This package provides grasping functionality using RealSense Camera and custom grippers. The grasping stack is designed to be modular and can be used with different manipulators, grippers, and grasp pose detectors.
This stack has extensions to support following manipulators:
- UR Robots via CollaborativeRoboticsLab/grasping_ur
- TM Robots via CollaborativeRoboticsLab/grasping_tm
- Omron Mobile Manipulator Robots via CollaborativeRoboticsLab/grasping_omron_moma
And also supports grasping pose detectors like
In this grasping framework, we evaluate different gripper types. Due to this we focus on custom built grippers and our gripper controller revolves around different types for servos used to build the grippers and are availble in CollaborativeRoboticsLab/grippers. Instructions related to setup, configuration and customization are in the linked file.
In this grasping framework, we have tested with the UR10, TM12s manipulators and LD250 & TM12x mobile manipulator. Instructions related to setup, configuration and calibration are in the linked files.
- UR10 Manipulator
- TM12s Manipulator
- LD250 & TM12x Mobile Manipulator
- Attaching new gripper and components
- Moveit Servo and Keyboard Teleop
This component now separates three concerns:
- scene activation and active planning-scene ownership in
scene_manager_node - arm execution in
motion_execution_node - arm-only feasibility in
feasibility_service_node
The grasping runtime transforms grasp poses, applies workspace obstacles through the scene manager, visualizes the active workspace area, rejects poses outside that area, and exposes explicit feasibility services distinct from execution actions.
- Workspace Creation
- Scene Manager
- Workspace Loading
- Arm Control
- Control stack overview
- Simple Grasping
An external stack is typically used to detect or calculate a grasp pose, and this grasping stack is used to execute motion to that pose. One example of such an external system is listed below.