Perception Pipeline

Goal: Fast (~30Hz) perception pipline for generating control parameters and workspace approximation for multi-agent cooperation (PMAF, Laha et al 2023) in tight-constrained scenes.

Generate Data

cd $PERCEPT_ROOT
source peract_env/bin/activate
cd src
scripts/run_datacap.sh

Run Pipeline

cd $PERCEPT_ROOT
source peract_env/bin/activate
cd src
python src/pipeline/3cam_pipe.py

Run PMAF

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$COPPELIASIM_ROOT
export QT_QPA_PLATFORM_PLUGIN_PATH=$COPPELIASIM_ROOT
cd $COPPELIASIM_ROOT && ./coppeliaSim.sh -h $PMAF_ROOT/src/bimanual_planning_ros/vrep_scenes/dual_arms.ttt

roslaunch bimanual_planning_ros vrep_interface_dual_arms.launch task_sequence:=dual_arms_static3
roslaunch bimanual_planning_ros planning_moveit_dual_arms.launch

Debug

Run Coppelia

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$COPPELIASIM_ROOT
export QT_QPA_PLATFORM_PLUGIN_PATH=$COPPELIASIM_ROOT
cd $COPPELIASIM_ROOT && ./coppeliaSim.sh

Credits

Collaborators

• Riddhiman Laha
• Tinayu Ren

Projects

• CuPoch
• Open3D
• CoppeliaSim