Changes
This commit is contained in:
@@ -1,29 +1,71 @@
|
||||
\section{Related Work}
|
||||
\label{sec:relatedWork}
|
||||
|
||||
\begin{itemize}
|
||||
%\begin{itemize}
|
||||
%
|
||||
% \item Thema: Wi-FI
|
||||
% \item klassiche auf Wi-Fi RSSI verweisen, wo kommt es her und was gibt es da für Lokalisierung
|
||||
% \item RTT Verfahren grob erläutern und viel zitieren
|
||||
% \item Wi-Fi RTT bisher schwer, gab einige Hacks für Lokalisierung... die zitieren
|
||||
% \item FTM im Standard von IEEE -> die 3 wichtigen Paper dazu
|
||||
% \item Mit welchem Methoden Schätzen wir für FTM und RSSI eine Position? (Position Estimation)
|
||||
% \subitem Trilateration klassischer Ansatz
|
||||
% \subitem Kalman Filter
|
||||
% \subitem Probabilistische methode
|
||||
% \subitem Particle Filtering macht bla und blub und ist bla und blieb - Hier: Ganz simples Bewegungsmodell für PF, damit Vergleichbarkeit gegeben bleibt.
|
||||
% \item Klassische Lokalisierungssysteme nutzen meist Particle Filter mit RSSI haben Genauigkeiten von xx und FTM verspricht hier besser Genauigkeiten weil... und hat die und die Vorteile.. deswegen soll diese Arbeit den grundlegenden Mehrwert von FTM gegenüber RSSI untersuchen.
|
||||
%\end{itemize}
|
||||
%
|
||||
%
|
||||
%\etal{Ibrahim} lays the required groundwork to use the still experimental FTM standard and verifies the general accuracy \cite{ibrahim2018verification}.
|
||||
%\etal{Yu} present a system using FTM measurements and multisensor multi-pattern-based dead reckoning based on a Unscented Kalman filter sensor fusion \cite{yu2019robust}.
|
||||
%\etal{Xu} \cite{xu2019locating}.
|
||||
%
|
||||
%Compared to the above state of the art our work...
|
||||
%\begin{itemize}
|
||||
% \item szenario realistischer und nicht nur quadrate die man läuft
|
||||
% \item direkter vergleich mit RSSI
|
||||
% \item in der praxis erprobte verfahren darauf laufen lassen.
|
||||
%\end{itemize}
|
||||
|
||||
\item Thema: Wi-FI
|
||||
\item klassiche auf Wi-Fi RSSI verweisen, wo kommt es her und was gibt es da für Lokalisierung
|
||||
\item RTT Verfahren grob erläutern und viel zitieren
|
||||
\item Wi-Fi RTT bisher schwer, gab einige Hacks für Lokalisierung... die zitieren
|
||||
\item FTM im Standard von IEEE -> die 3 wichtigen Paper dazu
|
||||
\item Mit welchem Methoden Schätzen wir für FTM und RSSI eine Position? (Position Estimation)
|
||||
\subitem Trilateration klassischer Ansatz
|
||||
\subitem Kalman Filter
|
||||
\subitem Probabilistische methode
|
||||
\subitem Particle Filtering macht bla und blub und ist bla und blieb - Hier: Ganz simples Bewegungsmodell für PF, damit Vergleichbarkeit gegeben bleibt.
|
||||
\item Klassische Lokalisierungssysteme nutzen meist Particle Filter mit RSSI haben Genauigkeiten von xx und FTM verspricht hier besser Genauigkeiten weil... und hat die und die Vorteile.. deswegen soll diese Arbeit den grundlegenden Mehrwert von FTM gegenüber RSSI untersuchen.
|
||||
\end{itemize}
|
||||
% complexe filter etc machen eine Aussage schwer
|
||||
|
||||
|
||||
\etal{Ibrahim} lays the required groundwork to use the still experimental FTM standard and verifies the general accuracy \cite{ibrahim2018verification}.
|
||||
\etal{Yu} present a system using FTM measurements and multisensor multi-pattern-based dead reckoning based on a Unscented Kalman filter sensor fusion \cite{yu2019robust}.
|
||||
\etal{Xu} \cite{xu2019locating}.
|
||||
|
||||
Compared to the above state of the art our work...
|
||||
\begin{itemize}
|
||||
\item szenario realistischer und nicht nur quadrate die man läuft
|
||||
\item direkter vergleich mit RSSI
|
||||
\item in der praxis erprobte verfahren darauf laufen lassen.
|
||||
\end{itemize}
|
||||
The FTM protocol was introduced in the \ieeWifiFTM standard but it only started recently to get more and more attention in the scientific community.
|
||||
One of the earliest work was presented by Intel \cite{banin2016wifi} using a Bayesian Filter to reduce the error in the position estimate.
|
||||
Two years later \etal{Ibrahim} \cite{ibrahim2018verification} gave a comprehensive verification study of the precision and accuracy of FTM with outdoor and indoor measurements in both the \SI{2.4}{GHz} and \SI{5}{GHz} frequency band.
|
||||
They conclude that FTM is capable to provide meter-level ranging in open space scenarios, but in the presence of multipath effects the accuracy drops to \SI{5}{m} indoors.
|
||||
Additionally, \etal{Ibrahim} include a detailed description on how to setup FTM on Linux with \intelOld cards in their work which presumable lead to more publications because off the shelf FTM compatible hardware was now available to a broader audience.
|
||||
|
||||
\etal{Yu} \cite{yu2019robust} developed a complete indoor positioning system based on a FTM ranging model combined with a robust dead reckoning algorithm.
|
||||
A Unscented Kalman filter is used to fuse the sensors
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
Previous work has only focused on verifying the ranging accuracy of FTM and integrating FTM as part of a larger system.
|
||||
It is not yet known how FTM performs in indoor positioning scenarios.
|
||||
The application of complex filters and the combination of several sensors makes it hard to reason about the impact of FTM on a indoor positioning system.
|
||||
In addition, most experiments are limited to rather optimistic test setups, where the results hardly applicable to real world situations with long and relative complex walks.
|
||||
For these reasons the goal of this work is set to directly evaluate the indoor positioning performance with FTM.
|
||||
This is archived by intentionally using a simple particle filter system without a complex movement model and without integrating other sensors.
|
||||
To evaluate the findings we compare the FTM results to RSSI based positioning.
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
@@ -103,7 +103,6 @@ Additionally, the computation of the wall-attenuation factor model requires cost
|
||||
%- walls
|
||||
%, also dynamic obstacles like persons can interfere with the signal.
|
||||
|
||||
|
||||
\subsection{Fine Timing Measurement}
|
||||
%Time-based distance measurements are intuitively based on the delay the signal took to travel from the sender to the receiver.
|
||||
%Multiplied by the propagation speed of light results in the distance between the two nodes.
|
||||
|
||||
@@ -218,31 +218,81 @@ However, the overall error of the device combinations is reasonable small and it
|
||||
|
||||
|
||||
\subsection{Range Measurements in NLOS Scenario}
|
||||
Während der posionierung ist an gewissen stellen eine systematische abweichung zum gt aufgefallen dies trat immer an zwei stellen auf, daraus folgender die hypthese das es irgendwie an dieser gegebenheit liegen muss. als folge daraus ist dann eben entstanden das wir die brandschutztüren gesehen haben und dann eine experiment aufgebaut haben um die hypthese zu messen, inwiefern sich die türen oder besser gesagt wie sich unterschiedliche wandmateriellien, insbesondere wenn diese so extrem wie hier sind, auf die FTM Messungen auswirken.
|
||||
%Während der posionierung ist an gewissen stellen eine systematische abweichung zum gt aufgefallen dies trat immer an zwei stellen auf, daraus folgender die hypthese das es irgendwie an dieser gegebenheit liegen muss. als folge daraus ist dann eben entstanden das wir die brandschutztüren gesehen haben und dann eine experiment aufgebaut haben um die hypthese zu messen, inwiefern sich die türen oder besser gesagt wie sich unterschiedliche wandmateriellien, insbesondere wenn diese so extrem wie hier sind, auf die FTM Messungen auswirken.
|
||||
%
|
||||
%rssi grafik lassen wir weg (bst2rssi.png) und schreiben das nur im text. müssen ja nur den sprung von 7 auf 8 beschreiben. interessanter sind die FTM Plots mit der Streuung (bst2ftm.png)
|
||||
|
||||
rssi grafik lassen wir weg (bst2rssi.png) und schreiben das nur im text. müssen ja nur den sprung von 7 auf 8 beschreiben. interessanter sind die FTM Plots mit der Streuung (bst2ftm.png)
|
||||
|
||||
During the analysis of the recorded data of the test walks presented below systematic and reproducible deviations of the estimated position to the groundtruth were found.
|
||||
These effects increased the error significantly and are bound to specific locations in the building.
|
||||
It is likely that environmental factors of the building at these locations affect the FTM distance measurements process.
|
||||
%Therefore, environmental factors of the building structure at these locations are likely to affect the FTM distance measurements.
|
||||
While it is well known that the environment will affect measurements, especially indoors, it is nevertheless interesting to analyses the underlying cause.
|
||||
It was noticed that the ranging measurements to one access point suddenly started to heavily overestimate the true distance.
|
||||
This effect occurs as soon as the pedestrians walks by a fire door.
|
||||
These heavy doors are about \SI{12}{cm} thick and made of metal.
|
||||
In the case of a fire outbreak these doors are automatically closed, but normally they are not closed and tucked away between walls.
|
||||
Whenever such fire door is in the line of sight between the access point and smartphone the ranging error increases significantly.
|
||||
|
||||
To quantify the impact of these fire doors on the FTM measurement we created two test setups as seen in \autoref{fig:BSTExp}.
|
||||
In the first experiment as shown in \autoref{fig:BSTExp:a} we used the same access point position as in the test walks described below.
|
||||
We placed seven measurement points on a circle so that most of these points are in the main hallway.
|
||||
The radius of the circle is \SI{10}{m} and measure point 1,2 and 3 are located in the shadow of the fire door while points 4 to 7 are not.
|
||||
At every point we placed the \pixelOld on a metal stand \SI{1.05}{m} above the floor and recored FTM distance measurements for \SI{60}{s} which results in around 255 distance measurements per point.
|
||||
|
||||
\begin{figure}[ht]
|
||||
\centering
|
||||
\begin{minipage}{.5\textwidth}
|
||||
\centering
|
||||
\subfloat[]{\label{fig:BSTExp:a}\includegraphics[width=\textwidth]{VersuchsaufbauBST1.png}}
|
||||
\end{minipage}%
|
||||
\begin{minipage}{.5\textwidth}
|
||||
\centering
|
||||
\subfloat[]{\label{fig:BSTExp:b}\includegraphics[width=0.8\textwidth]{VersuchsaufbauBST2.png}}
|
||||
\end{minipage}\par\medskip
|
||||
\caption{Test setups to evaluate the impact of fire doors (red lines) compared to regular walls (black lines). In \ref{fig:BSTExp:a} the measurement points are placed on a circle to keep the distance constant. The setup of \ref{fig:BSTExp:b} allows more measurement points. }
|
||||
\label{fig:BSTExp}
|
||||
\end{figure}
|
||||
|
||||
Notice that point 2, 3, 5 and 6 are located near a stairways with massive metal railings.
|
||||
While it is expected that the stairways also disturb the measurement the measure points are actually included in the test walks.
|
||||
|
||||
|
||||
In order to evaluate the effect of the fire door exclusively, we build a second test setup at a corner office located next to a fire door.
|
||||
As seen in \autoref{fig:BSTExp:b} it was not possible to keep the distance to the AP constant, due to structural limitations.
|
||||
The groundtruth was obtained by carefully measuring the distances to walls and taking the line of sight distance from a true to scale map.
|
||||
|
||||
\begin{figure}[ht]
|
||||
\centering
|
||||
\begin{minipage}{.55\textwidth}
|
||||
\centering
|
||||
\subfloat[]{\label{fig:Bst1Results:a}%
|
||||
\resizebox{\textwidth}{7cm}{\input{plots/BSTPlot1.pgf}}%
|
||||
% \includegraphics[width=\textwidth, axisratio=1.5]{plots/BSTPlot1.pgf}%
|
||||
}
|
||||
\end{minipage}\hspace{.09\textwidth}
|
||||
\begin{minipage}{.35\textwidth}
|
||||
\centering
|
||||
\subfloat[]{\label{fig:Bst1Results:b}%
|
||||
\resizebox{\textwidth}{7cm}{\input{plots/BSTPlot1Rssi.pgf}}%
|
||||
% \includegraphics[width=\textwidth, axisratio=0.86]{plots/BSTPlot1Rssi.pgf}%
|
||||
}
|
||||
\end{minipage}\par\medskip
|
||||
\label{fig:Bst1Results}
|
||||
\caption{}
|
||||
\end{figure}
|
||||
|
||||
|
||||
\begin{figure}[ht]
|
||||
|
||||
\begin{minipage}{.5\textwidth}
|
||||
\centering
|
||||
\subfloat[]{\label{main:a}\includegraphics[width=\textwidth]{VersuchsaufbauBST1.png}}
|
||||
\end{minipage}%
|
||||
\begin{minipage}{.5\textwidth}
|
||||
\centering
|
||||
\subfloat[]{\label{main:b}\includegraphics[width=0.8\textwidth]{VersuchsaufbauBST2.png}}
|
||||
\end{minipage}\par\medskip
|
||||
\centering
|
||||
\subfloat[CDF of error]{\label{main:c}\includegraphics[width=\textwidth]{bst1ftm.png}}
|
||||
\par\medskip
|
||||
\centering
|
||||
\subfloat[CDF of error]{\label{main:c}\includegraphics[width=\textwidth]{bst2ftm.png}}
|
||||
|
||||
\caption{die einzelnen aufbauten können wir in die grafiken der ergebnisse embedden rechts oben ins eck}
|
||||
\label{fig:main}
|
||||
\input{plots/BSTPlot2.pgf}
|
||||
% \includegraphics[width=\textwidth]{plots/BSTPlot2.pgf}
|
||||
\label{fig:Bst2Results}
|
||||
\caption{Results for setup as seen in \figref{fig:BSTExp}. While the groundtruth distance only slightly varies (black line) the mean measured distance (blue line) varies greatly depending on the relative position to the fire door.}
|
||||
\end{figure}
|
||||
|
||||
|
||||
|
||||
\subsection{Positioning Environment}
|
||||
\begin{itemize}
|
||||
\item Beschreibe Gebäude - inkl HLS Räume!
|
||||
@@ -330,21 +380,8 @@ der einsatz der probablistischen methode sieht weitaus besser azs als multilater
|
||||
\caption{Path 2 Error FTM / RSSI, Pixel 2}
|
||||
\end{figure}
|
||||
|
||||
Unterscuhen ob die Abweichung nach untne daher kommt weil die Brandschutztür stört
|
||||
|
||||
|
||||
\begin{figure}[ht]
|
||||
\centering
|
||||
\includegraphics[width=1\textwidth]{VersuchsaufbauBST1.png}
|
||||
\caption{}
|
||||
\end{figure}
|
||||
|
||||
\begin{figure}[ht]
|
||||
\centering
|
||||
\includegraphics[width=1\textwidth]{VersuchsaufbauBST2.png}
|
||||
\caption{}
|
||||
\end{figure}
|
||||
|
||||
\subsection{Comparison and Discussion}
|
||||
Hier vergleichen wie sich die einzelen Verfahren untereinander unterscheiden und welche Vor / Nachteile sie haben. Jeweils für RSSI und FTM. (Vorher haben wir RSSI und FTM gegenübergestellt innerhalb der Verfahren und jetzt stellen wir die einzelnen Verfahren gegenüber und diskutieren deren Unterschied in Bezug auf die WI-Fi methoden)
|
||||
|
||||
|
||||
@@ -3751,3 +3751,11 @@ month={Sep.},}
|
||||
howpublished={\url{https://developer.android.com/guide/topics/connectivity/wifi-rtt}},
|
||||
year={2019}
|
||||
}
|
||||
|
||||
|
||||
@inproceedings{banin2016wifi,
|
||||
title={WiFi FTM and map information fusion for accurate positioning},
|
||||
author={Banin, Leor and Schatzberg, Uri and Amizur, Yuval},
|
||||
booktitle={2016 International Conference on Indoor Positioning and Indoor Navigation (IPIN)},
|
||||
year={2016}
|
||||
}
|
||||
20
tex/plots/BSTPlot1.pgf
Normal file
20
tex/plots/BSTPlot1.pgf
Normal file
@@ -0,0 +1,20 @@
|
||||
\begin{tikzpicture}
|
||||
\begin{axis}[
|
||||
scale only axis,
|
||||
xlabel ={Measurement point},
|
||||
ylabel ={Distance in \si{\meter}},
|
||||
legend pos=north east,
|
||||
xmajorgrids=true,
|
||||
%xminorgrids=true,
|
||||
ymajorgrids=true,
|
||||
%yminorgrids=true,
|
||||
enlarge x limits=false,
|
||||
%enlarge y limits=false,
|
||||
ymin=0,
|
||||
xtick={1,2,3,4,5,6,7},
|
||||
]
|
||||
\addplot [color=black] table[x=X,y=GT,col sep=comma]{plots/data/BSTMean1.csv}; \addlegendentry{Groundtruth}
|
||||
\addplot [color=blue] table[x=X,y=MeanDist,col sep=comma]{plots/data/BSTMean1.csv}; \addlegendentry{Mean distance}
|
||||
\addplot [color=orange, mark=x, only marks, mark size=1pt] table[x=X,y=Y,col sep=comma]{plots/data/BSTData1.csv}; \addlegendentry{Measured distance}
|
||||
\end{axis}
|
||||
\end{tikzpicture}
|
||||
19
tex/plots/BSTPlot1Rssi.pgf
Normal file
19
tex/plots/BSTPlot1Rssi.pgf
Normal file
@@ -0,0 +1,19 @@
|
||||
\begin{tikzpicture}
|
||||
\begin{axis}[
|
||||
ybar,
|
||||
scale only axis,
|
||||
xlabel ={Measurement point},
|
||||
ylabel ={RSSI in \si{\dBm}},
|
||||
legend pos=north east,
|
||||
xmajorgrids=true,
|
||||
%xminorgrids=true,
|
||||
ymajorgrids=true,
|
||||
%yminorgrids=true,
|
||||
%enlarge x limits=false,
|
||||
%enlarge y limits=false,
|
||||
%ymin=0,
|
||||
xtick={1,2,3,4,5,6,7},
|
||||
]
|
||||
\addplot [blue, fill=blue] table[x=X,y=RSSI,col sep=comma]{plots/data/BSTMean1.csv};
|
||||
\end{axis}
|
||||
\end{tikzpicture}
|
||||
21
tex/plots/BSTPlot2.pgf
Normal file
21
tex/plots/BSTPlot2.pgf
Normal file
@@ -0,0 +1,21 @@
|
||||
\begin{tikzpicture}
|
||||
\begin{axis}[
|
||||
scale only axis,
|
||||
width=0.8\textwidth,
|
||||
height=0.4\textwidth,
|
||||
xlabel ={Measurement point},
|
||||
ylabel ={Distance in \si{\meter}},
|
||||
legend pos=north east,
|
||||
xmajorgrids=true,
|
||||
%xminorgrids=true,
|
||||
ymajorgrids=true,
|
||||
%yminorgrids=true,
|
||||
enlarge x limits=false,
|
||||
%enlarge y limits=false,
|
||||
]
|
||||
%\draw ({axis cs:7,0}|-{rel axis cs:0,1}) -- ({axis cs:7,0}|-{rel axis cs:0,0});
|
||||
\addplot [color=black] table[x=X,y=GT,col sep=comma]{plots/data/BSTMean2.csv}; \addlegendentry{Groundtruth}
|
||||
\addplot [color=blue] table[x=X,y=MeanDist,col sep=comma]{plots/data/BSTMean2.csv}; \addlegendentry{Mean distance}
|
||||
\addplot [color=orange, mark=x, only marks, mark size=1pt] table[x=X,y=Y,col sep=comma]{plots/data/BSTData2.csv}; \addlegendentry{Measured distance}
|
||||
\end{axis}
|
||||
\end{tikzpicture}
|
||||
BIN
tex/plots/data/BSTData1.csv
(Stored with Git LFS)
Normal file
BIN
tex/plots/data/BSTData1.csv
(Stored with Git LFS)
Normal file
Binary file not shown.
|
BIN
tex/plots/data/BSTData2.csv
(Stored with Git LFS)
Normal file
BIN
tex/plots/data/BSTData2.csv
(Stored with Git LFS)
Normal file
Binary file not shown.
|
BIN
tex/plots/data/BSTMean1.csv
(Stored with Git LFS)
Normal file
BIN
tex/plots/data/BSTMean1.csv
(Stored with Git LFS)
Normal file
Binary file not shown.
|
BIN
tex/plots/data/BSTMean2.csv
(Stored with Git LFS)
Normal file
BIN
tex/plots/data/BSTMean2.csv
(Stored with Git LFS)
Normal file
Binary file not shown.
|
Reference in New Issue
Block a user