Future aerial robots have to deal more intelligently with uncertainty. For example, flying in a city exposes you to unpredictable, turbulent air flow, which is particularly troublesome for very small aircraft like Micro Air Vehicles (MAVs). If you're flying indoors to survey a damaged building, say, you have to deal with unexpected obstacles. In both scenarios, you've probably lost your GPS reception, so the aircraft needs to track its position and surroundings using its own on-board sensors. And, of course, all the computers and sensors have to be lightweight so they fit on the aircraft.