Modern manufacturing is becoming increasingly data-driven, due to the ever-increasing amounts of data generated by IIoT devices and constant improvements in computing power required to process these vast volumes of data. Data generated through manufacturing activities are utilised to enhance manufacturing quality and thus enrich the flexibility and autonomy of the system. Data is used to develop data-driven models, that are powered by AI and machine learning algorithms, to make smart decisions governed by the data. Most Tier 1 and large OEM companies have in-house data-driven modelling capabilities. The challenge is to encourage mid-level companies and SMEs to adopt the state-of-the-art IIoT, connectivity and data-driven modelling strategies that would enable them to fully exploit the data generated during the manufacturing process to improve their productivity and business gains.
Data-driven, or autonomous, manufacturing systems, are capable of identifying solutions inconceivable to humans and thus open up new routes to improve productivity, resource efficiency and business gains. However, a lack of interpretability presents questions about trust and accountability. To overcome these barriers will require the development of new methods that make use of emerging data visualization technologies, such as Augmented and Virtual reality, to illuminate the data and decision-making process. Furthermore, the functioning of certain manufacturing processes is deeply reliant on operators’ experience. How we can incorporate that human knowledge into these data-driven models is currently unclear but may prove fundamental for the digitisation of certain manufacturing industries.