A new approach to plan manual assembly

Abstract

Today’s methods for planning manual assembly processes have been developed for many decades. Besides technical advances, organisational innovations such as concurrent engineering have made a major contribution to both cost and quality of the final product. However, mass customisation and personalisation of products have imposed additional requirements. Planning should be performed for many different product variants while at the same time planning cost and time to market should be kept at a minimum level. The objective of this paper is to present a new approach for planning of manual assembly processes that goes beyond established industry practices, which has the potential to provide planning teams with tools to evaluate different assembly process scenarios faster, eliminating the need for physical prototypes. The approach is based on controlled natural language textual descriptions of assembly tasks, automatic generation human motions from the textual descriptions based on data-driven, statistical motion models and interactive process optimisation. The proposed method is evaluated in a pilot case from automotive industry regarding ease of use, extensibility, adaptability, motion realism and motion diversity. Results suggest that while the approach addresses practical needs of production planning, technical challenges remain in order to make it ready to be implemented into digital planning environments.