[Studies in Computational Intelligence 481] Artur Babiarz, Robert Bieda, Karol Jędrasiak, Aleksander Nawrat (auth.), Aleksander Nawrat, Zygmunt Kuś (eds.) - Vision Based Systemsfor UAV Applications (2013, Sprin
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
220 K. Daniec et al.<br />
The construction and service for unmanned aircraft is fairly expensive and it<br />
should be added that the cameras are usually the most expensive part of the aircraft.<br />
Furthermore, prototyp<strong>in</strong>g autonomous flight algorithms and vision based<br />
algorithms is time consum<strong>in</strong>g and requires multiple tests and a priori assumptions.<br />
In prototyp<strong>in</strong>g the algorithms helpful appear to be a real-world simulators. There<strong>in</strong><br />
can be tested the control algorithms <strong>in</strong>clud<strong>in</strong>g the dynamics of fly<strong>in</strong>g object, GPS<br />
system simulation and among others earth model WGS-84. Us<strong>in</strong>g the simulator<br />
allows to run the tests without fear of destruction of the real object, and what goes<br />
with it, from possible accidents and exposure to the risk of third parties. Us<strong>in</strong>g a<br />
simulation of the real world we are mak<strong>in</strong>g a large f<strong>in</strong>ancial sav<strong>in</strong>gs result<strong>in</strong>g from<br />
no need for rebuild<strong>in</strong>g the model <strong>in</strong> the case of (undesirable) breakdown. We also<br />
save the time that we could spend much more, dur<strong>in</strong>g the design of mechanisms<br />
for autonomous flight. Us<strong>in</strong>g a simulation environment Prepar3D® Lockheed<br />
Mart<strong>in</strong>, we are able to get data from all k<strong>in</strong>ds of sensors and also to control the<br />
fly<strong>in</strong>g object model. This environment has mapped physics, not only <strong>in</strong> the air, but<br />
also on land and <strong>in</strong> the water. Importantly, the simulation allows us to change the<br />
weather, which greatly <strong>in</strong>fluences the dynamics of mov<strong>in</strong>g objects <strong>in</strong> it. The simulation<br />
<strong>in</strong>cludes, for example, build<strong>in</strong>gs and artificial traffic, which can be used to<br />
prototype the vision based algorithms. In addition, it should be mentioned, that<br />
realistic is also a terra<strong>in</strong>, as well as locations of all airports and cities.<br />
Ma<strong>in</strong> contribution of this paper are control algorithms for <strong>UAV</strong> on the set trajectory.<br />
This trajectory can overlap with the exist<strong>in</strong>g highway network, waterways<br />
and air routes. Thanks to this, it is possible to use a video stream from the onboard<br />
virtual camera for prototyp<strong>in</strong>g vision based algorithms.<br />
2 Literature Review<br />
The Control of the unmanned objects based on vision is an active area of research.<br />
In many publications we met with topics devoted just to control of fly<strong>in</strong>g objects<br />
us<strong>in</strong>g a flight simulators. One of them, worth mention<strong>in</strong>g is the work of Eric R.<br />
Mueller [1], which was based on the studies carried out with Hardware-<strong>in</strong>-the-<br />
Loop simulation. Simulation was based there on Matlab/Simul<strong>in</strong>k® mathematic<br />
model, while the control values was calculated by us<strong>in</strong>g LQR controllers [2].<br />
Another important publication is the work of Eric N. Johnson and Sumit'a Mishra<br />
[3], which uses simulation both <strong>in</strong> SIL (Software In the Loop) and HIL (Hardware<br />
In the Loop) and the simulation visualization. In the will of the explanation, the<br />
SIL-loop simulation controls a physical process model (the simulator) us<strong>in</strong>g the<br />
software, which is a model of the regulator, which was run on the same hardware<br />
platform. In the HIL-simulation process is <strong>in</strong>volved a real regulator, which communicates<br />
with the simulator through the appropriate <strong>in</strong>terfaces. In the articles [4]<br />
and [5] the <strong>auth</strong>ors use a simulation environment implemented <strong>in</strong> the "Matlab/Simul<strong>in</strong>k®<br />
6-DOF", and visualization, "FlightGear". The simulator that was<br />
used there is characterized by a greater reality, because it <strong>in</strong>cludes the basic model<br />
of atmospheric, gravity and magnetic field.