An Interpretable and Explainable AI Framework for Urban-Suburban Traffic Analysis and Understanding

Studying and understanding road traffic can be crucial for supporting the activities of many stakeholders. It may allow automated supervision of the ongoing situation, raising warnings or alarms in case of anomalies. It may be used to plan interventions on road and town organization. It may provide advanced support in decision-making. A manual approach is infeasible, due to the number and complexity of the involved entities and tasks, and the automated systems and their outcomes must not be black boxes. In this paper we propose TrAnSIT, an AI-based framework that can cover a wide range of traffic-related issues, from overall urban or suburban traffic management to surveying specific road segments. Most of its functions allow a human-level understanding of outcomes. We also describe various use cases showing how TrAnSIT can be useful for the management of emerging traffic problems and the identification of critical areas, offering timely and data-driven solutions to optimize traffic flow and road safety and facilitating informed strategic decisions.