Britain has developed software that enables UAVs to operate
independently, but cooperatively, under the command of a manned aircraft, on
combat missions. The UAVs would, most of the time, take care of themselves,
following the plan of the operation they were programmed with. But the pilot of
manned aircraft could order any of the UAVs to deviate from the plan. The key
to all this is flight control software, and artificial intelligence, that is
capable enough to make it all work. The UAVs also constantly communicate with
each other, and function "cooperatively" (to avoid collisions, or two UAVs
attacking the same target, and so on).
computers are running all this, the reactions are quicker than with human
pilots. The UAVs can perform a number of dangerous missions, like attacking air
defenses, searching for targets on the ground, or defending the entire
"package" (the manned aircraft and all the UAVs) from enemy aircraft. The UAVs
can be used more aggressively, because you are not risking the lives of pilots.
If the manned aircraft is shot down, the UAVs have their programmed orders to
complete the mission, or immediately break off and return to base. It's also
possible to use this software so that one human operator on the ground can
control a swarm (half a dozen or so) UAVs for, say, a recon mission, to scour
an area for enemy targets. This saves a lot of skilled manpower, as all you
need is someone to review the video. And increasingly powerful software is
capable of doing that as well, leaving only for the human analyst to confirm
that something important has been found.
kind of software has been around, in a conceptual form, for decades. But it
took powerful enough hardware (in terms of computing power, and high quality
digital vidcams), and decades of tweaking the software, to reach the point
where the stuff actually works.