Signal Kinetics: Year One

MIT Media Lab/Edmond Awad

It’s been a year since I launched my research group, called Signal Kinetics, at the MIT Media Lab. In that time, we’ve developed new wireless technologies for drones, virtual reality, and smart environments. We’ve tested prototypes of our technologies, often flying our drones inside the Lab. And we’ve written and submitted research papers about our work.

When our first paper was accepted, we flew to Los Angeles to present it at SIGCOMM, the premier international research conference in computer networking. After that event, I wanted to treat my researchers so I took them to nearby Universal Studios and the Wizarding World of Harry Potter theme park. When we returned to Boston, we saw our research featured on the front page of MIT’s website, and we fielded numerous calls from the scientific and popular press wanting to report on our work.

Now, as I take stock of our first year as a group, one theme stands out: It has been a rollercoaster like the one we rode in the Harry Potter theme park—exhilarating, exciting, and magical.

Building a Team

Flashback to May, 2016: I had just accepted a faculty position at the MIT Media Lab. My main goal was to establish a research group that would explore new ways to sense the world, beyond what people have traditionally seen as possible. I wanted the word “impossible” to inspire my group’s mission and be a focal point for my team members—to bring us together and to make our minds and hearts race.

This yearning for “impossible” technologies had emerged during my graduate studies at MIT. My PhD research showed how we could use wireless signals to deliver new capabilities— ranging from seeing through walls to capturing people’s breathing, heartbeats, and even their emotions. You can see what I mean in this TEDx talk, which I gave as a doctoral student.

I knew that my first step as a professor would be to build a team of those who’d share my vision. I believe that when brilliant minds work hard together, we can make magic, and there were two crucial qualities I was looking for in candidates: intelligence and dedication. I was also interested in hiring postdoctoral researchers who were not in my exact research area but in an adjacent field so that, together, we could work on innovative ideas at the intersection of our areas of expertise. By the end of the summer, I’d hired a postdoc and a graduate student.

Naming the Group

My team was an integral part of launching my lab, and I wanted all members to be involved in giving it a name so they could feel a sense of ownership and belonging. And, I felt that the very name of my group should reflect both its research mission and dynamics—to have a personality of its own that would represent what we do and who we are.

I made long lists of potential group names, discussed them with my team, asked my friends and colleagues for feedback, shortened the list, and iterated. Eventually, I chose “Signal Kinetics.” This encapsulates my vision to invent new ways of sensing the world by relying on invisible signals, which are in constant motion (that is, kinetics) in space and time. Motion is also a part of how my team operates. Anyone who comes to our lab space will see a buzzing group of excited scientists tinkering with hardware, designing new devices, manipulating them with new algorithms, and fantasizing about ways in which we could make positive change in the world.

Bootstrapping a Research Agenda

Once I formed the founding team, my next goal was to identify a research problem that we could work on as a group. It had to be exciting and intellectually stimulating while also holding potential for real-world impact. In the first few weeks, I met every day with my team to brainstorm different ideas and directions. I cast a wide net for ideas that spanned topics ranging from brain-machine interfaces to cybersecurity, and I asked my team to prepare presentations for discussions on specific directions in which I believed we could contribute.

Within a couple of weeks, I focused the team on an exciting question: Could we use drones to find objects that are otherwise “invisible” and hard to track? It had all the features of a great research problem—it was exciting, technically difficult, and intellectually stimulating. As well, it promised real-world impact because finding misplaced items and keeping track of inventory is a multi-billion dollar challenge facing many industries, ranging from pharmaceuticals to the U.S. Army. Walmart, for instance, reported that in 2013 it lost $3 billion in revenue due to mismatches between its inventory records and its stock. In addition, between 2003 and 2011, the U.S. Army lost track of $5.8 billion of supplies among its warehouses.

Most importantly, we had a plan for how to solve such problems! In essence, it was to leverage battery-free wireless stickers (just five cents each) and place them on any object we wished to locate. The flying drones could power up these battery-free stickers from a distance and use their wireless signals to find their locations accurately. In the video below, you can see that this idea required us to build new circuits that we could mount on commercially available drones, and to develop new algorithms that would enable us to perform accurate localization through the drones. 

I divided this project, called RFly, into different parts so that my team could work in tandem. Every day we met and shared our progress. This was a great opportunity for me to give my researchers timely and frequent feedback and make sure we were all on track towards our deadline.

Collaboration was key, especially during my group’s first couple of months. While our permanent lab space was getting set up, we only had one cramped office in which we all sat and worked together. But by mid-December, our lab was ready and we moved into a new space with high ceilings. We could finally fly our drones there and start testing our ideas! It was all very exciting.

Overcoming Hurdles

But there were some hurdles, and the first one arose too soon. The airborne circuit we had sent for fabrication was delayed. Upon contacting the fabrication company, we found out that they’d made an error and hadn’t started fabricating our circuit. This meant that we would not be able to submit our paper on time.

So, I pivoted my team to work on a complementary research question: Could we transform the battery-free wireless stickers into powerful radars without any hardware modification? Answering this question would allow us to obviate the need for the delayed circuit. While it would not solve the same problems as our drone-based design at the scale of warehouses, it could still have enormous implications for smart environments, robotics, and virtual reality.

During the following stressful week, my brilliant team was able to make significant progress on this new idea. At the same time, I contacted the fabrication company and requested expediting the assembly to make up for their delay. The director promised to honor my request, but I did not share his response with my researchers as I did not want to raise their hopes prematurely. Fortunately, the fabricated circuits arrived the week after, and we were all very happy to resume working on the drone-based project. We were two weeks behind schedule, but I promised my team that we could still make the deadline if we pushed hard enough. And we did.

More importantly, the progress that we’d made as we were waiting for the delayed circuit meant that by the time we had submitted our first paper, RFly, our stickers-to-radar project, called RFind, had already taken off. In the subsequent weeks, we built on our initial progress and designed algorithms that exploit the physics of battery-free stickers to enable precise positioning.

The video below shows how RFind can solve common problems in our everyday lives—like finding misplaced keys, wallets, or glasses. We also demonstrated how the same algorithms can enable various applications in virtual reality, robotics, and smart environments.

Move Fast, Aim High

By April, we had already submitted two papers to the top venues in our field of research, and I had also written and submitted my first grant as a faculty member to the National Science Foundation. We submitted our first paper on our drone-based technology, RFly, to SIGCOMM; and we submitted our second paper on our stickers-to-radar technology, RFind, to MobiCom, the premier international conference in mobile computing.

While I was enthusiastic and confident about the merit and potential impact of all our submissions, I knew that the review process could be random, so I was not sure all our submissions would be accepted. Nonetheless, I wanted my entire group to go through those experiences. I believed they should know that, even if we do not succeed on the first try, we would learn from the process, recover quickly, and take aim again from more steady grounds by making our technologies even more powerful. As it turned out, both our papers were accepted.

Around that time, we were also nearing an opportunity to demonstrate our research—at the Media Lab’s spring event where representatives from more than 80 member companies visit the Lab and find out about our latest projects. These member events enable collaboration between academia and industry. We demo our cutting-edge technologies, and the member companies share with us their toughest obstacles. This is central to our research mission: we love inventing practical solutions for impossible problems.

So we were excited to demo our research updates to the Lab’s member companies.  We worked day and night to polish the demonstration of our technological developments. It was a huge success, attended by almost every member company! It also provided a great opportunity for us to identify impactful avenues for collaboration with industry.

Another wonderful aspect of the Media Lab member meeting is that we get to learn about work going on in every other research group in the Lab. And they also learn about our work. This is a fantastic way to bootstrap collaborations across research groups and explore ideas at the intersection.

Transcending discipline boundaries

The Lab houses over two dozen research groups, which span everything from neuroscience and biomechatronics to human-computer interaction and computer vision. Why is this so exciting? Because if we identify a research problem at the intersection of any two groups, we may end up stumbling upon an entirely new scientific field.

To some extent, I’d experienced this before joining the Media Lab as faculty. I was a PhD student in the Computer Science and Artificial Intelligence Lab (CSAIL) at MIT, where I thoroughly enjoyed my collaborations with faculty across different subfields of computer science, including theory, computer vision, robotics, and human-computer interaction. One of the most beautiful features of academia is that it is not a zero-sum game: When scientists collaborate, they all win, and so do science and humanity.

I had this philosophy in mind when, during the final year of my PhD, I interviewed and received offers for faculty positions from over a dozen of the top 20 schools around the country. Each school was unique, and my choice of where to go was very personal. My love of exploring research that transcends discipline boundaries drew me to the Media Lab, where I could tap into initiatives ranging from outer space to extreme bionics.

Our First Anniversary

It’s thrilling to realize how much I learned and how much I was able to teach my researchers during my first year of being a professor. I feel fortunate that both papers my group submitted were accepted, and that my National Science Foundation grant proposal was successful. As well, our research was featured on the cover of the MIT homepage and was picked up by major news outlets around the world including the BBC, The Verge, Engadget, Sina, International Business Times, and IEEE Spectrum.

The team of Signal Kinetics researchers I lead has expanded from two to five in less than a year. What started as a small, energetic group is now a growing cadre that keeps buzzing. Nothing brings me more joy than intellectually stimulating discussions with my students—conversations that range from the physics underlying an equation to how our ideas can change the world.

I am profoundly excited about the scientific journey ahead of us. I have no doubt that we will encounter many obstacles along the way. We will learn the numerous ways in which our ideas will not work. We will fail fast, and we will fail often. Along the way, we will learn and I hope our discoveries will teach the world how to make impossible things not only possible, but achievable. I am confident that our journey will be magical as we invent ways to sense the entire world by moving invisible signals.

Fadel Adib directs the Signal Kinetics research group at the MIT Media Lab. 

Related Content