Years ago, in 1979, Ella Atkins was one of the first students to participate in a Johns Hopkins University SMPY program, which began in 1971. In this program, she would study mathematics. A grad student from Johns Hopkins was assigned to her as a mentor.
“In the very informal correspondence that went on between us, he was offering materials that would give me the opportunity to explore options,” said Atkins. “Some assignments were about math and science, some were about engineering, including aerospace. As I read those, I liked the look of airplanes.”
Atkins grew up in a small, rural community in West Virginia that had a fighter aircraft training route over her house. She enjoyed watching the planes fly overhead. She was also influenced by the NASA space programs.
“I don’t know that I ever really put astronauts as number one on my career list. I was more interested in designing the spacecrafts and have someone else fly them,” she said.
Atkins is now a Professor of Aerospace Engineering at the University of Michigan — also a core member of faculty in the Robotics Institute. Currently, she is on a yearlong sabbatical to work as a technical fellow for Collins Aerospace, which is now part of Raytheon. She is a senior member in IEEE and a fellow with AIAA, the aeronautics and astronautics group.
Her undergrad and masters were in AeroAstro (aeronautics and astronautics) from MIT. In her master’s studies, she did research in space robotics. Then she went into the aerospace industry for three years.
“One of the things you learn in space robotics,” she said, “is that the electronics and the software are a tremendous part of what you’re doing. The mechanical systems support the electronics and the software. So that knowledge led me to want more foundational knowledge in computer science.”
So, she received her PhD in computer science and engineering from the University of Michigan.
Taking things apart
Ever curious, Atkins tended to take things apart to see how they worked. One thing in particular stands out in her memory: dismembering Barbie dolls.
“I didn’t care how they looked, I cared how their legs connected. So, I would take all of the pieces and parts off of them, I’d behead them and remove the arms and legs. This frustrated my family; they thought I was quite destructive. I tried to explain to them that I was figuring out how they worked.”
This experience, though, led her to study robotics, especially robot arms used in space.
“Zero gravity is a big deal. If you have gravity, then when you lift an arm out to the side, you need a certain amount of strength in each of the joints of the robot to counter the weight of each of the pieces. In space, you don’t need that. If you look at a robot arm on Earth (NASA Goddard might have one that emulates what was on the space shuttle in the space station) you’ll find that they have to put pulleys and counterweights on the arm because it can’t hold itself up, but that’s not the goal,” she said.
“And there are additional challenges in space — for example, when one part moves, the whole unit moves. So, if a robot arm is servicing a satellite, when you move the arm out, the spacecraft moves in response because there’s no grounding. [Plus], the conditions in space are so harsh. The vacuum, the coldness, the temperature change from being in the sun to being in the shadow of Earth, for example, when orbiting. And you have to carry all of your own energy, there’s no plugging it into the wall to recharge it,” Atkins said.
As problem solvers, engineers enjoy a challenge as it’s an opportunity to learn something new. Atkins challenges the notion that once a problem is solved, it’s “done.”
In academia, many of the engineering challenges students or professors write about are not always done. Explained Atkins, it’s hard to be “done” as a researcher. You may have developed a proof, an analysis, or an algorithm, and you write about it, but you’re not done.
One example Atkins gives involves UAV autonomy. Atkins and some of her students have been working on autonomous flight, analyzing the use of multiple quadcopters to hold a payload. One of the goals is to have a person guide the payload by pushing it.
“That’s a natural haptic interface if you have a payload that’s carried next to a human. We started this project right after the hurricane disaster in Puerto Rico. We saw supplies sitting on docks because roads were blocked,” she said. “People were drinking contaminated water, they didn’t have enough food or medicine, and so on. With medicine, the payloads are often small. With water, you actually need to carry a pretty heavy payload to get enough to where you want to go.”
“We looked at the notion of carrying something that might be a 24-pack of plastic water bottles somewhere. Rather than have a big vehicle with large propellers, we looked at whether we could have many quadcopters and attach them with tethers to a payload. With that, you have redundancy, one quadcopter could fail and the rest of them could still carry it and drag it along, it wouldn’t crash.
“We looked for a way to guide the copter without the need to look at a cell phone or a laptop. Because the reality is, you don’t want an emergency services worker with their head down looking at a laptop. So, we explored the idea of pushing the quadcopters. You can guide one along until it gets out of a crowded area and then it can move on its own to the needed area. I think the next big thing for unmanned aircraft will be the small version. That combines the best of the robotics with the best of the small UAVs, maneuverability, and stability.
“If you think about any particular domain and ask the question, such as lifting something that’s relatively small off the ground and carry it somewhere, the functions of cutting, grasping, releasing, attaching, carrying a tool, and so on are the same problems we had many years ago with space robotics. The ability to do something dexterous with manipulation in space is now appearing with unmanned aircraft.”
What’s the rush?
As engineers know, when it comes to product design you can usually achieve at least two of three factors: you can have it fast, you can have it cheap, or you can have high quality. Most engineers find that their bosses pick one — get to market fast. Thus, engineers must deal with the pressure to rush products to market. Sometimes the rush has few consequences. Other times, the rushed product turns into a news story.
“In the software world, we offer beta versions of software. Many times, that software gets out into the community, who have mostly learned to accept problems with software because we want the latest things. Well, if it’s a beta version of a video game, who cares? Right? It’s exciting for it to fail.
“If it’s software in the automation of a Tesla vehicle, that’s unfortunate and we’ve seen crashes because that software has been rushed into the vehicle before it actually proved itself. It was a beta version. I’m not saying that Tesla is any worse than a lot of new companies. I’m just saying that they are an example of a company that has chosen to put early software systems into passenger vehicles so that they could collect data from it. They have manuals that say, ‘Warning, this is not a well-tested software. Don’t trust the autopilot. Always sit there with vigilance, with your hand over the wheel, be ready.’ But people don’t always listen to those kinds of instructions. They don’t always read those instructions.
“So, it’s a question of ‘Is that okay?’ Should Tesla be able to put such products on the road? We have different government standards — and the reality is, safety doesn’t sell. I’ve been so frustrated with that. Over my entire career, I’ve wanted to do research to try to make things safer. Autonomy, not to make the wiz-bang gadget more efficient or cooler, but to make it safer. And the reality is, customers don’t buy safety, they buy cool stuff. If government sets minimum standards, then the companies and the customers go after those minimum standards. There’s a perception that that level is enough.
“There are two questions here. One is, is there a rush to market problem? In which case, you can’t really push a person to be perfect. We make mistakes. So there needs to be enough people, enough time, and enough tests to find those mistakes. And it doesn’t matter what the company’s policy is, there has to be that time and effort.
“On the flip side, one of the reasons these rush-to-market things happen is because we’re running faster than we can catch up with standards. For example, I’m not going to defend Boeing. The 737 Max clearly had some issues, but we also have put Boeing in the crosshairs. And somehow a lot of other companies are sliding by, without ever actually coming into play for those crosshairs. I’m not saying one company is better than another, but a lot of companies rush to market.”
Dealing with diversity
In the past couple of years, issues of diversity have come to the forefront of company discussions. As a professor, Atkins still frequently sees diversity challenges in the classroom. But she also sees progress.
“We still, today, have problems with female and minority students in a classroom and lab setting. The students who are underrepresented tend to be ignored. As an instructor, you have to work hard to make sure that you advocate for those students, that you stop by and say, ‘I think this student had a good idea. I heard something about it.’ And then you talk to them and everybody is like, ‘Well, the professor’s talking to them, so maybe we should listen.’
“I’ve been happy to see an improvement in the university’s level of activism in student-led groups. There’s a group of students at Michigan that call themselves the Black Graduate Students. They not only support each other, but they are playing a more active role in engaging with faculty, serving on search committees, being in Dean’s committees, and so on.
“I think it’s really good for a group that normally might be unintentionally marginalized to be in a position where they have sufficient authority that it would be difficult to marginalize them. We also have women in an aeronautics and astronautics organization that has been forming over the last few years, not just at Michigan, but many other places. At a conference, I would say 15% to 20% of the technical attendees are women. It’s nice to see a group that is large enough to be confident.
Speaking of confidence, as well as diversity, women engineers still struggle with society’s expectations of them. Atkins related a story on how she dealt with expectations.
When Atkins was younger, she “lost points” by not dressing to society’s expectation of how a young woman should look. So, she conducted experiments.
“The first time I ever played games with that was when I was interviewing out of MIT. So, here’s the game that I played. I actually knew how to wear makeup. I would fix my hair and wear makeup and fancy clothes.
“When I got my master’s degree and I was interviewing for jobs, I had more than eight interviews. I was pretty confident that I was going to get at least a couple of them. For half of them, I showed up with makeup and fancy hair and nice clothing and high heels. The other half, I looked like I look today, hair barely combed, casual, maybe not too casual, but not fancy. I received interviews for every position that I had not ‘dressed’ for, but did not receive any onsite interviews or job offers when I was in makeup. That experience taught me something — that there’s an implicit bias against women that spend time looking nice in hiring and I have never worn makeup a day in my career since then, because it’s easier,” she said.
As Atkins shows, women engineers do not need to conform to some perception of what a women engineer is. They just have to be themselves and focus on the love of engineering.
Filed Under: Women in Engineering