Skip to main content

Keeping satellites on track


by Nicole Gaynor

Researchers use Blue Waters to make satellites more efficient and effective.

We rely on global satellite service for communications, navigation, and environmental monitoring, but have you ever thought about how satellites stay in orbit?

Over time outside forces push and pull a satellite away from its original path. Scientists try to put the satellites in places that will minimize the need for course corrections, but lack of computing power has limited the accuracy of the algorithms that guide their decisions. Blue Waters changes that.

The problem to be solved

“In reality, you launch your satellite system and from day one it starts to degrade from the idealized assumptions that you made because there are these perturbations from what you assumed,” says Patrick Reed, a professor of civil and environmental engineering at Cornell University and principal investigator on a project funded by the National Science Foundation that is using Blue Waters.

The “perturbations” are inaccuracies in the model that result from ignoring things like the non-spherical shape of the Earth and the gravitational pull of the sun and moon. In the simulation, the Earth is a perfect sphere. In reality Earth is an oblate spheroid, a ball that is a bit squished from pole to pole, which changes the gravitational pull on the satellites. Because gravity in the model is a little different from reality, the simulated movement of the satellite will also be a little off.

Reed says these differences are small day to day, but they accumulate over time and can reduce the useful lifetime of a constellation (a group of satellites that work together). In order to correct the satellite’s course and avoid compromising the global coverage of a constellation, the satellite has to burn some of its limited fuel.

Why do scientists assume things they know are not realistic? Available computational capabilities, or lack thereof. Before Blue Waters, open research supercomputers were too slow and had too little memory to handle the massive datasets needed to process some of the smaller scale jostling satellites encounter.

Reed and Matt Ferringer, a project leader at The Aerospace Corporation, are leading a team of researchersthat is adding these smaller pushes and pulls to the old, simplified model with the goal of making satellite systems more efficient and effective. The team involves Reed’s group from Cornell, Eric Wood’s group at Princeton University, and scientists from The Aerospace Corporation, a federally funded research and development center in Chantilly, Va.

Applied systems engineering

The Global Precipitation Mission (GPM) is the project’s test case. The team chose GPM because of its high scientific impact, particularly related to hydrology.

“Precipitation was one of the very first questions that people asked when we started to design space-based earth observation,” says Reed.

Space-based weather observation got its start watching clouds. The National Aeronautical and Space Administration (NASA) launched the first weather satellite, TIROS I, in 1960, a few years after Russia successfully put Sputnik 1 in orbit. TIROS I only lasted 78 days, but it started the world on the road toward global environmental observations.

Initially, Earth-observing satellites used passive sensors to look at different wavelengths of energy that bounced off or were emitted by the Earth and its atmosphere. The quality of sensors has come a long way over the last few decades and the number and variety of satellites has increased, Reed says. The Union of Concerned Scientists estimates that 15 to 20 percent of the more than 1,000 satellites that were in operation at the end of May 2013 take scientific observations of the Earth and space.

GPM’s predecessor, the Tropical Rainfall Measuring Mission (TRMM), contains the first active space-based precipitation radar, somewhat like a floating version of the instrument that shows current rainfall on the nightly news. It focuses on moderate to heavy rainfall in the tropics. GPM will extend coverage of these sorts of measurements to higher latitudes and lighter rain. It will also add more capability to look at the shape and size of frozen and liquid precipitation, which can tell scientists something about how the particles interact with the nearby atmosphere. This greatly enhances the ability to monitor the water budget around the world.

Blue Waters takes a hack at the problem

A more complex satellite system naturally requires more memory to keep in mind all the forces acting on the satellites and more calculations to optimize the satellite locations.

“Every time you account for a deviation from [the simplified simulation] the time to evaluate one design goes up substantially. Then if you account for all of the combinations of these, it goes up dramatically,” Reed says. “Now the question is, can you actually exploit this non-idealized behavior such that the satellite system can correct itself without you having to actively manage it?”

Reed’s team calls this passive control. Exploiting these perturbations to keep the satellite on track can save fuel, which will save money and make the satellite lighter at launch time. Economics are at the heart of this question.

The team’s other question grew from scientific concerns.

“What if GPM was well coordinated and we could optimize all of the precipitation satellite systems together?” Reed says. “So, perfect international coordination. What could we attain?”

Their methodology accounts for multiple objectives, like the cost and lifetime of the satellite system, using a combination of complex mathematics and knowledge of how the parts of the Earth’s water cycle interact.

Reed likened it to buying a car. A simplified model may seek to minimize cost within a set of constraints, like the car must have four doors. Once the consumer lists all the cars that have four doors, he only has to choose the cheapest one in order to fulfill his search criteria. This is like modeling a geostationary satellite orbit by taking into account only a spherical Earth and the weight of the satellite at launch. Both are very simplified problems.

If the consumer also considers a car’s reliability, upkeep costs, gas mileage, and a host of other characteristics, his decision will become more complex. Which characteristic is more important: cheap upkeep, low initial cost or good gas mileage? He will have to strike a balance between these, and there are many cars that fit the bill. For the satellites, this is akin to adding forces like the gravitational pull of the sun and the moon while trying to minimize the cost and maximize the satellite’s useful lifetime.

Reed says that his team will be the first to combine high-resolution astrodynamics design with its practicalimpact on Earth science through a complex global optimization model.

“The only reason that [this approach] is emerging is our team’s allocation on Blue Waters, posing questions which have traditionally just been classified as intractable,” says Reed.

First steps

Before the researchers can address GPM, though, they are using something called the Draim problem to show that their methodology indeed works.

The Draim problem involves only four satellites, less than half of the nine that will be part of GPM. Draim is a patented benchmark for global coverage of satellites. The team is using it as a means to test their perturbation hypotheses before moving on to the bigger GPM system. The original Draim severely degrades because of the perturbations the team is studying, said Reed.

Once the team begins looking at precipitation satellites, achievement will be measured in terms of spatial coverage of the data and frequency of measurements to assess the hydrological impact of changes to the satellite constellation. The team has already looked at the effect of the frequency of satellite rainfall measurements on different parts of the water cycle, specifically precipitation, runoff on the surface, evaporation, and near-surface water in the soil. They have found that the frequency of measurements significantly affects each of these quantities, but in different ways.

Looking forward

The new algorithm may make the constellation more sustainable and economical by reducing perturbations in the satellites’ orbits while maintaining whole-Earth coverage.

Most Earth-observing satellites are only expected to last five to seven years, but some are expected to remain operational for a decade or more and many are used well beyond their expected lifespan. The TRMM satellite, for instance, launched in late 1997 with an expected life span of three years. It is still in service. Because GPM satellites might follow this trend, a more precise orbit may improve the constellation’s useful lifetime.

Reed says that preliminary results from the Draim problem are very promising.

Over the next year the team will apply their work to GPM and other more complex systems that provide a variety of services. Reed says the ultimate goal is to show that his team’s approach improves the longevity of a satellite constellation and quality of data it provides.

“We show the example for one specific application, but it could be global climate change, it could be telecommunication, it could be some other scientific question,” says Reed. “You can plug in the same kind of concepts of blending the on-the-ground science and modeling all the way up to the in-the-sky aerospace engineering and design.

“We will hopefully get new designs that are fundamentally different and innovative, and Blue Waters is the engine to that kind of discovery,” says Reed.

Disclaimer: Due to changes in website systems, we've adjusted archived content to fit the present-day site and the articles will not appear in their original published format. Formatting, header information, photographs and other illustrations are not available in archived articles.

Back to top