Deep Reinforcement Learning for High Precision Assembly Tasks

TL;DR

This paper presents a reinforcement learning approach using recurrent neural networks to enable robots to perform high precision assembly tasks like peg-in-hole, reducing manual tuning and increasing robustness.

Contribution

It introduces a novel reinforcement learning method with neural networks for precise assembly, demonstrating improved robustness and reduced manual tuning compared to traditional methods.

Findings

Successful training of neural networks for peg-in-hole tasks

Robustness against position and angle errors demonstrated

Validated on a 7-axis articulated robot arm

Abstract

High precision assembly of mechanical parts requires accuracy exceeding the robot precision. Conventional part mating methods used in the current manufacturing requires tedious tuning of numerous parameters before deployment. We show how the robot can successfully perform a tight clearance peg-in-hole task through training a recurrent neural network with reinforcement learning. In addition to saving the manual effort, the proposed technique also shows robustness against position and angle errors for the peg-in-hole task. The neural network learns to take the optimal action by observing the robot sensors to estimate the system state. The advantages of our proposed method is validated experimentally on a 7-axis articulated robot arm.