\section{Using odometry to identify urban canyons} \label{sec:meth}

\subsection{A basic understanding of GPS} \label{sec:}
GPS is a common way to provide global geolocation to eg. a vehicle, using satelites. It uses a lowfrequency signal recieved from a series of satelites which consists of among other things their position, along with the timestamp from their extremely accurate atom clocks. Using this timestamp one can measure the time it has taken the series of signals to reach you and from that calculate their distance to you. Using a minimum of four different satellite signals like these that have a clear line of sight to the vehicle in question, one can then solve for its 3D position on the globe. However this calculation has many complications, among other things the effects of relativity, as well as the signal bending(?) when penetrating the athmosphere. Due to all of the complications like these, all GPS use several satellite signals along with Kalman filters and optimization algorithms to remove noise and bias. One of these problems that are harder to simply correct are what is called urban canyons.

\subsection{Urban canyons} \label{sec:}
GPS can not be used indoors as the low frequency signals do not penetrate walls, and objects like it. When you get a GPS signal when indoors, it is more likely due to the signal bouncing around until you recieve it. This would off course take the signal longer time than for a signal taking the straight route. The straight route is what is taken into account when calculating its distance, making a bouncing GPS signal not one to be trusted. In other words, GPS wont be as accurate indoors. However this effect is also met when simply being around other objects that can disrupt the measurers field of view from the satelite - which is common in urban environments. Buildings in cities create what are called urban canyons, that create a hard time for a GPS to position itself.

\subsection{Choosing satelites of a certain angle to you} \label{sec:}
GPS commonly allready disregard satelites that have an angle below 15 degrees to you, because these would have a very long time travelling through the atmosphere, making the signal take longer and from that result in a less precise result. The train of thought here is that there are more satelites to choose from, given that we know their position we can use the to choose the satellites that give us the best result.