FHWS/Fusion2016
Archived
3
0
Fork 0
Code Issues 1 Pull Requests Releases Wiki Activity

updated tex. sensors done.

Browse Source
This commit is contained in:
Toni
2016-02-12 15:40:52 +01:00
parent f0215731ce
commit 6b02336277
6 changed files with 40 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
tex/chapters/sensors.tex
View File

@@ -42,7 +42,7 @@ The evaluation following the transition then compares the predicted relative pre
\subsection{Wi-Fi \& iBeacons}
For additional absolute location hints, we use the smartphones Wi-Fi and iBeacon sensor to measure the signal-strengths
of nearby transmitters. As the positions of both \docAP{}s and and \docIBeacon{}s are known beforehand, we compare
of nearby transmitters. As the positions of both \docAP{}s and \docIBeacon{}s are known beforehand, we compare
each measurement with its corresponding signal strength prediction which is defined by the 3D distance $d$
and the number of floors $\Delta f$ between the \docAPshort{} and the particle
%
@@ -72,24 +72,12 @@ Therefore, a smaller $\mPLE$ can be chosen to model the signal strength predicti
\subsection{Step- \& Turn-Detection}
\commentByToni{da muessen wir nochmal drueber reden. das problem ist glaube ich. das die state transition vorher zu wenig gestreut hat und dadurch auch zu wenige particles in "hochwinkligen" bereichen hatte. dann haben die turns immer entsprechen einen delay gehabt. bin mir also nicht sicher ob es wirklich das downvoting / sample impoverishment ist. hast du da vielleicht noch ein paar infos für mich? weil unser sigma war ja immer rießig... quasi fast gleichverteilt.}
A big disadvantage of using the state transition as proposal distribution is the high possibility of sample impoverishment due to a small measurement noise. This happens since accurate observations result in high peaks of the evaluation density and therefore the proposal density is not able to sample outside that peak \cite{Isard98:CCD}. This causes a downvoting of particles with increased heading deviation. ...
To prevent degradation within the particle-filter \cite{??} due to downvoting of particles with increased heading deviation, we incorporate the step- and turn-detection within the transition step.
A big disadvantage of using the state transition as proposal distribution is the high possibility of sample impoverishment due to a small measurement noise.
This happens since accurate observations result in high peaks of the evaluation density and therefore the proposal density is not able to sample outside that peak \cite{Isard98:CCD}.
Additionally, erroneous or delayed measurements from absolute positioning sensors like \docWIFI{} may lead to misplaced turns.
This causes a downvoting of particles with increased heading deviation.
Therefore, we incorporate the turn-detection, as well as the related step-detection, directly into the state transition $p(\mStateVec_{t} \mid \mStateVec_{t-1}, \mObsVec_{t-1})$. This leads to a more directed sampling instead of a truly random one.
directly into the transition
$p(\mStateVec_{t} \mid \mStateVec_{t-1}, \mObsVec_{t-1})$.
\cite{thrun?}\cite{lukas2014?} to get a more directed sampling instead of a truly random one.
This happens since accurate observations result in high peaks of the evaluation density and therefore the importance density is not able to sample outside that peak [IB98b].
\commentByFrank{todo: wie wird die unsicherheit in der transition eingebracht, sigma, ..}