Engineers at the University of California San Diego's Jacobs School of Engineering and Qualcomm Institute have developed a more efficient, scalable way to handle satellite communication traffic, one that could significantly increase throughput by rethinking the ground station to work with satellites already in orbit.
"The fundamental bottleneck in scaling satellite connectivity today is not in space, it is on the ground," said Dinesh Bharadia, associate professor in the Department of Electrical and Computer Engineering at UC San Diego, an affiliate of the UC San Diego Qualcomm Institute and senior author of the paper. "Every bit of data a satellite transmits still needs to travel through a ground station to reach the end user, and that infrastructure has simply not kept pace with the growing number of satellites."
Instead of relying on the large parabolic dishes used in many ground stations today, the researchers propose using multiple small, flat antenna panels that can be spread across rooftops and other sites and be coordinated to work together. The approach could offer a lower-cost, easier-to-deploy alternative to traditional ground-station hardware.
"This work enables the industry to scale ground stations rapidly and cost-effectively, even through crowdsourced deployment," said Bharadia. "Any rooftop owner or enterprise can install our solution and carry satellite data back to the internet."
"Each panel on its own isn't powerful enough to maintain a high-speed link to a satellite," said UC San Diego Electrical and Computer Engineering Ph.D. student Rohith Reddy Vennam, who was first author of the paper. "But combining enough of these small, laptop-size arrays will match and exceed the data rates that a large dish can deliver. This system is also extremely cost-efficient, built from off-the-shelf panels, and easy to deploy on rooftops or other accessible locations."
The research will be presented at IEEE International Conference on Computer Communications in Tokyo, Japan, May 18-22.
Beyond the dish
Satellite communication has become essential for telecommunications, navigation, secure financial transactions, internet connectivity from remote locations and numerous other features of modern life, including connecting first responders during emergencies and enabling remote healthcare.
"Satellites are exploding in number," said Vennam. "SpaceX alone has thousands of satellites in orbit, and OneWeb and Amazon will soon be launching their own constellations. With reduced manufacturing and launch costs, many companies are now deploying satellites for different applications: broadband, satellite-to-mobile connectivity, Internet of Things, Earth imaging and more. However, ground station infrastructure is not scaling at the same pace. That will be the next bottleneck."
Vennam (left) and Associate Professor Dinesh Bharadia demonstrate ArrayLink, with old satellite dish technology behind them. (Photo by Areli Alvarez, UC San Diego Qualcomm Institute)
Current ground stations often rely on large parabolic dishes. Those dishes are powerful, but they can track only one satellite at a time and must physically move to follow fast-moving low Earth orbit satellites. This mechanical steering is slow, and valuable connection time is lost every time the dish swings from one satellite to the next. That makes them slower and less flexible than the growing satellite landscape demands. In addition, acquiring and maintaining a property to house these ground stations can be expensive.
The UC San Diego researchers thought they could come up with a better way.
Spreading out
The team explored the approach of using phased arrays, which are flat antennas made of many small elements that can steer radio beams electronically, without moving parts.
But building one giant phased array powerful enough for ground-station use would be expensive and complex. So, instead, the team used many smaller panels and made them work together.
The researchers recognized a practical limit: beyond a certain number of panels, the added gain starts to flatten out, so piling on more hardware provides much less benefit for much more cost and complexity. Their solution was to accept a small tradeoff, settling on 16 panels instead of the 45 theoretically needed to match a large dish, while still reaching signal strength close to dish-class performance. This made the system far more realistic to deploy.
At that point, the design was practical, but the researchers still needed a way to boost overall capacity. Then came the breakthrough.
By accepting a small reduction in raw signal strength, the team unlocked something far more valuable: the ability to run multiple parallel data lanes between satellite and ground simultaneously, moving significantly more total data than a single high-powered link ever could.
The technique worked because of a fundamental property of how radio waves work.
"In a line-of-sight link to a satellite, if your receiving panels are too close together, they all see essentially the same signal and you cannot separate multiple data streams," Vennam explained. "But as you spread the panels farther apart, each one starts to see the signal slightly differently. And if you spread them far enough apart, about a kilometer in this setup, the separation creates what physicists call a near-field effect, where each panel sees a distinct enough version of the signal to receive multiple data streams simultaneously at satellite-scale distances.
"What makes this surprising is that near-field physics is typically associated with short-range systems," he continued. "This work shows it can be deliberately engineered even for satellite links hundreds of kilometers away, by controlling how far apart you place your panels on the ground."
Using that insight, the team found that the system could support up to four simultaneous information streams under the right conditions. This translates to up to three times the total data throughput compared to a traditional dish.
"One key point is this number of possible streams is not random," Vennam continued. "By positioning your arrays on the ground and on the satellite and choosing the operating frequency, you can precisely control where multiple streams are possible and how many you can enable."
"Rather than dedicated, large, bulky equipment at a limited number of locations, with our method arrays could be placed on existing 5G cell towers to receive the satellite data," says Ish Kumar Jain, a UC San Diego alumnus who is now a faculty member at Rensselaer Polytechnic Institute.
The researchers then tested their system, which they called ArrayLink (patent pending), on the ground, obtaining experimental results that neatly matched the theoretical predictions.
What comes next
Vennam and co-author Ish Kumar Jain, a UC San Diego Jacobs School of Engineering Ph.D. alumnus who is now a faculty member at Rensselaer Polytechnic Institute, are excited to present their paper at IEEE INFOCOM.
At previous conferences, they have interacted with members of the satellite industry who are clear about their priorities: any new ground station technology must be cost-efficient and easy to scale.
"One cool thing is that ArrayLink can be readily implemented with commercially available arrays," said Vennam. "You don't need any custom hardware."
Jain noted that ArrayLink is also compatible with existing 5G systems.
"Rather than dedicated, large, bulky equipment at a limited number of locations, with our method arrays could be placed on existing 5G cell towers to receive the satellite data," he said. "5G base stations have everything available for a satellite ground station in terms of ground connectivity, and the location is already leased out. That solves a lot of problems."
The researchers are continuing to refine ArrayLink, with satellite testing on the horizon.
In addition to Vennam, Jain and Bharadia, authors of the paper, " Satellites are closer than you think: A near field MIMO approach for ground stations ," included Luke Wilson, a master's student in electrical and computer engineering at UC San Diego.
