this is for holding javascript data
Paul St-Aubin edited Methodology Measurements TTC.tex
almost 10 years ago
Commit id: 4e2dfb4625d11c7725cca2e0bf10c88958bdee79
deletions | additions
diff --git a/Methodology Measurements TTC.tex b/Methodology Measurements TTC.tex
index 011bfbe..a2160b9 100644
--- a/Methodology Measurements TTC.tex
+++ b/Methodology Measurements TTC.tex
...
Time-to-collision is measured instantaneously: a new value of TTC may be computed for every interaction-instant. Thus, a pair of users may have a time series of TTC observations evolving over time. Some efforts have been made to study these evolutions \cite{Mohamed_2013}. Other approaches have focused on quantile or threshold observations \cite{Svensson_2006}, while others still attempt to examine instantaneous risk and significance of TTC.
TTC distributions are generally gamma-like-shaped across the literature \cite{Ismail_2010} \cite{Autey_2012} \cite{St_Aubin_2013}. Quantifying collision risk based on TTC is the remaining puzzle piece. TTC thresholds have been popular, though are subject to arbitrary selection. One recent approach proposed a shifted gamma-generalised pareto distribution model \cite{Zheng_2014}. In the meantime, some qualitative analysis is possible in some circumstances, for example
when shifts in center of with a continuous mass
shift of
distributions are clear a probability distribution function as demonstrated in Figure~\ref{fig:distro-comparison}. This approach has been tried in some early applications of the methodology, e.g. in \cite{Ismail_2010}, \cite{Autey_2012}, and in \cite{St_Aubin_2013}.