An Advanced Human-inspired Compliant Peg-in-hole Solution for the Industrial Assembly of Almost Arbitrarily Shaped Lithium-ion Battery Packs

Lithium-ion battery packs based on the common 18065 cell vary in shape, size and other geometric parameters. If produced in small batch production, they are still mostly assembled manually. This paper completes our former publication concerning an advanced compliant human-inspired approach to solve the battery-based peg-in-hole problem, where the peg is represented by the battery and the holes being provided by the battery pack. While the original contribution focuses on bringing the peg to the hole if displacement errors of up to several millimeters occur, the here presented strategies are about centering and lowering the peg into the hole itself. The algorithms were implemented on an industrial robot plant, specially developed in cooperation with the German battery pack manufacturer ANSMANN AG. By only using the inbuilt coarse force torque sensors of a UR5e collaborative robot and geometric constraints, it can assemble a variety of different battery packs. The plant has been tested for several months on customer site, mounting thousands of batteries with a success rate of more than 99.5%.