Grasp candidates
Sample or analytic poses on faces, edges, or taught points relative to the part model. Clutter reduces the set of reachable surfaces.
Knowledge · Robotics
Grasp planning in 3D bin picking
After pose estimation finds the part, grasp planning chooses where and how the gripper contacts it, and whether the robot can reach that pose without collision.
Pipeline
From grasp candidates to executable motion
Grasp planning generates many candidate contact poses on the part surface, filters them by gripper kinematics and scene geometry, then ranks survivors for the robot controller.
Gripper constraints
Finger opening, stroke, suction cup diameter, vacuum zone, and payload limit which candidates remain valid.
Collision checks
Gripper body, fingers, and approach path tested against bin walls, neighboring parts, and fixtures.
Approach vectors
Pre-grasp pose offset along the tool axis; retreat path after pick to clear the bin rim.
Practice
What engineers specify early
- 01
Gripper CAD in the scene
Collision models must match the physical tool, including hoses and brackets.
- 02
Quality scoring rules
Prefer upright grasps, maximize contact area, penalize shallow approach angles into clutter.
- 03
Fallback ordering
When the top-ranked grasp fails IK or collision, the planner tries the next without a full re-scan when possible.
- 04
Tie to pose error budget
Grasp tolerance must absorb combined pose estimation and hand-eye error.
Grasp planning for your parts
Send gripper type, part geometry, and bin layout. We outline a feasible grasp and collision strategy.