UAVs (Unmanned Aerial Vehicles) are thought of as "robotic aircraft." In reality, they are very manpower intensive. On the high end, we have the $30 million dollar Global Hawk, a UAV that can fly across the Pacific without any human intervention. But there is always someone on duty as a backup. When flying across the Pacific, the Global Hawk requires the services of 2-3 operators on the ground. Plus there is a maintenance crew of about two dozen technicians, waiting to go to work when the UAV lands. The Global Hawk is a lot more automated than the smaller, like Predator, and more widely used UAVs. Predator has a two man flight crew on the ground (a "pilot" and a "sensor operator.") There's also technical people on call in case there are communication problems. Moreover, most UAVs can stay in the air for a long time, requiring flight crews to work in shifts. Worse, it's actually harder to fly a UAV than it is an aircraft you are in. That's because you can see less (just what's viewed by the nose cam) and can't "feel" what the UAV is doing. While UAVs can be programmed to fly a course by itself, the human pilot often has to take over when the weather gets rough, or the sensor operator spots something that requires the UAV to break away from its pre-programmed course. While much smaller than Global Hawk (and costing a tenth as much), a Predator platoon of four UAVs still requires the services of 63 pilots and technicians. It's also proving difficult to create UAVs that are really "smart," and this is making it difficult to create UAVs that can be controlled in groups by one operator. There is already several decades of research on software and sensors to make "autonomous" vehicles (land and aerial) possible. So the scope of the problem is known, and those developing UCAVs (combat UAVs) feel that they will have a computer program smart enough to fly a UCAV or UAV without a lot of human intervention.