Motion planning for peg-in-hole insertion

Establish motion planning methods that can generate executable robot trajectories for peg-in-hole insertion tasks in industrial automation, particularly under inevitable grasp pose errors and limited joint motor precision, in order to enable reliable insertion without relying on fixed fixtures or highly constrained initial configurations.

Background

Peg-in-hole assembly is a foundational capability for industrial robotic automation but remains challenging due to the diversity of object shapes and connection mechanisms. Traditional approaches often assume fixed holes and rigidly held pegs, limiting practical deployment.

The paper emphasizes that motion planning for insertion has been less studied compared to pick-and-place and is complicated by grasp pose uncertainties and robot joint precision limitations, which can cause planned trajectories to fail at the insertion stage. The authors propose integrating combined task and motion planning with compliant control as a step toward addressing this challenge, highlighting the broader open problem of planning motions for peg-in-hole tasks.