Today's military satellites are $500 million miracles of modern technology. They're also idiots; they don't even have enough sense to get out of the way when space junk or an enemy weapon is about to crash into them. The Air Force is looking to change that. Soon, those satellites might get minds of their own.
A recent call for research proposals outlines three main tasks for this space automaton: detecting threats, identifying malfunctions, and scanning for missile launches. Traditionally, all of those processes take a lot of (human) brain power – most satellites require at least a dozen humans on Earth to keep them in check, guard against space debris collisions, and analyze satellite data for missile threats.
But doing these jobs from the ground isn't exactly efficient. Figuring out what's wrong with the satellite is like long-distance detective work, according to Brian Weeden, a former U.S. Air Force officer and space security expert. For example, it took months to figure out why Galaxy 15 suddenly stopped responding in April 2010. The incident was initially blamed on a a geomagnetic storm. But it turned out that a software malfunction due to electrostatic discharge was to blame. A satellite that could diagnose and fix its own problems – or at least relay the specifics back to humans – could save valuable time and resources.
Satellites are already becoming more independent. With a technology called Sciencecraft, they can collect and analyze images, sending only the most relevant scientific observations back to Earth. But snapping images is one thing – detecting and responding to threats is quite another. What the Air Force has in mind would involve giving satellites the "capability to detect and isolate non-deterministic events, characterize these events, plan resulting actions and then execute activities." In other words, satellites wouldn't need much human input at all.
But even a single one of those four functions is a hugely complex task. Each particular event – whether a component failure or oncoming shard of space trash – requires a different response. So the first step is simply gathering and synthesizing enough information to accurately assess the situation. That would involve processing sensor data, historical information stored on board, environmental conditions, and mission objectives.
While the Air Force solicitation concentrates on threat detection, it also claims to be on the lookout for a more fully autonomous flight system. Which begs the question – how autonomous would this system actually be? Even with gobs of information, any automated process is bound to make mistakes. Car airbags are designed to respond automatically to a threat, but they occasionally mess up.
"Having an airbag accidentally fire is an annoyance," Weeden said, "having an automated threat response/action system misfire on your $500 million satellite that is 36,000 km away is something else entirely."
Photo: U.S. Air Force
See Also:- Military Wants to Read Satellite Info by Playing Dance Dance Revolution
