Intelligent Authoring and Management System for Assembly Instructions

Abstract

Continuously increasing complexity and variance within high variety low volume assembly systems causes a vast amount of work instructions. As the amount of new models and variants increases, the need of efficient generation of unambiguous instructions rises. Continuous instruction modifications are unavoidable due to design, customer or process changes.Case based research in cooperation with four manufacturing companies with manual assembly environments points out that assembly instructions authors currently are combining different authoring tools for creating and updating work instructions. Consequently, keeping the rising amount of work instructions up to date becomes less trivial. Furthermore, authors often create work instructions from scratch while instructions of product variants are mostly identical. This causes a large amount of similar work instructions stored as separate documents. As a result, the amount of inconsistent and outdated assembly instructions increases. Poor assembly instruction quality causes frustration and a lower performance of assembly operators.An automatic authoring system and intelligent operator feedback must eliminate these problems. The automatic authoring system provides the author with an overview of preprocessed information and related historical assembly instructions that can serve as a basis for the newly created instructions. In this way, the creation of instructions can be significantly accelerated and work instructions will become more consistent. An experimental lab setup is built in order to test the presented framework. Based on the first tests, the authoring process was significantly accelerated. Further tests within production environments are required in order to validate the presented framework.