A digital twin framework for urban parking management and mobility forecasting

Area of Science:

• Urban planning and transportation science
• Data science and artificial intelligence
• Environmental management

Background:

• Rapid urbanization presents challenges in urban mobility, including traffic congestion and pollution.
• Inefficient parking management and forecasting hinder sustainable urban development.
• Existing systems lack integrated data for comprehensive mobility solutions.

Purpose of the Study:

• To develop and implement a digital framework for integrated urban parking management and mobility forecasting.
• To leverage diverse data sources for predictive and generative modeling of urban mobility.
• To enhance urban planning and resource allocation through data-driven insights.

Main Methods:

• Integration of historical and real-time data (parking, sensor, weather, temporal patterns).
• Application of descriptive statistics, Spatial-Temporal Identity model for prediction, and Conditional Variational Generative Adversarial Network for digital twin generation.
• Utilizing Generative Artificial Intelligence for simulating 'what-if' scenarios.

Main Results:

• The framework successfully forecasts parking demand and generates spatial data for planning.
• Generative AI enables virtual testing of mobility strategies, improving efficiency and accessibility.
• Validation in Caserta demonstrated the framework's robustness and adaptability.

Conclusions:

• The digital framework offers a powerful tool for enhancing urban mobility management and sustainable planning.
• Improved parking meter placement and reduced inefficiencies are key benefits.
• Further data expansion and component refinement are recommended for full potential realization.