One of the major challenges facing the industry today is the loss of knowledge as experienced operators leave their jobs. Demographic changes over the generations, combined with recruitment difficulties, are creating a so-called “inverted” age pyramid in the workforce. It is estimated that over the next few years, the number of new retirees will rise from 70,000 to well over 100,000.

Training the next generation of industrial workers is therefore a crucial issue, and faces a number of challenges: in-house training is costly in terms of time and resources, repetitive and highly unscalable. What’s more, operators’ knowledge is made up of diverse, unstructured information that is difficult to collect, standardize, search and, ultimately, exploit.

The aim of this project is to create a tool that can facilitate the non-intrusive collecting and analysis of operators’ experience, making it easily exploitable (known as “actionable knowledge”) for training the next generation of industry professionals.

Our research will focus on two aspects:

• A knowledge base adapted to collecting, managing and capitalizing on operators’ experience. These data, although in small quantities (small data), are varied, dynamic, variable and unstructured; they therefore have characteristics similar to those of “Big Data”. The innovative idea is to adapt Big Data approaches to aggregate this information and make it usable.
• An advanced user interface (Tutor) to record, retrieve and make easily accessible operational skills for using software and industrial machines. The Tutor will therefore have a dual role:

1. To help experienced operators annotate the operations they carry out in the field in a simple,  non-intrusive way, by recording them using eye-tracking glasses and a think-aloud process.
2. To make this information available to novice operators; this will enable them to follow the annotated operations step by step, or search for key words in the knowledge base.

The outcome of the project will be a proof-of-concept (POC) which will integrate the knowledge base and our Tutor into a knowledge management system. This will make it possible to collect the experience of operators (e.g. during initiation sessions for novices) and make it available for any future training in an incremental way (i.e. with the possibility of adding new content or enriching existing content).

The innovation potential of this research is very broad, and these same approaches and tools could help to solve the problems associated with “multi-shift manufacturing operations”. In other words, it could be used to improve the communication of errors, problems, best practices and processes between teams sharing the same tasks but on different shifts.