Human Factors, Human Machine Interaction and AI

Description

The research activity focuses on developing AI-based pipelines for the automatic recognition of emotions, attention, and cognitive states from facial expressions, as well as on methods and tools to support continuous, real-time monitoring of physical ergonomics in factory settings and to foster musculoskeletal risk prediction models, leveraging low-cost systems based on RGB cameras, wearable sensors, and automatic posture-recognition models.

The goal is to integrate ergonomics, safety, and efficiency within the Industry 4.0 and 5.0 framework, such as in projects on agile factories, predictive maintenance, and smart retrofitting. These results feed into platforms for continuous user-experience analysis and for adapting interfaces, content, and services. This area includes both methodological contributions (definition of architectures, datasets, and metrics) and experimental applications, such as a platform for analyzing student–teacher interaction or toolkits for automated User Experience analysis. Recent developments involve the research, structuring, and application of synthetic training sets for neural network training to improve the accuracy of human-factor prediction even in highly complex contexts (factory and automotive). These datasets, defined as synthetic because they do not derive from data acquisition on real subjects but from computational procedures, make it possible to drastically speed up the creation of datasets for training Artificial Intelligence algorithms, while also achieving an extremely high quality level (reduced bias and acquisition noise). Algorithms trained on these synthetic datasets are applied to physical ergonomics monitoring in industrial environments, attention and gaze-direction monitoring in the automotive domain, and emotion detection for affective computing.

Publications
  • Agostinelli, T., Generosi, A., & Mengoni, M. (2026). A novel approach for monocular RGB-based ergonomics monitoring in industrial workspaces employing synthetic datasets to train a deep learning model. The International Journal of Advanced Manufacturing Technology, 1-24. DOI: https://doi.org/10.1007/s00170-025-17168-1
  • Generosi, A., Ceccacci, S., Faggiano, S., Giraldi, L., & Mengoni, M. (2020). A toolkit for the automatic analysis of human behaviour in HCI applications in the wild. Advances in Science, Technology and Engineering Systems Journal, 5(6), 185–192. DOI: https://doi.org/10.25046/aj050622
  • Generosi, A., Ceccacci, S., Tezçi, B., Montanari, R., & Mengoni, M. (2022). Nudges-based design method for adaptive HMI to improve driving safety. Safety, 8(3), 63. DOI: https://doi.org/10.3390/safety8030063
  • Agostinelli, T., Generosi, A., Ceccacci, S., & Mengoni, M. (2024). Validation of computer vision-based ergonomic risk assessment tools for real manufacturing environments. Scientific Reports, 14, 27785. DOI: https://doi.org/10.1038/s41598-024-79373-4
  • Generosi, A., Agostinelli, T., & Mengoni, M. (2023). Smart retrofitting for human factors: A face recognition-based system proposal. International Journal on Interactive Design and Manufacturing (IJIDeM), 17(1), 421-433. DOI: https://doi.org/10.1007/s12008-022-01035-4
Staff

Prof.ssa Maura Mengoni 
Tel. +39 071 220 4969
E-mail: m.mengoni@staff.univpm.it