A Design Science Research Approach to Architecting and Developing Information Systems for Collaborative Manufacturing: A Case for Human-Robot Collaboration

Information generated from the conceptualization, design, manufacturing, and use of a product has immense potential in transforming both the business and manufacturing processes of the manufacturing enterprise. The digital transformation at the heart of the fourth industrial revolution has acknowledged this with a special emphasis on weaving a thread of this information to support functions and systems throughout the life cycle of the product with what is known as a digital thread framework. This dissertation aims to develop and use one such framework in the context of human-robot collaborative assembly. The overarching problem that the framework aims to solve can be attributed to the abstract qualities of adaptability and flexibility. The human-robot collaboration (HRC) systems of today are built predominantly as static systems and ignore the intuitive role of humans by having their roles in collaborative tasks pre-defined. Furthermore, their ability to switch between products during product changeovers is also limited. This is especially problematic in the current era of product variety, stemming from the customised requirements of customers.

To this end, this dissertation employs the design science research methodology to design, develop, and deploy predominantly three artefacts in a human-robot work cell in a laboratory setting. The first is the digital thread framework that integrates the product design environment using state-of-the-art knowledge-based engineering systems, as an agent of a multi-agent system, which provide the collaborative human-robot agents with access to product design models at run time. The second is a constituent mixed-reality model that provides an interface for the foregoing framework for the human operator engaged in collaborative assembly. The third is a supporting information model that the agents use as their knowledge base to fulfil adaptively the goals of collaborative assembly.

Together, these developed artefacts were employed in case studies involving a real diesel engine assembly during which they were observed to provide utility and support the cause of adaptability for which the framework was designed. The identification of bounding boxes as a scalable information construct, that approximates the part geometry of the sub-assembly components, demonstrates the utility of the developed artefacts for spatially augmenting them as projections as intentions of collaborating agents.

In summary, this dissertation contributes with an approach towards realising intelligent and adaptive robotics within the realms of information flows and modelling in the context of human-robot collaboration. The lack of intelligently adaptable HRC systems reported by the industry in part motivated the work undertaken in this dissertation. As future products and production systems become more complex, information systems are expected to assume greater responsibility to compensate for the inherent limits of the human working memory and enable transition towards a human-centred manufacturing, the current likes of which are labelled as Operator 4.0 and Industry 5.0. Thus, the expectation is that information systems research, such as this dissertation, can help take significant strides forward in this direction.