FHWS/IPIN2018
Archived
4
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

eval opti nervt mich...

Browse Source
This commit is contained in:
toni
2018-09-14 13:33:41 +02:00
parent ade1677fa4
commit 9e2bb6fba9
4 changed files with 20 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
tex/chapters/experiments.tex
View File

@@ -102,6 +102,8 @@ This is more difficult using the mash and requires the handling of baricentric c
\subsection{\docWIFI{} Optimization}
\label{sec:exp:opti}
\commentByToni{Work in Progress... Irgendwie passt die Grafik nicht so wirklich. Im Gegensatz zum 2017 Paper würde ich gerne ein wenig über die geschätzten Positionen reden. Die Unterschiede zwischen Local und Global dabei. Warum machne Schätzungen gar so weit weg von der Realität sind und das es oft auch gar nicht so schlimm ist, falls das passiert. Tipps sind Willkommen}
%wie viele ap sind es insgesamt?
As described in section \ref{sec:wifi} we used \SI{42}{} WEMOS D1 mini to provide a \docWIFI{} infrastructure throughout the building.
Within all Wi-Fi observations, we only consider the beacons, which are identified by their well-known MAC address.
@@ -121,17 +123,24 @@ Fig. \ref{fig:apfingerprint} illustrates the results of the global and the per-f
\def\svgwidth{\columnwidth}
\input{gfx/optimization/side/sideplot2.eps_tex}
\caption{Side view}
\label{fig:apfingerprint:b}
\end{subfigure}
\caption{Position of Ap's optimized with global and per floor and real.}
\label{fig:apfingerprint}
\end{figure}
%
It can be seen that the red...
Here, the respective optimized positions are connected by a grey line with the corresponding ground truth.
In order not to be confused by overlaps, the side view in fig. \ref{fig:apfingerprint:b} only shows the eight \docAPshort{}'s installed on the right side of the building.
Looking at fig. \ref{fig:apfingerprint}, it can be seen that most optimized positions differ in the lower single-digit meter range.
gives a good idea on how optimizing a simplified signal-strength prediction model behaves.
of course, the position alone does not provide sufficient information of the overall performance of the optimiziation since they give no information about the other optimized parameters (bla, blub and bliib).
Nevertheless, fig. \ref{} gives an idea on how optimizing a simplified signal-strength prediction model behaves.
%Nevertheless, fig. \ref{} gives an idea on how optimizing a simplified signal-strength prediction model behaves.
By only considering ceillings, the attenuation factore... and thus most z coordinates for the ap's are similiar.
The main message from this is, that wherever the ap's are optimized they are optimized to perfectly fit the underlying signal-strength model.
thus the optimized parameters provide far better localization results compared to just using the real ones. simply because modelling the realistic incidents is very time consuming.

37
tex/chapters/transition.tex
View File

@@ -121,40 +121,5 @@
Likewise, just using some random position, omitting heading/steps might be viable as well.
\commentByFrank{es gaebe noch ganz andere ansaetze etc. aber wir haben wohl nicht mehr genug platz :P}
like
max. 1 Seite
\subsection{Mapping}
%\begin{figure}
% \centering
% \includesvg[width=8.7cm]{gfx/map1}
% \caption{
% wonderful image of a wonderful floorplan within a wonderful building
% \commentByFrank{vlt ein cooles mittelalter beispiel? schwerter oder bilderrahmen an der wand?}
% }
%\end{figure}
\commentByFrank{
was haben wir bisher verwendet (graph)\\
was sind die probleme (speicher, diskret)\\
was verwenden wir deshalb jetzt (triangles)\\
was sind die vorteile / nachteile? (weniger speicher, teils schneller, teils langsamer)
}
\begin{itemize}
\item Karte wird manuell über ein Tool erstellt bei dem wände, türen, fentser etc. eingezeichnet werden
\item Karte unterscheidet bereiche, treppen etc.
\item dort werden auch metainformationen wie ap etc. eingetragen
\item daraus wird ein mesh generiert über bla blub mit bla blub. cite cite cite :)
\end{itemize}
\subsection{PDR}
\begin{itemize}
\item aktuelle bewegungsmodell
\item ...
\end{itemize}
\commentByToni{ich denke aber auch, es langt.}

8
tex/gfx/optimization/heightPlot.gp
View File

@@ -1,7 +1,7 @@
reset
#set terminal x11 size 2000,1500
set terminal windows size 2000,1500
set terminal x11 size 2000,1500
#set terminal windows size 2000,1500
#set terminal epslatex size 3.5,2.0
#set output "plot.tex"
@@ -19,4 +19,6 @@ set key inside left top box opaque
plot \
'real_with_zoffset.csv' using 1:4 pt 1 ps 1 lc rgb '#000000' title 'Ground Truth' ,\
'perFloor.csv' using 1:4 pt 2 ps 1 lc rgb '#FF0000' title 'Per Floor' ,\
'global.csv' using 1:4 pt 10 ps 1 lc rgb '#0000FF' title 'Global' ,\
'global.csv' using 1:4 pt 10 ps 1 lc rgb '#0000FF' title 'Global' ,\
pause -1

6
tex/gfx/optimization/plot.gp
View File

@@ -1,7 +1,7 @@
reset
#set terminal x11 size 2000,1500
set terminal windows size 2000,1500
set terminal x11 size 2000,1500
#set terminal windows size 2000,1500
#set terminal epslatex size 15cm,22cm
#set output "plot.tex"
@@ -63,5 +63,5 @@ splot \
'perFloor.csv' using 2:3:4 with point pt 7 ps 1 lc rgb '#FCAF3E' notitle ,\
'global.csv' using 2:3:4 with point pt 7 ps 1 lc rgb '#0D6DCD' notitle ,\
pause -1