Augmenting camera images for operators of unmanned aerial vehicles
J.A. Veltman & A.B. Oving
TNO Human Factors
|
Controlling the camera of an unmanned aerial vehicle (UAV) is a difficult task. The part of the world that is viewed by the camera is
often very small due to the high altitude of the UAV and large the zoom of the camera. Furthermore, operators become disoriented easily
because the viewing direction is often not the same as the flight heading of the UAV. Moreover, the data links between the control unit
and the UAV are hampered often by time delays and reduced update rates. The consequence is that steering the camera solely on the basis
of the camera pictures is very difficult.
TNO Human Factors developed a simulation environment that can be used to demonstrate possible problems with camera control and to test solutions for these problems. For instance, new principles to augment the view of the operators are tested in this simulation environment. One of these principles is the use of a 3-D database of the environment that is drawn from the viewpoint of the camera. The advantage is that the "footprint" of the camera can be calculated real time by a computer near the ground control station. This overcomes the problems of time delays and reduced update rates. Furthermore, the viewing angle is much larger than the field of view of the camera which might improve the situation awareness. An experiment was conducted to test this augmentation principle. Subjects had to locate targets under a) optimal camera conditions (no time delay and 30Hz update rates), b) with 1-second time delay and c) with a 3Hz update rate. Furthermore, the workload was increased by adding a secondary memory task in one half of the conditions. Subjects performed all the tasks with and without the 3-D database. The performance was measured extensively and the workload was measured with subjective (RSME) and physiological workload measures (HR, HRV, BP, respiration and eye activity). |
Lezing, te presenteren op het symposium 'Echt of Namaak: simulatie en modellering in beweging', 5 oktober 2001, Rijksuniversiteit Groningen.