Can you recommend a grasp planning software for object sorting on a multi-DOF robot arm?

Yes — for object sorting with a multi-DOF robot arm, I’d recommend looking at a grasp planning stack rather than a single “grasp planner”, because you usually need:

• grasp pose generation
• collision checking
• reachability / inverse kinematics
• motion planning
• execution integration with perception

Strong software options

1. MoveIt 2 (best general-purpose choice)

• Why: Widely used in ROS 2 for robotic manipulation.
• Includes:
  • motion planning
  • collision checking
  • IK integration
  • grasp/pose handling
• Good if your sorting system uses ROS/ROS 2.
• Pair it with:
  • MoveIt Task Constructor for pick-and-place workflows
  • a grasp generation method or perception pipeline

Best for: production-like industrial sorting and research systems.

---

2. GraspNet / Contact-GraspNet (best for vision-based grasp generation)

• Why: Excellent if you want to generate grasps from point clouds or RGB-D data.
• Useful for:
  • cluttered bin picking
  • object sorting by visual perception
• Usually combined with a motion planner like MoveIt.

Best for: perception-driven grasp selection in cluttered scenes.

---

3. OpenRAVE + grasping plugins (older, but useful in some research setups)

• Why: Good for grasp analysis and reachability studies.
• Less modern than MoveIt, and community support is weaker now.

Best for: legacy or academic environments.

---

4. PyBullet or CoppeliaSim

• Why: Good for simulation and prototyping sorting behavior.
• Both support:
  • robot models
  • grasp testing
  • scene simulation
• You can test grasp strategies before deploying to hardware.

Best for: rapid prototyping and sim validation.

---

My practical recommendation

If you want a robust setup for a multi-DOF arm sorting objects:

Use:

• MoveIt 2 for planning/execution
• Contact-GraspNet or GraspNet for candidate grasp generation
• ROS 2 for integration
• optional MoveIt Task Constructor for full pick-sort-place pipeline

This combination is very common and flexible.

---

If you want the simplest answer:

• ROS 2 + MoveIt 2 if you need a dependable manipulation framework
• Add GraspNet/Contact-GraspNet if you need vision-based grasp proposals

If you tell me:

1. your robot arm model,
2. whether you use ROS/ROS 2,
3. whether the objects are in bins or on a conveyor,
4. and whether you have RGB-D or just a camera,

I can recommend a more specific software stack.