Research Areas

Mathematical Foundations of Urban Complexity

We explore cities as complex systems, integrating theoretical and computational perspectives from mathematical physics. Our work investigates the multiscale spatial-temporal dynamics of urban processes, the physical underpinnings of digital twins, and the structural principles of urban phenomena. Using tools from multilayer network theory, dynamical systems, statistical mechanics, and both data-driven and analytical modeling, we aim to build a rigorous foundation for understanding and simulating urban complexity.

Mobility and Network Dynamics

Network Theory and Random Walks

We investigate the dynamics of complex networks through random walk models to analyze node centrality, community structure, and diffusion processes. Applications include modeling of urban infrastructure, mobility systems, and energy networks.

Urban Vehicular Mobility Modeling

We develop large-scale simulations of vehicular traffic in cities such as Bologna to predict congestion patterns, optimize routing, and improve traffic light coordination. Our models integrate real-time data from multiple sources and use adaptive mechanisms for control and forecasting.

Railway Delay Prediction from Italian Train Data

Using national railway data, we construct models to reproduce and predict train delay distributions. The aim is to understand systemic inefficiencies and provide data-driven insights for enhancing network reliability and punctuality.

Urban Energy and Vegetation

Building Energy Simulation

We simulate the energy performance of buildings based on their morphology, construction period, and usage. This enables the identification of energy- intensive areas and supports urban decarbonization strategies through informed policy recommendations.

Urban Vegetation Analysis

Using airborne LiDAR, orthophotos, satellite imagery, and remote sensing, we detect and classify urban trees, estimate structural attributes like height and crown width, and monitor vegetation changes over time. These analyses support green infrastructure planning and biodiversity management.

Urban Heat Island (UHI) Analysis

We study the UHI effect by integrating spatial indicators such as Sky View Factor (SVF), land cover, and built morphology. The goal is to assess microclimate variability and identify zones most vulnerable to heat, enabling targeted climate adaptation measures.

AI and Urban Modeling

Urban Language Models and Vision-Aware Agent

We explore how large language models (LLMs) can be trained or fine-tuned to understand spatio-temporal urban processes and respond to complex urban queries. Our research investigates generalization capabilities of LLMs, their integration with vision transformers, and their role as autonomous agents navigating urban data.

Physics-Informed Neural Networks (PINNs) for Urban Modeling

To ensure that AI models reflect underlying physical laws, we use physics- informed neural networks that embed dynamical constraints into learning architectures. These models allow for more accurate and interpretable simulations of urban processes, from energy flow to climate dynamics.