This article originally appeared in the May 14, 2018 issue of SpaceNews magazine.
Clusters of satellites that scan the globe at previously unimaginable speeds will make it possible to detect subtle changes on the ground so that earthquakes, floods and avalanches can be more accurately predicted. And they will give governments and corporations access to unprecedented intelligence such as the precise identity of ships at sea that may be illegally fishing or trafficking arms.
“You can see changes at a centimeter level, from space. It’s amazing,” said Robert Laudati, managing director of commercial products at Harris Space and Intelligence Systems. Laudati was speaking about space-based synthetic aperture radar, or SAR — one of several forms of data that companies are collecting and combining with other types of information, sparking a boom in geospatial intelligence services.
Harris works with Capella Space, a Silicon Valley startup whose small satellites with SAR sensors cover the planet once every six hours. “SAR has been a very important data source in Europe. It’s one of the few sensors that can penetrate clouds and operate in day and night.”
Whereas the space SAR industry is now starting to pick up speed in the United States, the European Space Agency has been giving out high-resolution SAR data for free, Laudati said in an interview.
What is making SAR more desirable is the magic that happens when the data is fused with optical images, radiofrequency and hyperspectral data. A demand for maritime surveillance is fueling investments in SAR constellations. Radar can identify ships that turn off their tracking Automatic Identification System, (AIS) beacons intentionally. “SAR has turned out to be very good at detecting ship outlines regardless if the AIS is turned on,” Laudati said. “You overlay data sets and you see a ship.”
Julie Baker, co-founder and vice president of Ursa Space Systems, a startup that analyzes radar satellite data, said that by 2021 there will be over 100 SAR satellites on orbit. “More frequent revisits mean more timely response for first responders,” she said at the GEOINT 2018 symposium last month. Weather is not an obstacle to SAR, she said. “Tornadoes come through at dusk after optical satellites have gone over, and it’s too windy for aerial systems. Radar can cut through clouds and work at night.”
Radar images look like a 1950s TV show in black and white, said Walter Scott, chief technology officer at Maxar Technologies. The company’s MDA business unit operates the Radarsat 2 satellite.
“It’s actually easier for machines to pull data out of radar imagery than to pull it out of electro-optical,” Scott told SpaceNews. “Radar brings its own flash bulb, its own light source. It’s more predictable than optical imagery.” For most applications, it’s less about a picture but using the radar to detect things.
Using machine learning algorithms, MDA combines AIS and radar data to draw information about ships, Scott said. “You can get a tip out of the radar and feed the software in the electro-optical sensor to go take a picture.”
Machine learning lets you deal with the “show me where” problems, said Scott.
Maxar spokesman Turner Brinton said the company is building a three-satellite Radarsat constellation mission for the Canadian Space Agency scheduled to launch in late 2018. The agency is interested in repeat imaging of strategic areas such as coastal zones and Arctic waterways at different times of the day. The new constellation will have AIS technology, which when combined with radar images, can be used for immediate detection and identification of ships at sea.
The Canadian Department of National Defense owns the rights to the Radarsat constellation mission and, for now, will be its sole user, Maxar CEO Howard Lance told analysts May 9. The constellation might be expanded in the future beyond the three satellites to allow Maxar to sell data commercially. “We continue in discussions with the government of Canada on what they call the RCM Radar Continuity program. This would be additional satellites to provide additional capacity, some of which could be available for commercial use by Maxar.”
SAR firms increasingly are teaming up with new commercial players in the geospatial industry that do radiofrequency sensing and analytics.
“We are partners with a number of SAR companies,” said John Serafini, CEO of HawkEye360, a three-year-old startup that is about to launch its first cluster of RF-tracking satellites.
“The wave of the future is not to commercialize pixels. It’s to commercialize answers to problems. You need multiple modalities of information: SAR, RF, electro-optical imagery, open source information.”
HawkEye360 plans to introduce a new data-analytic product this summer. “It’s maritime domain awareness for tracking dark ships,” said Serafini. “We are very complementary to SAR and optical systems. We can detect interesting behavior of vessels and other actors, and provide that information as a tip to SAR or [electro-optical] systems.”
The company’s satellite payload is a software-defined radio that detects, geolocates and tracks RF signals. Satellites will fly in clusters of three. The first is scheduled to go into orbit in October. Data from on-orbit RF sensors will be fused with terrestrial RF data provided by Kratos Defense & Security Solutions.
“We are building a new industry similar to how DigitalGlobe created an industry for Earth observation,” said Serafini.
Kratos is also marketing its RF-monitoring data for space situational awareness and traffic management. “Traditionally that has been done by optical sensors and with ground based radars,” said John Monahan, senior vice president of Kratos Defense. “Optical sensors are great at looking at higher GEO orbits, but they struggle seeing through clouds and can’t see anything during the day.”
Radar, in contrast, can see in all weather but only to low Earth orbit. “There’s a gap,” he said. “And you can’t have gaps in traffic management.”
Greg Caicedo, vice president of Kratos Defense, said the business today is about data analytics and partnerships with as many data providers as possible. HawkEye360 is a case in point. “They’re flying in low Earth orbit; they are not going to persist over a larger area like we would,” he said. “Once they’re fully populated they can fly over an area once every 40 minutes.”