active_vision_grasping

To Grasp

STEPS:

• setting up environment - all

• point cloud library learning - all

• stitching the point cloud - do we really need it? (bonus)

• segementation - Janie & Hasitha

• get normals - Ramana & Tript

• spawn objects - Laddu & Saurabh

  • sphere on a table
  • cuboid on a table
  • glass vase on a table

• grasp calculation -

  • find optimal grasp
  • moving the camera
  • probably do it in another node

• move camera optimally (optional)

• move robot to grasp (optional)

• filtering - (optional needed only in real world scenarios)

• downsampling - (optional needed only in real world scenarios)

Steps to run environment

1. cd $workspace_dir$
2. colcon build --packages-selct segmentation
3. ros2 launch segmentation simulation_can.launch.py
4. Publishing joint position via topic

ros2 topic pub /forward_position_controller/commands std_msgs/msg/Float64MultiArray "data:
- 0.0