Sensors and Methods to Handle UAV

Unmanned Aerial Vehicles (UAVs) play an increasingly important role in many military scenarios. For the Navy to incorporate these vehicles into its critical missions, there has to be an efficient protocol to handle them on the decks of aircraft carriers.Current handling approaches and procedures have substantial limitations. For example, GPS-based taxiing via pre-set coordinates assumes that the area is clear of all other traffic, while joystick remote control is cumbersome and relies on specially trainedpersonnel. Although an elaborate sensor network positioned on the deck could alleviate these problems, its cost would be prohibitively high.We propose to develop a UAV-based intelligent navigation system that relies solely on image cues. Our sensor suite will be configured using commercial off-the-shelf video cameras and mounted in the nose section of the UAV. The cameras have small formfactors and support both daytime and nighttime operation. The video processor consists of object detection, tracking, and range estimation modules. These low-level modules support two higher level functions-recognizing taxiing gestures given by a humancontroller and path planning for obstacle avoidance. An intelligent controller fuses information from these modules with scenario knowledge to achieve robust situation assessment throughout the UAV's taxiing. The intelligent vision-based command recognition and navigation system developed under this effort would find immediate use in numerous defense and civilian scenarios. Examples include operating robotic vehicles in hazardous or dangerous environments andemergency management (finding survivors in wreckages). Initiatives to provide UAV access to the national airspace, such the UAV National Industry Team (UNITE), have already identified the need for this type of vision technology.