Consideration of obstacles in density calculation

[chraibi]

In Gitlab by @gjaeger on Mar 12, 2018, 19:14 [origin]

obstacles are not currently included in the analysis.

[chraibi]

In Gitlab by @gjaeger on Mar 12, 2018, 19:30

UNI_CORR_500_06.zip

[chraibi]

In Gitlab by @chraibi on Mar 12, 2018, 21:59

changed the description

[chraibi]

In Gitlab by @chraibi on Mar 12, 2018, 22:07

Actually, if the geometry is constructed "correctly" it should work, as the following screenshot shows:

JPSreport uses the geometry lib from JPScore. So it expects a certain structure.
It creates a polygon made of walls. This big polygon, contains "holes", which are in our case the obstacles.

The problem here are the 4 adjacent obstacles left. Seems like a bug in Boost. (No chance to fix)

Here a small fix in the geometry, which solves the problem: geo_UNI_CORR_500_062.xml

[chraibi]

In Gitlab by @gjaeger on Mar 13, 2018, 08:49

What is the different between wall and B1 to B4?

<obstacle id="0" caption='wall' height="3.0" >

<obstacle id="0" caption='B1' height="3.0" >

The parameter caption is only for visualization (see manual, page 20). Can you explain it?

[chraibi]

In Gitlab by @chraibi on Mar 13, 2018, 09:31

Yes, it is just a "name". Useful, to debug the geometry with JPSvis and know where is what..

[chraibi]

In Gitlab by @chraibi on Mar 13, 2018, 09:33

changed title from {-Analytics.cpp | consideration of obstacles-} to {+Consideration of obstacles in density calculation+}

[chraibi]

In Gitlab by @gjaeger on Mar 13, 2018, 11:27

The problem is still there.

[chraibi]

In Gitlab by @gjaeger on Mar 13, 2018, 12:19

with

...
    <rooms>
        <room id="0" caption="hall">
          <subroom id="0" caption="Room 0" class="Hall" A_x="0" B_y="0" C_z="0">

            <polygon caption="wall">
                <vertex px="-20.00" py="-2.00"/>
                <vertex px="-20.00" py=" 7.00"/>
                <vertex px=" 20.00" py=" 7.00"/>
                <vertex px=" 20.00" py="-2.00"/>
                <vertex px="-20.00" py="-2.00"/>
            </polygon>

            <obstacle id="0" caption='B1' height="3.0" >
                <polygon>
                    <vertex px="-9.0" py="-1.0"/>
                    <vertex px="-9.0" py="0.0"/>
                    <vertex px="8.0" py="0.0"/>
                    <vertex px="8.0" py="-1.0"/>
                    <vertex px="-9.0" py="-1.0"/>
                </polygon>
            </obstacle>
...

There is still a problem.

[chraibi]

In Gitlab by @chraibi on Mar 13, 2018, 13:25

I see that. The disturbing thing here, is it does not work some times. To rule out any Boost-bugs, I need more tests.

This is the plot of the polygon calculated within jpsreport.

(to be continued!)

[chraibi]

In Gitlab by @chraibi on Mar 13, 2018, 14:27

Ok here is my conclusion: The problem is once again in the two adjacent obstacles. Boost seems not able to handle this kind of "holes" in the geometry. Therefore, the cutting function does not work reliably. Smoothing everything out, shows better behavior:

PS: Look at that wave!! :sunglasses: geo_UNI_CORR_500_06v3.xml

[chraibi]

In Gitlab by @gjaeger on Mar 14, 2018, 13:57

My conclusion. It looks also interesting.

[chraibi]

In Gitlab by @chraibi on Mar 14, 2018, 13:59

Yes indeed!

[chraibi]

In Gitlab by @chraibi on Mar 14, 2018, 13:59

closed