Colaborative Transportation Planning

Preparing

• Install OMPL (tested with 1.6)

• Install Crocodyll

  sudo tee /etc/apt/sources.list.d/robotpkg.list << EOF
  deb [arch=amd64] http://robotpkg.openrobots.org/packages/debian/pub $(lsb_release -sc) robotpkg
  EOF
  
  curl http://robotpkg.openrobots.org/packages/debian/robotpkg.key | sudo apt-key add -
  sudo apt update
  sudo apt install robotpkg-py310-crocoddyl -y
  

Building

Tested on Ubuntu 22.04.

mkdir build
cd build
cmake .. -DCMAKE_PREFIX_PATH="/opt/openrobots/"
make -j

Running

cd examples/2cfs_pointmass
mkdir stats/
touch ompl_stats.yaml
cd ../../build
./nUavsPayloadPlanner --input ../examples/2cfs_pointmass/2cfs_pointmass.yaml --output output.yaml --stats ../examples/2cfs_pointmass/stats/ompl_stats.yaml

Scenarios

• you can run it using the commands above or for maze, forest and empty:

   cd tests/
   ./maze 0 3
   ./forest 3 4
   ./empty 5 8

• The first argument is the experiment number (you can have multiple experiments to compare).
• The second argument is the number of uavs. Now we support from 3 uavs to 8.
• If you want to run any of the other tests you do not need any arguments (i.e., it is only needed for the scenarios):

  cd tests/
  ./2cfs_payload
  ./3cfs_payload
  ./3cfs_pointmass

Visualize

python3 ../scripts/visualize.py output.yaml pointmass

• For different payloads: currently we support: rod, cuboid, triangle and pointmass.

Shell scripts to run examples and visualize with one command

cd tests/
./2cfs_pointmass.sh

Plot

 python3 ../scripts/plot.py output.yaml

Compute Statistics

cd scripts/
python3 compute_stats.py

• To change configuration, check scripts/compute_stats.yaml. Three scenarios (the configs are in examples folder) are provided: empty, maze and forest. All of them have the same config parameters (i.e., runs, rangeofRobots: which number of robots to include) except forest, has some extra ones like the number of Obstacles, [min_env, max_env]: the minimum and maximum space where the random obstacles can be.
• The stats for the latest run is found inside exampls/forest/stats
• The output of each run to visualize can be found inside examples/forest/stats/run3
• The script generates a dict output: stadtdict.yaml.
• For the forest example, we provide a random generation obstacles script `compute_obs.py, that generates and visualizes obstacles.yaml.
• In case the obstacles.yaml is a good example rename the file to obstacles_final.yaml.

cp -a obstacles.yaml obstacles_final.yaml

• Otherwise, the compute_stats.py will complain about not finding the obstacles_final.yaml
• To visualize any result, choose scenario and which run and which number of robots (i.e., output6.yaml means for 6 robots, output3.yaml, 3 robots and so on):

cd scripts
python3 visualize.py ../examples/maze/run4/output6.yaml triangle

Dynoplan Notes

deps/dynoplan/main_optimization --env_file ../deps/dynoplan/dynobench/envs/unicycle2_v0/parallelpark_0.yaml --models_base_path ../deps/dynoplan/dynobench/models/ --init_file ../deps/dynoplan/dynobench/data/unicycle2_0_parallelark_guess_0.yaml --results_file opt.yaml

python3 ../deps/dynoplan/dynobench/utils/viewer/viewer_cli.py --robot unicycle1 --env ../deps/dynoplan/dynobench/envs/unicycle2_v0/parallelpark_0.yaml --result opt.yaml.trajopt.yaml -i

python3 ../deps/dynoplan/dynobench/utils/viewer/viewer_cli.py --robot acrobot --env ../deps/dynoplan/dynobench/envs/acrobot_v0/swing_up_empty.yaml --result ../deps/dynoplan/dynobench/envs/acrobot_v0/trajectories/swing_up_empty_init_guess.yaml -i

quad2dpole

deps/dynoplan/main_optimization --env_file ../deps/dynoplan/dynobench/envs/quad2dpole_v0/test.yaml --models_base_path ../deps/dynoplan/dynobench/models/ --init_file ../deps/dynoplan/dynobench/envs/quad2dpole_v0/down/idbastar_v0_db_solution_v0.yaml --results_file opt.yaml

python3 ../deps/dynoplan/dynobench/utils/viewer/viewer_cli.py --robot quad2dpole --env ../deps/dynoplan/dynobench/envs/quad2dpole_v0/test.yaml --result opt.yaml.trajopt.yaml -i

python3 ../deps/dynoplan/dynobench/utils/viewer/viewer_cli.py --robot quad2dpole --env ../deps/dynoplan/dynobench/envs/quad2dpole_v0/test.yaml --result ../deps/dynoplan/dynobench/envs/quad2dpole_v0/down/idbastar_v0_db_solution_v0.yaml -i