Traffic Simulation for the paper "Bayesian dynamic linear model with adaptive parameter estimation for short-term travel speed prediction" Tai-Yu Ma, and Yoann Pigné.

We use version 0.26 as it is compatible with the LUST dataset.

Download the archive from sourceforge as the svn Checkout of tag v0_26_0 fails.

export CPPFLAGS="$CPPFLAGS -I/usr/local/include -I/opt/X11/include"
export LDFLAGS="$LDFLAGS -L/usr/local/lib -L/opt/X11/lib"

./configure CXX=clang++ CXXFLAGS="-stdlib=libc++ -std=gnu++11" --with-xerces=/usr/local --with-proj-gdal=/usr/local --with-fox-config=/usr/local/bin/fox-config -with-proj-gdal=/usr/local --with-xerces=/usr/local --prefix=/opt/sumo
make -j8

We use the Luxembourg SUMO Traffic (LuST) Scenario as a base scenario for mobility scenario.

git clone https://github.com/lcodeca/LuSTScenario.git

Then we artificially bloc road segments in order to simulate unpredicted traffic hazards.

Output data is generated from a set of detectors materialized in this map :

• https://www.scribblemaps.com/maps/view/Luxembourg_traffic/7tdzwHfuLz

Detectors are configured in the mydetectors.xml file in each scenario folder.

http://sumo.dlr.de/wiki/Simulation/Output/Induction_Loops_Detectors_(E1)

Data are recorded between 7AM (25200s) and 9AM (32400s)

There are two scenarii. One with normal traffic. One where an accident occurs at 7.30AM and blocs the road for 30 minutes.

Finally, results are presented in files :

• scenarios/results/lust.normal-taffic/lust.normal-traffic.out.xlsx
• scenarios/results/lust-accident-pt-duchesse/lust.accident-pt-duchesse.out.xlsx

These table contains lane-by-lane speed time series on each detector defined earlier. This data is used as an input for the estimation method.