Problem
Research questions and friction points this paper is trying to address.
Robotic Insertion
Kinematic Feasibility
Object-in-Gripper Reorientation
Post-Grasp Repair
Inverse Kinematics
Post-Grasp Kinematic Repair for Robotic Insertion via Object-in-Gripper Reorientation
🤖 AI Summary
This study addresses the issue that object poses following a stable robotic grasp often lead to kinematic singularities or joint limits along insertion trajectories. To overcome this, we propose an in-gripper reorientation strategy that fine-tunes the object pose via tactile extrinsic manipulation to restore insertability. Furthermore, we construct a branch-aware inverse kinematics graph and integrate a task-conditioned learning prior to optimize search ordering, thereby enabling efficient motion planning. Experimental evaluations on a UR5e platform, conducted in both simulation and real-world settings, demonstrate that the proposed approach reduces the average reorientation angle by 51.5% compared to a grid-search baseline. Additionally, incorporating the learned ranking further decreases planning time by 51.1%, highlighting the effectiveness of our method for robust peg-in-hole assembly tasks.
📝 Abstract
A stable grasp does not guarantee kinematically feasible robotic insertion because the object-in-gripper transform may force the robot towards singularities or joint limits along the prescribed insertion path. We study post-grasp kinematic feasibility repair through object-in-gripper reorientation. Given an achieved grasp and a fixed insertion path, we seek a small reorientation that restores kinematic feasibility. Sequential IK can miss such candidates by following an unfavorable joint-space path, while the nonsmooth feasibility landscape makes the search computationally expensive. We evaluate candidates using a branch-aware IK graph that maximizes the minimum feasibility margin over the discretized insertion path and use a learned task-conditioned prior to improve query ordering. The selected reorientation is executed through tactile-based extrinsic manipulation. In UR5e simulations, the planner without learned ranking reduces mean reorientation over successful trials by 51.5% compared with grid-based sequential IK. Adding learned ranking reduces this planner's mean planning time by an additional 51.1%. Real-robot experiments validate the complete pipeline.
Innovation
Methods, ideas, or system contributions that make the work stand out.
post-grasp kinematic repair
object-in-gripper reorientation
branch-aware IK graph
task-conditioned prior
tactile-based extrinsic manipulation
🔎 Similar Papers
No similar papers found.
