This project creates an autonomous Jackal robot that can avoid moving obstacles. Using this algorithm, this repository provides code for 3 scenarios:
- Navigating around a highway with many moving obstacles
- Racing against a manually controlled robot in a highway
- Racing against a manually controlled robot in a racetrack
Need ROS Kinetic, Gazebo, catkin_pkg, and ros keyboard
Setup the catkin_ws folder (needs to be done every time the code in catkin_ws is changed)
cd catkin_ws
catkin_make
Setup the jackal_ws folder (needs to be done every time the code in jackal_ws is changed)
cd ../jackal_ws
catkin_make
Source the setup file
cd ../
source catkin_ws/devel/setup.bash
source jackal_ws/devel/setup.bash
The source commands need to be done for every new terminal. To avoid this, .bashrc can be modified to include these source commands
Launch the launch file
roslaunch multi_jackal_tutorials jackal_highway_long_6obs.launch
Open Gazebo
gzclient
Run the code
rosrun jackal_nodes particle_planner_highway
This runs the highway scenario with 6 obstacles. To test the scenario with 12 obstacles, run roslaunch multi_jackal_tutorials jackal_highway_long_12obs.launch instead.
Additionally, edit particle_planner_highway.cpp line 65 to be 12. Also uncomment lines 79 and 80 while commenting lines 82 and 83. After code changes, catkin_make is needed.
Launch the launch file
roslaunch multi_jackal_tutorials jackal_highway_long_6obs.launch
Open Gazebo
gzclient
Run the code
rosrun keyboard keyboard
rosrun jackal_nodes jackal_teleop_racing
rosrun jackal_nodes particle_planner_highway_race
Launch the launch file
roslaunch multi_jackal_tutorials jackal_highway_long_6obs.launch
Open Gazebo
gzclient
Run the code
rosrun keyboard keyboard
rosrun jackal_nodes jackal_teleop_racing
rosrun jackal_nodes particle_planner_racetrack
planner_visualization_v2.m is a matlab visualization that shows the trajectories that the planner examines during the highway scenario. This visualizer only looks at the very first plan that is called when all the objects are at there starting positions. It uses recorded ros bag files.
To record another ros scenario, edit particle_planner_highway.cpp line 652 to be if(true) to stop the obstacles from physically moving. (This may not be needed, but this is what was done when testing the visualizer)
Then catkin_make and run the commands for scenario 1. Then go into the bagfiles folder and run
rosbag -a
After a second, all the needed data will be recorded and the newly created bag file can be used by the visualizer.
The list of waypoints was created by running Scenario 3 and running jackal_teleop_racing. The manually controlled car was driven from the starting point around the track and its positions were recorded in a file. To do this, modify jackal_teleop_racing by uncommenting lines 45, 55, and 93. Then run drive the manual controlled car around the racetrack. A new text file will be made in the catkin_ws folder. Use plot_track.py to visualize the waypoints created.
jackal_ws was taken from https://github.com/NicksSimulationsROS/multi_jackal
Email [email protected] for any questions or corrections needed