The project, funded by the Government of Navarra, is working on a software platform designed to reduce errors in production processes.

The AI system will continuously learn from feedback from people, who will be able to enter different types of information such as images, numbers or text in natural language.

 

ZEO Technology, the NAITEC Technology Centre, BORG Automotive Group and Bacaicoa Industrias Plásticas have launched the HUCAI-S4 project, which aims to integrate artificial intelligence (AI) with human experience to optimise industrial processes.

The project, funded by the Government of Navarre as part of the 2024 call for strategic R&D projects, aims to develop a software platform that will help predict potential production defects and provide recommendations to help factories work more efficiently, reduce errors and improve their competitiveness.

The HUCAI-S4 project is based on the human-in-the-loop concept, an approach in which artificial intelligence does not work autonomously, but with human supervision and support. This means that humans contribute their experience and knowledge to adjust and improve AI systems, allowing for continuous adaptation to production conditions and more accurate prediction models.

The software platform proposed by HUCAI-S4 includes multimodal artificial intelligence, continuous learning and recommender systems.

Multimodal artificial intelligence combines different types of data (images, sensor readings, human feedback in natural language), leading to the generation of more complex but also more accurate and reliable models.

Continuous learning allows the predictive models to be constantly optimised through operator feedback, adapting to new situations and plant conditions. In addition, recommendation systems help operators to make efficient decisions in real time, improving quality and reducing error rates.

The HUCAI-S4 platform will be used in two pilot tests to prove its effectiveness: one in a manufacturing environment with automated processes and the other in an environment with manual processes. These tests will allow the development to be validated in different environments, making it easier to adapt to different contexts.

HUCAI-S4 is being coordinated by ZEO Technology, a technology company specialising in the design and implementation of products that improve industrial productivity, the digitalisation of factories and the overall efficiency of industrial plants, adding value to decision-making by capturing and processing relevant information in production processes. In the project, it is developing the pilot platform, technologies for data collection and processing, and solutions for human interaction.

NAITEC, a technology centre specialising in mobility and mechatronics in Navarre, is taking the lead in developing the artificial intelligence models. It will design and develop the multimodal AI platform for anomaly detection and recommendation generation, which will include machine learning algorithms. In addition, NAITEC will develop a component verification module based on advanced sensors and embedded AI.

Borg Automotive is Europe’s leading independent remanufacturer of automotive parts. Remanufacturing is the reconditioning of used parts or components to original equipment conditions. Borg is participating in the project as a use case for AI to optimise the production flow in a discrete industrial process and to inspect and sort parts. The ultimate goal is to seek improvements in productivity, efficiency and quality.

And Bacaicoa Industrias Plásticas is a company dedicated to the extrusion of low and high density polyethylene film for highly demanding sectors such as the agri-food and automotive industries. Bacaicoa is participating in the project as a use case with the aim of optimising the production process of three-layer film, which is affected by a significant volume of variables, through the advanced exploitation of information by means of AI models and recommenders.

HUCAI-S4 aims to be one of the first industry-oriented multimodal AI platforms that involves people, both by providing them with recommendations and by collecting their contributions and supervision to continuously improve the AI models. Furthermore, HUCAI-S4 aims to be easily applicable in different industrial environments through learning transfer and retraining. In this way, it overcomes one of the main barriers SMEs face in adopting AI: the need to develop specific and customised predictive models.