Automated Robot Skill Parameterization Based On Task Specifications

Abstract

High-mix, low-volume manufacturing still relies on manual labor despite the potential for automation. This is due to the high cost of robot programming, which is required on a frequent basis. Skill-based frameworks for robot programming have reduced the expertise level and effort required for automation to meet the demands of high-mix, low-volume manufacturing. In these frameworks, the capabilities of robots are encapsulated in skills. By parameterizing these robot skills, they are adapted to new variants or products. This paper presents a concept for automating the parameterization process based on task specifications to reduce worker efforts by combining natural language processing and knowledge models.