Thursday, May 08, 2014

Combat Augmentation Drones--Another Take on Autonomous Killers

The future of drones in the US is murky and uncertain, but what does seem clear is that they are here to stay.  Robots, unmanned aerial vehicles, pilotless aircraft, and a slew of other forms of autonomous or remotely controlled machines will, throughout the 21st century, become an increasingly omnipresent technology.  They have already been used to investigate and track criminals, spy on nude celebrities, and deliver fresh craft beer to ice fishermen. Future users will include everyone from Amazon to firefighters to insurance companies.  The US military has been on the cutting edge of drone technology, and will likely remain there, with the Air Force in particular looking to continue expanding the numbers and types of unmanned aircraft it operates.

The FAA, spoilsports that they are, have already banned the practice.



In my presentation, I focused on some of the more realistic and current development projects that are currently underway with the US and foreign militaries, as well as technology currently under development in the civilian research sector.  To recap, one of what I believe are the most important developments is drone swarm technology, or the ability of drones to coordinate their actions and solve problems according to prior programming without human input at the time that actions are taken.  This can be seen in the video presented here:


While this may appear incredible, it's important to realize that this is not any sort of artificial intelligence.  Instead, it's complex programming to allow the drones to coordinate and appear to solve problems independently.  While the technology and algorithms will no doubt become more advanced with time, they are merely coordinating rather than acting intelligently.

As this ability for coordination between drones and other aircraft becomes more complex, I believe that it will revolutionize the way that air combat works.  Many people take an absolutist view of autonomous killing robots:  The Terminator, acting completely independently, choosing its own targets, killing them and anything else that gets in its way.

It's ok, I'd vote for him too.
What if, instead of flying independent combat patrols, drones were integrated into the fighter control network?  Robotic wingmen that autonomously attack enemy planes . . . but only those specifically targeted by their lead fighter or an AWACS operator.  In a theoretical post-RMA air battle, with integrated command and control structures and radar coverage of the entire battle space, F-22s and F-35s would transmit targeting data not only to each other, but also to robot fighters operating independently of human control.  These would be capable of autonomously coordinating tactics with their human counterparts, attacking targets identified as hostile without direct human input beyond the initial targeting data.

While it sounds far-fetched, the technical hurdles are rapidly approaching the realm of being problems of implementation.  The networking capabilities of the F-35 should be capable of interfacing with unmanned aircraft if they were to exist, and the programming could borrow heavily from autopilots and AIs from flight simulation video games, some of which are capable of advanced competition against human pilots (albeit in a video game environment).

The concept has actually been explored already in a number of science fiction works and video games.  One of these is Eve Online, a science fiction space combat MMO which features incredibly complex and realistic economics, including the destruction of spaceships and fighters.

An unmanned combat drone in EVE online.  Not pictured: the interceptor it's deployed from.
Many ships and fighters in Eve carry a small (or large) compliment of drones which patrol close to their mothership, attacking the enemies its ordered to by the controlling ship's pilot.  These greatly enhance the combat ability of many ships, overwhelming the countermeasures and defenses of opponents and often making major contributions to the outcome of dogfights.  There are of course a number of technological and procurement impediments to implementing such a system, but the technology is realistic and the tactical advantages may be numerous.  The command and control communications would also be much more robust, as short range line of sight communications are more difficult to jam, especially given multiple transmission technologies.

While the notion may seem far fetched and unrealistic at present, it may be less so than handheld wireless personal communicators must have seemed when Star Trek first aired in the 1960s.  And what's that? I've got a message coming in--my Android tells me it might be important, so I guess that's all for now.

No comments: