Magnus Egerstedt’s Robotarium at Georgia Tech features a bowl-shaped arena where the robots follow prompts and commands from online visitors. Photo: Christopher Moore/Georgia Tech

For the most part, laboratories are great places to work. While the weeknight and weekend hours may not be enjoyable, and you probably still don’t have a window in your office, the work scientists and researchers complete in laboratories is fundamental to society. 

Of course, there are outliers to the stereotypical vision of a scientific laboratory—there are those that conduct unique work both inside and out. 

One such lab at Georgia Tech celebrated its one year anniversary in June 2018. The Robotarium is a “robotics lab for all.” The 725-square-foot lab houses hundreds of robots that can be accessed remotely by anyone—a concept that has rarely been embraced in the past, leaving many robotics researchers with zero access to their main tool, and no further options. 

Two other examples include the University of Southern California’s two-year-old Stevens Hall for Neuroimaging, which boasts the largest brain data repository in the world. The facility’s director, Arthur Toga, actually assisted in the design and architectural development of the lab. There’s also the Satellite Laboratory at the University of Bristol (UK), which is the first in the region to have a satellite lab and a ground station that allows interaction with NASA’s International Space Station, as well as its data and astronauts. 

The Robotarium at Georgia Tech

The concept for the Robotarium largely came about due to Magnus Egerstedt’s belief and commitment to open research. Egerstedt—the executive director of Georgia Tech’s Institute for Robotics and Intelligent Machines—had already been teaching open, online courses to more than 200,000 students globally when he realized the courses allowed students to gain access to higher education, but not instrumentation. 

“It shouldn’t matter if you are a grad student at MIT or not, you should be able to test your theories regardless of where or who you are,” Egerstedt told Laboratory Equipment. “Fundamentally, innovation benefits from people having access.”

For robotics researchers, testing code and algorithms is especially important, if not critical. 

“It costs millions to maintain, operate and build world-class robotics labs, and the consequence of that is there are only a handful globally that truly matter, but there are certainly more people that have good ideas,” he said. “Swarm robotics has become largely a resource competition as opposed to a good ideas competition. My humble ambition was to solve that.”

Thus, the Robotarium was born—with the help of $2.5 million from the National Science Foundation (NSF) and the Office of Naval Research.

The small lab has motion capture cameras hanging from the ceiling that peer down into the lab’s centerpiece: a white, bowl-shaped arena that looks like a 12 x 14 feet hockey rink where the robots conduct their business. The lab also features a large window and seating for live viewing, as well as a safety net that protects viewers when flying robots are suddenly unleashed. Egerstedt refers to the lab as a rare “theatrical experience,” which is an apt description given the hundreds of people that migrate to the Georgia Tech campus on Fridays for the Robotarium’s weekly open house. 

“It’s become a destination,” Egerstedt said. 

To use the facility, researchers can sign up at, where Egerstedt and other members of his lab “approve” users—which they always do in their quest to be as open as possible. The value of the lab is that they provide swarm robotics capabilities, meaning users can control 50 robots at a time, rather than just one or even a handful. The Robotarium provides example code of what coordinated controllers typically look like. Users can adapt the code for their own purposes, run the algorithm on the provided simulator, and then submit that same code to test on the actual robots. Users will enter a queue rather than be given a scheduled time. Once the robots have completed their mission, an email with data and a video feed is sent from the Robotarium to the user. 

It’s not uncommon for Egerstedt and other lab personnel to be completing paperwork one minute, and the next have 50 robots flying over their heads due to a command from someone in the room, on campus or in a different country. 

“We’re giving people access to physical robots on the Georgia Tech campus, both ground and aerial quadcopters, and this is potentially unsafe,” Egerstedt said. “We don’t want to over-constrain what people can do because then it’s not an effective research instrument. You should be able to test code that’s crappy, because that’s research too. You should be able to test code that fails, but we really needed to build in safety guarantees so that whatever you do, you don’t break our robots—or even worse, have quadcopters attack my graduate students.”

So, in addition to the unique physical design elements The Robotarium required during building, special consideration was also given to the safety of viewers, staff and the robots themselves. 

According to Egerstedt, the lab developed low-level safety routines to allow users to conduct robots’ actions up to the point where catastrophic events are imminent—then the safety override system kicks in. The robots are still allowed to collide—as that is a valuable part of swarm robotic testing—but they cannot collide head-on at high velocities while on the ground, or at all if they are airborne.

Of course, none of this would be possible without Egerstedt’s custom-designed GRITSBots, which are the core of the Robotarium. These inexpensive, miniature differential drive robots are specifically designed to work in tandem with the swarm robotics lab through features that include 1) automated registration with the server and overhead tracking system, 2) automatic battery charging after experiment completion, and 3) wireless reprogramming. 

In just over a year, more than 300 research groups from all continents except Antarctica have used the Robotarium’s facilities. The groups come from different application domains, such as those investigating traffic patterns, precision agriculture, warehousing, and urban search and rescue to biologists and social scientists studying ant and human interaction. 
In that year-plus, there have been a few lessons learned, and thus a few enhancements planned to upgrade the Robotarium even further. 

First on Egerstedt’s list is to upgrade the GRISBots. 

“It turned out we have this beautiful arena floor, but when dust particles accumulate, the robots’ wheels get stuck every now and then, and then that whole experiment needs to be scrapped,” Egerstedt explained. “So we’re updating the robots with ‘beefier’ wheels to make them more effective.”

The team is also working on advanced safety methods that would allow aerial experiments even when lab personnel are not present—something that is not possible now. Lastly, Egerstedt is building an educational version of the online command center, one that will have more example code and will be easier to use and manipulate. 

“You don’t have to be a researcher to use this one, you could just be a clever fifth grader,” he said.

Stevens Hall for Neuroimaging at USC houses the world’s largest brain data repository in the world, which currently stores nearly 3 petabytes of data. Photo: Tyler Ard/USC Stevens Neuroimaging and Informatics Institute

USC’s Stevens Hall for  Neuroimaging

When laboratory architects are commissioned for a project, they usually interface with lab directors a certain amount in order to fully understand the project’s requirements. After all, constructing a laboratory is unlike any other project, given the complexity of scientific instruments, safety standards, data storage needs, etc. 

However, having the director of the institute literally design the new building right alongside architects is a little less common. But that’s exactly what happened at the Stevens Hall for Neuroimaging at the University of Southern California when Director Arthur Toga teamed up with renowned architectural firm SmithGroupJJR. 

Toga, a 40-year research veteran and provost professor, knew what would make his lab run as efficiently as possible—thus, he was the best person to consult at all design stages, even down to the selection of materials. 

“It’s something I have done all my adult life, and I just really enjoy it,” he told Laboratory Equipment. 

That much is evident in the physical results of his work. 

Stevens Hall for Neuroimaging is highlighted by warm wood tones and comfortable, soft materials Toga selected to make it feel less sterile and cold, as laboratories tend to be. The hall also includes as many curved spaces as possible since “there aren’t many straight lines in biology.”

“Building a space that is tailored to what we do and reflect what we do is very important,” Toga said. “It’s important that there’s communication between the place we work and the people that work in it. That’s an element that fosters people being fully engaged, and improves morale, enthusiasm and motivation.”

The result is an eye-catching, unique structure that serves to attract top researchers from all over the world to study the human brain using highly advanced imaging equipment, alongside additional scientific measures. 

The center is not only home to the most powerful Siemens 7T MRI scanner, but it also houses the world’s largest brain data repository in the world, which currently stores nearly 3 petabytes of data but can hold up to seven if necessary. It also has its own onsite high performance computing (HPC) cluster with 38 terabytes of memory—something that is essential to the everyday work of internal researchers, as well as for the extensive collaboration that takes place outside the university. 

When you have the largest brain data repository in the world and are obsessively studying the human brain, you’re bound to cross paths with Alzheimer’s disease, which explains why the Stevens Neuroimaging and Informatics Institute has about a dozen research programs dedicated to the debilitating disease. 

Some researchers at the institute are using imaging to try to understand whether exercise can influence the course of Alzheimer’s, while others are looking at dysregulation and the blood-brain barrier that’s associated with disease. There’s a large program examining how high-field imaging at 7T can improve scientists’ ability to look at the pathophysiology that can be seen using magnetic resonance, and another that studies the relationship between the accumulation of beta amyloid and tau, relative to imagery collected using magnetic resonance. Research efforts have also looked at gender differences in the disease, genetic risk factors and morphological characteristics, among others. 

Having access to a very large repository and massive amounts of data allows the institute’s scientists to leverage statistical power to understand the similarities and differences of healthy and diseased brains. 

“It permits us to ask very detailed questions that if you didn’t have a large enough sample, you wouldn’t know whether [something] was normal variability or whether cohorts were different because their genetic profile made them different or because they all had a particular type of disease, or whether some other variable distinguished them. Having vast numbers gives us enormous power for investigation,” Toga said. 

Part of housing all of this data is knowing how to utilize it. For that, the institute has the Data Immersive Visualization Environment (DIVE) presentation theater, which allows researchers to project massive data sets and highly magnified images on a 12 x 15 feet screen with 1.5 mm pixel display in ultrahigh definition 4K resolution. 

“The idea of this [room] was to use the power of these computers to create models that allow us to literally fly into these spaces and examine the complicated spatial relationships between the different measurements, i.e. what’s happening in this part of the brain versus how that part of the brain is organized,” Toga explained. “The theater really allows you to feel like you’re part of the data so that you can ask questions of it while you’re deeply being presented from all sides.”

While the neuroimaging and informatics institute is obviously on the cutting-edge of research, it’s also moving in the right direction when it comes to the clinic—that is to say, it’s moving into the clinical sphere, where it has the chance to be transformative. Already, the basic science that has emerged from neuroimaging has been hugely successful in medicine, helping clinicians make more informed decisions about the treatment of their patients. 

“Having computers help us discriminate the different populations and the response to various interventions is a hugely powerful tool, especially with something as complicated as neurological disease or psychiatric disorders where we have vast amount of information, including imagery, that can help clinicians make decisions, record their responses and determine whether any modifications are necessary,” Toga said. 

Eventually, this power could be leveraged to identify disease in certain populations earlier than is possible now, enabling the development of FDA-approved therapeutics that could either slow or eradicate the disease. 

Tim Gregory, Ph.D. student at the University of Bristol, and Lucy Berthoud, head of the satellite lab, inspire young students from Bridgewater College Academy. Photo: University of Bristol

Satellite Laboratory at the University of Bristol

Students and faculty at the University of Bristol wanted to get more hands-on with their space research, advancing the research discipline to new heights within the British university. They wanted to build their own space mission and satellite, so they asked for the resources necessary to do so, including a new laboratory and ground station. 

The university agreed, and in November 2017, a state-of-the-art satellite laboratory opened on the Bristol campus. 

One of the reasons the university so readily agreed to a new laboratory is the sudden democratization of space research, thanks to CubeSats, or miniature 10 x 10 cm open-access satellites that can be put into orbit by deployers on the International Space Station. As of April 2018, more than 800 CubeSats have been launched. 

The CubeSat’s miniature size allows it to keep a low cost of deployment, and its standard design specifically minimizes risk to the rest of the launch vehicle and payloads. Since most CubeSats are 10 x 10 cm, they can all be launched and deployed using a common deployment system called a Poly-PicoSatellite Orbital Deployer (P-POD), developed and built by California Polytechnic State University. 

Anytime a CubeSat is built, it must be open-access, meaning anyone with an antennae can download the data. Thus, “mission control” inside the University of Bristol’s satellite lab can control their own ultrahigh frequency and very high frequency antennaes, in addition to tracking and downloading data from other satellites. 

“As a starter lab, this is very exciting since we don’t have our own data yet,” Lucy Berthoud, director of the lab, told Laboratory Equipment. “We can start off by downloading other peoples’ data, including the ISS. We can send a small package of data up to the ISS, they timestamp it, then send it back down.”

The ground station on the roof of the building is where the antennas are stationed, and inside with mission control there is a cleanroom to ensure proper satellite construction.

Another important—and unique—component of satellite construction: the lab’s “air bearing” system. The system features a ball that is supported by air shooting upward. In turn, the ball supports a CubeSat, enabling researchers to test specific controls on their mini satellite, such as the altitude control function, or the wheels inside the satellite that control its direction. 

The lab’s first satellite mission, funded by the UK Space Agency, is a volcano monitoring satellite designed in collaboration with RAL Space. The project is anticipated to take several years to complete; but once designed and built, the satellite will observe volcanoes from space and take 3-D images of ash clouds, which can be toxic to humans and cause excessive damage to aircraft.  

“We’re hoping this is not just a one-off satellite, but a program where we’ll also have satellite 2, satellite 3 and so one,” said Berthoud. “Our aim is to push as much of the work as possible toward the students. Our staff is there to give research value, but the students are managing and building the satellite themselves.”

In addition to participating in live satellite missions, the university’s lab is also used for outreach to young students in the Bristol area. Berthoud says the lab has already had school-aged students visit, with their university counterparts around to help them understand the importance of space missions. 

“Space can help us inspire young students to choose science and engineering over other careers,” Berthoud said. “We need our young scientists and engineers, we’re going to need lots of them in the future. In a very practical way, they will contribute to the economy—and space research does contribute to the world economy. In a more idealistic way, space can help us solve global challenges, like medical, disaster management, sustainability, climate change, etc., so I feel then we are doing our part to help solve these challenges.”