communication were able to recover trust faster.

One detail I especially liked was learning about Hannah, a student who had pushed to make trust part of the project.

Apparently, Professor Gillespie hadn’t originally planned to focus on that question, but Hannah thought it was important enough to investigate.

I loved that. It showed me how a research project can change because someone notices something that wasn’t originally part of the plan.

Sometimes the unexpected question turns out to be one of the most interesting ones.

Two people, one joystick

Another experiment I learned about involved two people trying to move a cursor along a path using joysticks.

Each person could attempt the task individually, but they could also work together.

On average, the pairs were more successful together than the individuals were on their own.

The interesting part was how they were communicating.

They weren’t just looking at the cursor on the screen. They could also feel what the other person was trying to do through the physical feedback of the joysticks.

If one person started moving in a certain direction, the other person could sense that movement.

It was almost like having a second sense.

That idea kept coming up throughout the day.

We usually think about communication as speaking, listening, or seeing something. But touch can carry information too, even when nobody is consciously translating it into words.

I had mostly thought about haptics in terms of vibrations, like when your phone buzzes. But these experiments showed me how much more there is to it.

Making music more accessible

I also got to see a demonstration of TAMIE, a University of Michigan research project exploring how to make musical instruments more accessible to people with disabilities.

I helped cart equipment back and forth to get everything ready, and then I got to watch the demonstration.

The idea behind TAMIE is really cool.

Traditional musical instruments are designed around particular physical movements. Playing a piano, for example, requires certain finger and hand movements. But not everyone can make those movements in the same way.

TAMIE explores using machine learning to recognize movements that a person can make and translate them into musical sounds.

Instead of expecting the person to adapt to the instrument, the goal is to make the instrument adapt to the person.

I hadn’t spent much time thinking about accessibility in music before, so it was interesting to see researchers approaching a completely different problem from the ones I’ve been studying.

I also liked getting a little behind-the-scenes look at the work involved in putting together a demonstration. Research isn’t always just the exciting experiment or the finished technology. Sometimes it’s moving equipment around and making sure everything is ready.

And then you get to see what all that work was for.

A different way to think about touch

One of the biggest things I took away from the visit was how many different ways there are to communicate information through touch.

Vibration is one, but there is also pressure, resistance, tension, and physical movement.

We spent some time talking about the difference between telling someone to move in a direction and actually creating the feeling of being pulled in that direction.

When I’m guide running, there’s a physical connection between the runner and the guide. If the guide changes direction, that movement can be communicated through the tether.

A vibration telling someone to turn left is different from the physical sensation of being guided left.

That led to a conversation about whether some kind of wearable device could create a similar pulling sensation.

One idea involved connecting something at the waist to something at the wrist, with cables that could shorten or lengthen to create tension in different directions.

I have no idea how practical that would be, especially while running, but I thought it was a really interesting possibility.

What I liked most was the way we approached the question.

Instead of starting with a particular piece of technology and figuring out what to do with it, we started with the feeling we wanted to create and worked backward.

I hadn’t really thought about engineering that way before.

The connection I didn’t expect

We also talked a lot about the balance between human control and automation.

With a self-driving car, one extreme is that the computer controls everything. The other is that a person does everything.

But there are lots of possibilities in between.

The same problem comes up in robotics, accessibility technology, and systems where humans and computers have to work together.

How much should the computer do? When should a person take over? How does the computer communicate what it’s planning? And what happens when the person and computer disagree?

What surprised me was realizing how many seemingly unrelated research projects are actually asking similar questions.

A self-driving car, a musical instrument, and a wearable navigation device might seem like completely different things.

But all three involve figuring out how technology can work with a person rather than just doing something for them.

That was probably one of the most interesting connections I made during the visit.

What I learned from visiting a research lab

I’ve spent a lot of time reading research papers over the past few months. They’re useful, but they usually tell you about the finished experiment: what the researchers wanted to know, what they did, and what they found.

Visiting Michigan gave me a much better sense of everything that happens around that.

There are conversations that lead to new questions. Students who suggest changing the direction of a project. Experiments that don’t behave exactly as expected. Ideas that sound strange at first but turn out to be worth investigating.

And there are people from completely different fields working on problems that have surprising things in common.

I also realized how valuable it is to talk through ideas with people who have more experience than you. Sometimes they’ll point out a problem you hadn’t considered. Other times, a conversation about something completely unrelated will give you a new way to think about your own work.

I came away with a much bigger appreciation for how creative research can be.

More than just a lab visit

Looking back, what made the visit special wasn’t any one experiment or conversation.

It was the combination of everything.

I loved learning about the research, especially the projects that were completely different from what I’ve been working on. It was fascinating to see how researchers use touch to communicate information, how they study trust, and how something as unexpected as an inflatable bladder can become part of an engineering experiment.

But I also really enjoyed getting to spend more time with Professor O’Modhrain.

Our first meeting was just breakfast. This time, we went running, rode the bus, spent time on campus, and had conversations that went beyond my research project.

I got to see a little of what her everyday life at Michigan looks like, while also experiencing what it’s like to be around people who spend their days asking questions and trying to figure out how things work.

I went to Michigan expecting to learn more about haptics, and I definitely did.

But the best part might have been getting to know the people behind the research.

And I came home with a lot more questions than I arrived with.

I’m starting to think that’s usually a good sign.

This post is part of an ongoing research series at runlikeafish.blog exploring biomimetic approaches to assistive technology for visually impaired runners. If you’re a visually impaired runner, a guide, or a researcher working in this space, I’d love to hear from you. You can reach me at brooke@runlikeafish.blog