SLAM Evaluator

Evaluator to save and compare results from different SLAM algorithms/commits.

How does it work?

Important: the algorithms have to detect when the bag ends and end naturally in order to work.

Record a folder of rosbags

Launch the following command to record a folder with your rosbags: roslaunch slam_evaluator record_folder.launch bagfolder:="XXX" algorithm:="YYY" commit:="ZZZ" bags:="AA BB CC". In record_folder.launch you will see that you can also define the absolute paths for convinience and for each new algorithm, you have to define its topics in a YAML file.

Studying data

Run data/evalutate.py to generate CSV files containing difference with the ground truth data. Run data/compare.py [commit] to compare one commit's difference with ground truth with the others. Run data/plot_inter.py [algorithm] [bag1 bag2 bag3 ...] to plot the outputs of the algorithm vs. others and ground truth.