A Machine Learning Approach for Collaborative Robot Smart Manufacturing Inspection for Quality Control Systems

TLDR

The 4th industrial revolution drives automatic, zero‑defect product inspection, with collaborative robotics combining robot accuracy and human flexibility. The study proposes using a collaborative robot paired with a learning intelligent system to perform smart inspection and corrective actions in manufacturing quality control. Reinforcement learning enables the robot to adapt its inspection trajectory, with the UR3 robot equipped with a force‑torque sensor. Preliminary experiments trained a UR3 robot with a force‑torque sensor to perform a quality inspection path, demonstrating the feasibility of the approach.

Abstract

The 4th industrial revolution promotes the automatic inspection of all products towards a zero-defect and high-quality manufacturing. In this context, collaborative robotics, where humans and machines share the same space, comprises a suitable approach that allows combining the accuracy of a robot and the ability and flexibility of a human. This paper describes an innovative approach that uses a collaborative robot to support the smart inspection and corrective actions for quality control systems in the manufacturing process, complemented by an intelligent system that learns and adapts its behavior according to the inspected parts. This intelligent system that implements the reinforcement learning algorithm makes the approach more robust once it can learn and be adapted to the trajectory. In the preliminary experiments, it was used a UR3 robot equipped with a Force-Torque sensor that was trained to perform a path regarding a product quality inspection task.