From the thesis abstract: "The US military already uses some semi- and fully autonomous technology (in both UAS and manned systems) and, as UAS [ autonomous unmanned aerial systems] technology matures, will look to increase these systems' autonomy. Despite improvements in UAS technology, the USAF will not be able to rely exclusively on autonomous systems because of their significant technological limitations, political resistance within the United States, international obstacles, doctrinal and operational problems, and service-related impediments. These obstacles will require the USAF to maintain a diverse force of autonomous UAS, manned aircraft, and man-in-the-loop (MITL) controlled UAS. Future autonomous UAS will have several benefits. These are: the ability to eliminate risk to US personnel, the reduction in personnel requirements and the ability for new technology and capabilities to be quickly propagated across the entire fleet. To be able to capitalize on these benefits, autonomous UAS will need to accomplish certain technological milestones. The first is the ability to accomplish basic aircraft tasks (takeoff, land, in-flight refueling, etc). Additionally, autonomous UAS must be able to operate within future bandwidth constraints, as well as ensure the security of communications sent to/from UAS. Lastly, UAS will need to be able to correctly identify airborne and ground targets. This identification requirement will be most difficult during counterinsurgency missions where it is unrealistic for UAS to be able to correctly distinguish between civilians and insurgents."
Defense Technical Information Center (DTIC): http://www.dtic.mil/dtic/