13.08.2018 Views

[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

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Probabilistic Approach to Plann<strong>in</strong>g Collision Free Path of <strong>UAV</strong> 195<br />

Fig. 3. 3D MATLAB model of the real map<br />

2.4 Probabilistic Approach<br />

S<strong>in</strong>ce the search space <strong>in</strong> 3D models is large that implies hard computations have<br />

to be performed <strong>in</strong> order to f<strong>in</strong>d the solution. The probabilistic approach allows to<br />

simplify the problem. Instead of search<strong>in</strong>g through entire space a number of representative<br />

elements are selected. For <strong>in</strong>stance if the map has a resolution 100x100<br />

<strong>in</strong> the process of search<strong>in</strong>g the solution requires 10000 elements to be processed.<br />

The planner searches the optimal solution us<strong>in</strong>g the reduced space. Let us denote<br />

the randomly selected set of representative samples of the orig<strong>in</strong>al model as:<br />

where<br />

, ,…, (4)<br />

, , 1, 1, (5)<br />

In (5) the elements , are <strong>in</strong>teger values selected <strong>in</strong> a random way from the given<br />

range. An exemplary conta<strong>in</strong><strong>in</strong>g 25 elements, random realization of the subspace<br />

is presented <strong>in</strong> fig. 3. Us<strong>in</strong>g representative elements conta<strong>in</strong>ed <strong>in</strong> (4) the process<br />

of search the collision free path is performed. If the size of the (4) is not large<br />

enough or the distribution of po<strong>in</strong>ts does not allow to f<strong>in</strong>d admissible solution, the<br />

process of select<strong>in</strong>g the representative subspace is repeated or the size of the set<br />

(4) is <strong>in</strong>creased.<br />

2.5 Graph Model<br />

The next stage of the process of f<strong>in</strong>d<strong>in</strong>g the solution is creat<strong>in</strong>g the model that will<br />

be a base for the plann<strong>in</strong>g process. Here <strong>in</strong> this study the graph representation of<br />

the random subspace is proposed. Hence the model has a form of weighted graph<br />

def<strong>in</strong>ed as:

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!