Ausarbeitung/experiment2.tex

   1 \chapter{Experiment 2: Movement Behaviour with Mean Correction}
   2 \todo{}
   3
   4 As presumed in Section \ref{exp1:results}, errors in the Roomba's movements
   5 could originate from imprecise measurement of the Roomba's internal sensors or
   6 in the Wiselib implementation. So a natural approach to correct this sort of
   7 errors would be to average the results for each data point from Experiment 1,
   8 find a function that fits the mean measured error depending of the
   9 target velocity and target distance or angle as well as possible, and then
  10 adapting either one of the target parameters so that the resulting movement
  11 would most likely be the desired target value. In this experiment however, only
  12 the target distance resp. the target angle was adjusted, while the velocity
  13 remained unadjusted.
  14
  15 Fitting the function\index{fit function} was done with \acs{GNU} R\index{GNU R}
  16 through a wrapper script which is explained in section~\ref{sec:impl:eval}. In
  17 this experiment, a linear fit of the form $o = a*v+b*i+c$ was used, with $o$
  18 being the measured value, $v$ the input velocity, $i$ the target distance or
  19 angle, and $a,b,c \in \mathbb{R}$. The fitted values \todo{how? least
  20 square?} for $a, b, c$ were then used in the algorithm to calculate the adapted
  21 target distance or angle.
  22
  23 \section{Setup}
  24 The hardware setup was exactly the same as in Experiment 1. However, in this
  25 experiment the application \cmd{mean\_correction\_test} was used to measure
  26 data. It did exactly the same as the application from Experiment 1, except that
  27 it adapted the target distance resp. target angle according to the algorithm
  28 described above.
  29
  30 \section{Results}
  31 \begin{figure}[p!]
  32  \centering
  33  \includegraphics[width=\textwidth]{images/iz250flur_drive-mean_data.pdf}
  34  \caption{Behaviour with mean correction on laminated floor, straight drive
  35 tests}
  36 \end{figure}
  37 \begin{figure}[p!]
  38  \centering
  39  \includegraphics[width=\textwidth]{images/iz250flur_turn-mean_data.pdf}
  40  \caption{Behaviour with mean correction on laminated floor, turn tests}
  41 \end{figure}
  42 \begin{figure}[p!]
  43  \centering
  44  \includegraphics[width=\textwidth]{images/seminarraum_drive-mean_data.pdf}
  45  \caption{Behaviour with mean correction on carpet floor, straight drive
  46 tests}
  47 \end{figure}
  48 \begin{figure}[p!]
  49  \centering
  50  \includegraphics[width=\textwidth]{images/seminarraum_drive-mean_data.pdf}
  51  \caption{Behaviour with mean correction on carpet floor, turn tests}
  52 \end{figure}
  53
  54 results better than in experiment 1, very accurate for laminate floor, carpet
  55 floor more spread but still kind of in the middle and less deviation from ideal
  56 value.
  57 \todo{statistical values, stddev?}