Platform to test MRPP algorithms in SUMO via ROS

• SUMO

• ROS - Melodic

• Python Packages

  python3 -m pip install networkx configparser pandas seaborn numpy threading rospkg catkin_pkg
  

• Clone into your ros workspace

• Add the following to your .bashrc script

  export SUMO_HOME=/usr/share/sumo
  export PYTHONPATH=$SUMO_HOME/tools:$PYTHONPATH
  

• Ensure that the scripts are executable

  chmod -R 755 /path/to/mrpp_sumo/scripts/
  

• Enjoy!(?)

• Create a sumo compatible representation for the required environment with [graph_name].net.xml, [graph_name].nod.xml and [graph_name].edg.xml files and add it to the graph_sumo folder.

• Create a [graph_name].sumocfg file by copying an existing .sumocfg file and replacing the 'net-file' attribute value with [graph_name].net.xml

• Create a [graph_name].graphml file by running

  python3 scripts/sumo_to_graphml.py [graph_name]
  

  This file stores the graph information in a networkx compatible format.

Coming Soon

• Add a *.yaml file to the /config folder with the required rosparameters (here's a list of required parameters)

  • graph (name of the graph)
  • random_string (simulation id)
  • init_bots (number of bots)
  • use_sim_time: true
  • algo_name
  • sim_length (in secs)
  • init_locations: '' (if empty, locations picked at random)

• Add a *.graphml file in /graph_ml folder and a *.nod.xml, *.edg.xml, *.net.xml, *.sumocfg in /graph_sumo for the corresponding graphs

• Create an /outputs folder (if not already)

  roscd mrpp_sumo
  ./tpbp_new.sh

Find the following files in /outputs folder

• *_command.in Contains the sequence of walks sent by your algo_file. 4 lines per command (time_stamp, bot_name, node_id sequence, time of departure (can be empty))
• *_visits.in Contains the sequence of visits excuted by the bots. 3 lines per visit (time_stamp, node_ids, bot_ids)
• *_vehicle.xml Output from SUMO file

• Add scripts to generate plots for analysis
• Lots of other cool stuff