Skip to content

Latest commit

 

History

History
77 lines (67 loc) · 3.65 KB

README.md

File metadata and controls

77 lines (67 loc) · 3.65 KB

SkyScanner integration

Software integration for the SkyScanner project

Compilation

  1. Get ROS "Indigo" for your system following the instructions for Desktop-Full Install. (Newer versions should work but they haven't been tested) Don't forget to initialize rosdep and do the environment setup.
  2. Install the following python dependencies (not exhaustive list):
    • numpy
    • scipy
    • matplotlib
    • netCDF4 (MesoNH integration)
  3. Set up a SkyScanner ROS workspace:
    mkdir -p skyscanner_ws/src
    cd skyscanner_ws/src
    git clone https://github.com/rafael1193/skyscanner-integration.git
    cd ..
    echo "source $(pwd)/devel/setup.bash" >> ~/.bashrc
    source ~/.bashrc
    
  4. Build the ROS package:
    catkin_make
    
  5. Reload your bashrc and recompile again so ROS tools can detect the package properly:
    source ~/.bashrc
    catkin_make
    

Depending on the simulation backends you want to run you may have to install and configure some other dependencies:

Execution

Beforehand, run roscore in a dedicated terminal window and leave it running.

Launch files define the nodes and parameters that have to be run to perform different tasks.

  • Paparazzi
    • Select your aircraft, nps target and build
    • Launch a Simulation session adding -f127.0.0.1 -P50501 to the Simulator command line.-P should correspond with the port defined in launch/conf/pprz_ac*.yaml files and must be unique if several simulators are launched.
    • Run roslaunch skyscanner pprz_planned.launch
    • To command several paparazzi aircrafts go to the pprz_planned.launch file
  • FlightGear
    • Start FlightGear with the fly_malolo1.sh script. You can change the default UDP, Telnet and HTTP ports inside.
    • Run roslaunch skyscanner fg_planned.launch if you want to launch the whole architecture.
    • Run roslaunch skyscanner fg_guided.launch to start everything but the pathplanner so you can send tasks manually to the guidance node (by default it uses the VF alogrithm as guidance, but PLOS can be se too). In the following example rostopic is used to send to ac_1 a TrajectorySequence composed of a circle of 200m radius and origin at (0, 0):
    rostopic pub /ac_1/trajectory_sequence skyscanner/TrajectorySequence "header:
     seq: 0
     stamp:
       secs: 0
       nsecs: 0
     frame_id: ''
    ac_id: 0
    trajectories:
    - circle: true
     origin: {x: 0.0, y: 0.0, z: 0.0}
     destination: {x: 0.0, y: 0.0, z: 0.0}
     radius: 200.0
     duration: 100.0
     time_limit: 10000000000.0" -1
    
  • Additionally roslaunch skyscanner supervision.launch will show online stats using rqt_plot and rqt_plotxy (make sure you have them).
  • MesoNH
    • By default realistic MesoNH wind is deactivated as you have to declare the path where the NetCDF files are stored in the parameter mesonh_files_path located in launch/conf/mesonh.yaml.
    • If wind data is available you can activate it switching use_dummy_env to False in launch/conf/mesonh.yaml.

Other configuration parameters can be tuned in .yaml files or directly in the roslaunch files you are running.