Competitive high variance, low volume manufacturing with robot manipulators

Abstract

Competitive robotized manufacturing of high specter variance, low volume product lines represents market opportunities for manufacturing companies, but cost-efficient production is challenging. In this paper, we present two main industry use cases which represent key challenges to be solved for cost-efficient low-volume, high-variance production. The use cases are found in collaboration with three manufacturing companies. We identify and describe these challenges which include perception and manipulation with shiny/high-reflectivity parts, human-machine interfaces for robot reconfiguration and calibration between simulated and real-world environments. In this paper, we present new methods for meeting these challenges: machine vision for handling sensor data with low quality in robot manipulation, automated robot programming based on CAD-models and automated calibration. Moreover, we implement and demonstrate the methods on the two identified industry use cases for robotized assembly.