% --------------------------------------
% CHAPTER: Related Work
% --------------------------------------
\section{Related Work}
\label{sec:Related Work}


It is not possible to navigate the satellite since it has no rocket propulsion built in.
Therefore it is also not possible to aim the camera to a certain object.
This circumstance makes it quite hard to take pictures.

So the first consideration is just
to take any pictures and send them from the satellite down to the base stations on earth.
But this method is not very feasible for some reasons. First we need to take a lot of pictures
to get at least some good ones. This would need a lot of memory on the satellite, which is not 
such a big deal, but the writing of data needs electric energy which is quite limited on the 
satellite. The second problem is, that we also would need to transfer a big amount of data.
Data transfer is limited through transfer rate and also through the time windows that allow
communication with the satellite.

It is quite obvious that there is a need of preselection of the pictures which shall be
downloaded from the satellite. So we thought about a simple algorithm which could be able
to do some evaluation if a picture wether it is 'interesting' or not. Totally black and 
totally white pictures can be sorted out. Other pictures could show something intresting.
The satellite travels on a n orbit which is about 318 km above the earths surface. Therefore
the dihedral angle of earth in view is about \textbf{TOTO: check raumwinkel} which is rather 
big compared to the one which is showing the sun. This means, that the likelyhood that a
picture, which is not only black or white, is rather big that it shows some parts of the earth.

In order to raise the likelyhood of shooting a picture at the right moment, we were searching
for a way of how to find this right moment.
Using the magnetic field of the earth is not an appropriate approach since the magnetic force 
lines of the earth run approximately parallel to the earth surface (except of near the poles),
which would not provide enough information to calculate the satellites orientation in a rather
easy way. Furthermore, a teslameter that provides values of a usable accuracy would be to 
expensive as well as too damageable.

Finally we decided to use a 3D - light sensor. It shall help to find out the relative 
orientation of the satellite to the earth. One photodiode on each side of a cube would provide 
6 lines of different light intensities. When the satellite is spinning through space, the 
photodiodes will give periodical signals, such way that three different periodes will be seen.
The satellite will rotate around three axis, whereby each of these three periods corresponds to the rotation around one of these axes.

\begin{figure}[h]  % [top, bottom, h ... where defined, p ... own page]
\centering 
\includegraphics[keepaspectratio,width=\hsize]
%{drawings/Crc_Check.pdf}
{figures/ligth_sens_1_ggb.pdf}
\caption{Periodic output of 3 photodiodes.}
\label{fig:lightsens1}
\end{figure}

In Figure~\ref{fig:lightsens1} we can see, that these periodes have one common periode,
which is starting at t0. The point im time tp (here just an example) can the be taken as
the point in time which is perfect for taking a picture. So the idea is to cumulate the
photosensors measurements and download them to earth. After analysing and calculating a
certain point in time tp, the satellite can be asked to take pictures at tp. If these
pictures are not satisfying, tp can be tuned until the results fit to the expected outcome.