This project uses an Artificial Potential Field Algorithm in order to find a path around obstacles and towards a goal. The algorithm utilizes ROS and the simulation environment Gazebo.
Install PX4, Gazebo and ROS with this tutorial: https://docs.px4.io/master/en/simulation/ros_interface.html
ROS Melodic is needed, other versions were not tested.
Move the entire folder of the repo into ~/catkin_ws/src. Then run "catkin build" from ~/catkin_ws.
There are already a few example launch files in the folder ./launch. You can launch them with
roslaunch uav_path_planning path_planner_01.launch
or some other launch file of your choice.
Refer to the ADP-Documentation for detailed information on the ROS-communication-network and how the nodes work with each other. Alternatively, you can use the command rqt_graph to create a graph of the communication of nodes with each other.
uav_path_planning/
├── benchmarks/
│ ├── scenarios/
│ │ ├── 2d # Contains the 2D-worlds and launch files used for evaluation
│ │ └── 3d # Contains the 3D-worlds and launch files used for evaluation
│ └── classification_benchmark.csv # Contains the risik index for different obstacle types; used by the benchmark launch files
├── launch/ # Start various gazebo maps with our path planner
├── msg/
│ ├── obstacleListMsg.msg # Message type used by the obstacle_sensor and the obstacle_map nodes
│ └── obstacleMsg.msg # Message type used by the obstacle_sensor and the obstacle_map nodes
├── nodes/ # This folder contains python files that can be used as a shortcut to start the nodes from the python package in /src
│ ├── global_path_node
│ ├── local_path_planner_node
│ ├── obstacle_map_node
│ ├── obstacle_sensor_node
│ ├── path_logger_node
│ └── setup_path_logger_for_px4_node
├── src/ # Contains the python source code
│ └── py_uav_path_planning/ # The python package
│ ├── extras/
│ │ ├── evaluate_log_file.py # Plots the logs file
│ │ ├── path_logger.py # Logs the UAV position and velocity and the global waypoints for every time step
│ │ └── setup_path_logger_for_px4_avoidance.py # Kills the offb_node and starts the path_logger
│ ├── obstacle_map/
│ │ └── obstacle_map.py # Starts the obstacle map
│ ├── obstacle_sensor
│ │ ├── obstacle_sensor.py # Starts the obstacle sensor
│ │ └── helper
│ │ └── read_gazebo_xml.py # Reads all obstacles in a gazebo world file
│ └── path_planning/
│ ├── calc_apf.py # Service node for the potential field and gradient field
│ ├── global_path.py # Node that forwards waypoints from local planner and publishes global waypoints from a file
│ └── local_path_planning.py # Local planner that uses calc_apf for finding movement direction
└── srv/
└── potential_field_msg.srv # Service type used by calc_apf and local_path_planning
Leonard Smits
Daniel Piendl
