A drone launch usually draws eyes to the skies, but Tom Zambrano looks to the sea spray off of Catalina Island instead. Zambrano, an environmental scientist at Apium Swarm Robotics in Glendale, Calif., serves as technical director of Apium’s SeaSwarm project, which is leveraging autonomous vehicle technology in watercraft, the smaller of which resemble Minions boring through the waves.
But these capsules aren’t designed for entertainment. Apium’s aquatic drones are engineered to collect data for oceanographers and, soon, to clean up oil spills, deployed from shore to execute their mission with a fish’s intuition for others’ position in the water.
In a phone conversation with Crain's Los Angeles, Zambrano shared how he joined Apium founder Tyler MacCready in translating drone technology and swarm instincts to seafaring vessels. The story begins at AeroVironment, another California company that MacCready’s father founded in the 1970s and expanded with Zambrano’s help.
Q: What are the roots of Apium?
A: I worked with Tyler’s father, Paul MacCready, who started AeroVironment just after the Clean Water Act and Clean Air Act were passed. Paul recognized a need for a new branch of aerodynamics to deal with the built environment — how does smog move and how do buildings interfere with air movement. Paul was always an aviator and thought it would be great to put his love of airplanes to use for the environment. Lo and behold here I was with a degree in environmental aerodynamics.
I’ve been here for 35 years now, working on myriad things, including making specialty aircraft and drones.
Tyler joined the company with his father. He said, “I wonder if we can look into the ocean the way we looked into the air a generation ago.” So we started an ocean environmental division, which became sort of an unsanctioned spin off. We called it Ocean Lab initially.
Q: How did you come up with the idea for these aquatic drones?
A: Just as AeroVironment evolved into making drones for air, we thought maybe we can make specialty drones for measuring things in the ocean. We started working on these small Data Divers, 2- or 3-foot-long yellow water drones. They skim the surface and can dive into the water and come back up. We started using them for scientific measurement of things right off the coast. The reason we did that is, we are a small company. Large companies make equipment that does deep ocean stuff, but their drones are a million dollars apiece. So we focus from the beach to a mile or so offshore.
One of Tyler’s brothers, Parker MacCready, is a professor of oceanography at the University of Washington, and he gave us some good use cases about what to measure, such as tidal flows entering an estuary, salinity, pH and temperature of water. From that, scientists have a lot of ecological information they can use in developing their models. If we have a large storm, for example, they want to know how much organic material has entered the ocean, and they want this data as soon as possible, so they can forecast whether communities are going to have to close a beach and things like that.
Because of the connection with the University of Washington, one of the premier oceanographic schools, we got a little street cred, so to speak. We were dealing with bona fide oceanographers and addressing a measurement problem that is very useful for them. It’s a bit nerdy, but that’s what we’re doing.
Q: Where does the oil spill application enter the picture?
A: We started talking with the directors of the Senseable City Lab at MIT, Carlo Ratti and Asoff Biderman. They showed us something they worked on after the BP oil spill, a swarm of ocean drones that could clean up oil. They called it Seaswarm.
We are licensing some technology from the scientists who came up with the big picture, but we aren’t using the same vehicle design.
Q: How do the drones work for the oil spill application?
A: Imagine if you have liquid, and put oil on top of it, it’s pretty easy to get it off, but once that oil hits the beach, then it’s a whole different story; it gets really hard to clean it off of sand and rocks. So really all the effort in science is on how can you stop the oil before it hits the shoreline.
We saw that right here in southern California with an oil spill in the ocean a couple of years ago. Santa Barbara is probably the most environmentally advanced community in the U.S., with a great university that has some of the best oceanographers in the world. Swarms of people were showing up on the beach doing anything they could, but it still takes hours and hours for the work boats to come in. Once the oil started advancing onto the beach, it was a problem.
We’d like to have, at the first threat of oil approaching a beach or community, a 20-pound shipping container show up, and you open it up, and one or two emergency professionals or people could carry these things, each about the size of a kayak, and literally throw them in the water, and they’ll take it from there. Once the vehicles are out there, they would be able to skim the oil and ingest it. It would be great to have these as a first responder type of thing before the big guys come.
We make no illusions to being the only ones who would deal with this sort of thing. We’re not the Tesla of the ocean or anything. But we hope to have these on the market in 2018.
Q: What role does swarming play in these?
A: Swarming is how a simple fish or bee that possess very limited innate resources can reach destinations it never could on its own. While [the] movement of individuals within a swarm may look to us very chaotic, they never collide. Also, viewed as a whole, the collective motion of the swarm is incredibly fluid and seemingly unlimited in size and number.
While the science of swarming has been widely studied, it has been difficult to apply to robotic vehicles, mainly because our robotic vehicles most often are constrained to work in a structured environment. There are no roads or traffic laws that contain how and where a bird or bee moves through the sky or a fish through the ocean.
Our SeaSwarm vehicles will be operating in something very close to an unstructured environment. While they are programmed to stay within the perimeter of a designated region, there are no other preset rules limiting their motion within it. They are programmed to just follow a few simple rules with respect to their neighboring vehicles. This enables our vehicles to be very simple, not needing incredibly advanced computers and sensors, and their numbers can be in the tens, hundreds and potentially thousands if that is what is needed to surround an oil spill.
Operating in this unstructured environment is what will allow our SeaSwarm to move very efficiently over an oil spill and skim it up before it can drift onto the shore.
I just came from the Dodger Stadium parking lot yesterday; that was probably the biggest swarm locked in a structured environment I’ve ever seen. If we were free to swarm like bees, I would have been able to get home in 45 minutes as opposed to 2 hours.
Q: Why did the company change from Ocean Lab to Apium?
A: We had our first round of funding, private money, that we’ve been using the last few years. Our first round is when we changed our company name. Apium was one of those great names where you hire guys from Silicon Valley — well, it was Silicon Beach — to come up with five-letter words. Apium means “liked by bees.” We thought, we’ll go with Apium. We were working with a small company called Liftoff, and they would like us to become the next Tesla.
When we began talking to Liftoff, they said, “We know you like the ocean, and boating is a lot of fun, but does your technology relate to aerial drones and self-driving cars?” — which are the two things everyone is talking about right now. They asked, “The swarm autopilot, could this be used in cars?” We said, “Well, yes, it can, in certain applications.” They said, “Can it be used in aerial drones?” and certainly, it can be. So we realized if we are to be selling this technology, we should come up with a name that isn’t just ocean-focused.
We kind of had a little lesson like that in my hundred-year career at AeroVironment. When we started getting in to the serious business of drones, all of the new, young people called the company “AV,” whereas the old people called it AeroVironment. The young people said, “No one knows how to spell that. We need to make it more useful for the clientele and easier to say.”