SIRRIS
Using Advanced AI Solutions to Achieve “First Time Right” and “Consistent Product Quality” Throughout the Product Development and Manufacturing Cycle

Human AI

Aim

The aim is to develop advanced decision support systems to guide engineers and operators during the whole design and manufacturing cycle so that a production with high quality standards is achieved quickly.

What's at stake

A major contribution to the performance of the manufacturing sector. Manufacturing Cycle Efficiency (MCE) is key in this respect. It measures the percentage of the time spent in manufacturing products that are devoted to value added activities. Non-value-added activities are those notably devoted to moving, inspecting, testing products. Decision support systems based on AI technologies would significantly contribute to reduce the time dedicated to these tasks, thereby increasing the efficiency of the whole cycle, sparing ressources especially in a context of mass production.

Challenges

A first challenge will consist in identifying all relevant sources of data. A second will consist in linking two different types of data: those related to the processing of the product and those used for its quality control in order to detect, diagnose and predict production defects related to process parameters or machine condition. A third challenge could stem from a more ambitious aim to be able to reuse the decision support system in other manufacturing contexts than the one where it was originally developed. In other words, to create a generic model with parameters that could be easily revised and hence adapted to other sectors.

AI possible solutions

A number of AI solutions will contribute to achieving the aim above while addressing the identified challenges:

  • The development of digital twins: to simulate different scenarios before a real production;
  • Active learning: to give the human operator the possibility to use his/her expertise to teach the model in case of divergence;
  • Supervised learning: to explain certain AI-inferred behaviours to the operator as best as possible so he/she can continue to play a supervising role.

    • These constitute an exciting challenge for our AI researchers.

    Key AI tech topics

    People

    Lorenzo Colantonio