ZIDI

Scientific Excellence for Societal Challenges

Intelligent Models and Interfaces for Safety and Health

ZIDI

Scientific Excellence for Societal Challenges: Intelligent Models and Interfaces for Safety and Health

The project investigates innovative solutions based on virtual models, artificial intelligence and advanced multimodal interfaces to address key societal challenges in environmental risk management, cybersecurity and healthcare.

In environmental risk management, the project focuses on developing explainable models for predicting, recognising and managing the consequences of disasters caused by climate change and human activity, such as wildfires and pollution. In cybersecurity, a methodology is being developed for using large language models as digital twins of users for ethically acceptable testing of resilience to phishing attacks. In biomedicine, the project employs virtual models of the human body for the development and testing of medical devices and the optimisation of diagnostic and therapeutic procedures, reducing the need for costly and risky physical experiments. In telemedicine, the focus is on developing AR/VR, speech-gesture and BCI interfaces for rich, spatially aware and empathetic remote collaboration.

The project delivers innovative methodologies and technologies, fosters scientific excellence and lays the groundwork for socially beneficial and sustainable technological breakthroughs.

Project Goals

1. Environmental Risks Develop explainable and reliable artificial intelligence models for the automatic recognition, prediction and mitigation of the consequences of climatic and environmental risks, with the aim of increasing community resilience and environmental safety.

2. Telemedicine Develop advanced user interface models with the aim of creating a telemedicine platform based on extended reality (XR), enabling multimodal collaboration, personalised patient visualisation and remote diagnostics through interactive and spatial protocols.

3. Biomedicine Establish a digital twin methodology in biomedical experiments and medical device development, with the aim of optimising laboratory testing, reducing costs and creating a virtual patient.

4. Cybersecurity Develop AI simulation methods for cybersecurity testing based on virtual personality models, including LLM emulation of user behaviour and testing of responses to various types of phishing attacks under controlled conditions.

Work Packages

WP1
Virtual Models and Interfaces for Recognition, Prediction and Management of Environmental Risks Using Artificial Intelligence
Research into AI models for automatic recognition, prediction and mitigation of the consequences of climatic and environmental risks, with the aim of increasing community resilience and environmental safety. Activities include the use of satellite imagery, images and sensor data for risk analysis, development of simulations (e.g. what-if scenarios) for modelling wildfire and pollution spread, and creation of interactive tools for prediction and visualisation of environmental threats.
WP2
Advanced Multimodal Interfaces and Collaborative XR Environments in Telemedicine
Research into advanced user interfaces (AR/VR, speech-gesture and BCI interfaces) and integration of multimodal interactions (speech, gesture, space). Key activities: evaluation of collaborative XR environments for clinical decision support in telemedicine, analysis of emotional awareness and application in diagnostics, and optimisation of interfaces for fast and precise communication between medical professionals.
WP3
Digital Twins for Biomedical Applications
Development of digital twins for simulation of biomedical devices and physiological responses. Key steps: modelling of the human body using software tools (Sim4Life, Altair FEKO, CST), creation of virtual prototypes for testing medical devices and therapeutic approaches, and integration of biophysical data in real time for personalised simulations.
WP4
Virtual Models for Cybersecurity Testing
Research into user behaviour emulation for evaluation of cybersecurity vulnerabilities. Activities include the use of OSINT data for user profiling, application of large language models (LLM) for simulation of reactions, emotions and decisions, and development of a methodology for testing susceptibility to phishing attacks and other threats.
WP5
Project Management, Ethical Oversight and Open Science
Project coordination and promotion of transparency and ethical standards. Focus on progress monitoring, team and budget management, ensuring data and results availability through open repositories, preparation for commercialisation and collaboration with health and industry partners, and ethical evaluation (working with user data, health data and LLM simulations).

Project Team

Project Leader: Prof.Ljiljana Šerić, PhD

Ljiljana Šerić
Project Leader
Damir Krstinić
Researcher
Marin Bugarić
Researcher
Maja Braović
Researcher
Mladen Russo
Researcher
Marjan Sikora
Researcher
Matko Šarić
Researcher
Maja Stella
Researcher
Jurica Đerek
Researcher
Luka Kraljević
Researcher
Davor Meter
Researcher
Mario Čagalj
Researcher
Antonio Šarolić
Researcher
Darko Stipaničev
Researcher
Antonia Ivanda
Researcher
Dominik Žanić
Researcher
Antonia Bartulović
Researcher
Viktoria Pavlovski
Researcher
Mirna Marić
Researcher

Contact

For all enquiries related to the ZIDI project, please contact the project leader:

Prof.Ljiljana Šerić, PhD Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture University of Split Ruđera Boškovića 32, 21000 Split, Croatia

Email: Ljiljana.Seric@fesb.hr