Array Labs scans the Earth from space to provide self-driving vehicles with 3D maps
Image credits: Matrix laboratories
It’s an oft-told story: Today’s space startup boom can be traced back to the dramatic decline in the cost of launching and manufacturing satellites over the past 10 years. But Array Labs, a two-year-old startup based in Silicon Valley, is also taking advantage of other technological advances in its quest to build a 3D map of the Earth.
This includes computational gains, such as in advanced graphics processors (GPUs), and radar software development, explained Array CEO Andrew Peterson. The revolution in scientific computing has opened up new possibilities, said Peterson, an aerospace engineer who previously worked at General Atomics Aeronautical Systems and Moog’s aerospace and defense division.
“If we could take all this superpower that we’ve seen in radar and scientific computing, and we could couple that to really low-cost satellites… there might be a really interesting way to do a new kind of Earth observation,” he said. “That would have been better.” “An idea that never occurred to me.”
Array plans to map the Earth in 3D by flying constellations of radar satellites in low Earth orbit to image the same place at the same time. By taking photos of the same place from different viewpoints, the company hopes to capture a full, high-resolution 3D digital record of the world. It works similarly to synthetic aperture radar, but the precise technique is known as “multi-static radar,” so named because of how multiple radar satellites work cooperatively across spatial distance.
According to the company, the 3D data will be so high-resolution that it can be used by autonomous vehicle fleets, augmented reality headsets, insurance analytics, and in national security applications.
This is not the first time a group has tried to use space-based radar to create 3D global images. In 2003, the Air Force Research Laboratory made plans to launch a program called TechSat 21 that would do just that, Peterson said. But it was hampered by account restrictions. For example, the TechSat 21 team had to solve a bandwidth problem: How do you get enough bandwidth to store all the data the radar will generate?
“The system they came up with was ten spinning hard drives that were all sorted together,” he said. “It used to weigh maybe 20 pounds, and it used 150 watts (of power). Now, something the size of my thumbnail has 100 times better performance and 100 times less cost.
Peterson launched the startup in 2021. Array’s first angel investor was Brian McLendon, an engineer who oversaw the development of Google Earth. To expand further, Array quickly joined two accelerators: Seraphim Space Camp in fall 2021 and Y Combinator the following year. It closed a $5 million seed funding round last October, led by Seraphim Space and Agya Ventures, with participation from Republic Capital, Liquid 2 Ventures, Rebel Fund, and Y Combinator co-founder Trevor Blackwell.
The plan is to launch constellations of satellites, each consisting of about dozens of satellites. The ultimate goal is to get 10 to 20 groups up and running. Even one group would be able to take over 5% of the world’s surface, representing 95% of the world’s population, every two weeks, Peterson said. Array will launch its first test satellite on a SpaceX ride-sharing mission next year, the first in a series of test satellites designed to iteratively de-risk various aspects of the company’s technology.
Before it gets there, Array does as much testing as possible on the ground. The 10-person team set up an internal radio frequency (RF) test range to demonstrate the image generation algorithms and software before the first launch. The company plans to publish a blog post on the system, detailing the process of creating a SAR test render in CAD, to deliver the platform in a Silicon Valley parking lot, to create the first clean radar image.
“Building a little bit in public is really important,” Peterson said. “We’re doing something really new, and we’d like to talk to as many of these companies as possible, and as many people in these companies as possible.”