Efficient, sustainable and reliable transport in large urban areas is vital for the economic and social development of modern cities and represents a major research challenge involving multiple disciplines, ranging from information and data science to economics and urban planning. The daily mobility of large masses of people and goods in, from and to urban areas strongly depends on a seamlessly available multimodal transport infrastructure, and on proper knowledge of the mobility demand. Large cities are facing crucial issues in achieving these objectives.
In 2009, for the first time ever, more than half of the world population lived in urban areas, and, by 2030, cities are projected to house 60 percent of people globally. The growing concentration of people translates into dramatically increasing demand, frequent perturbations and high strains on the urban transport infrastructure, which degrade users’ mobility experience, decrease air quality and increase the risk of malfunctions and cascading failures, especially in the presence of adverse or extreme weather conditions. On the other hand, cities are becoming “smarter” and therefore able to collect huge amounts of data in real-time, via connected sensors, devices, people and infrastructures. Such massive data possesses the great potential to provide highly valuable insights on how citizens move, work and live within smart cities. However, this data is still far from being jointly mined in effective and efficient ways that can enact and promote more resilient and sustainable transport in city life.
In this context, the principal goal of PROMENADE is to devise a novel systemic, real-time data-driven platform for the efficient, resilient and smart management of multi-modal urban transport, by integrating innovative and sustainable solutions based on complex networks modelling, machine learning and big data technologies. This main goal is decomposed in three objectives.

The first objective of PROMENADE consists in proposing a novel, data-driven and dynamic graph-based modelling of multi-modal, large-scale urban transport networks. Such model aims at grasping the complex inter-relationships existing among transport modes and allowing for a more accurate estimation of their inner properties. Since most of the urban transport networks operate nowadays at capacity limits, any disruptions of single transport modes (e.g. rail transport) or components of the infrastructure network (e.g. bridges, tunnels) have negative effects on the mobility provided by other modes as well. For instance, if the railway system is disrupted, more people use their own vehicles, leading to congestion and mobility reduction overall. Traditionally, urban transport networks (e.g., roads, bus and tram lines, the underground system, etc.) have been studied separately, by exploiting different models and solutions depending on the analyzed mode. These mono-modal approaches are mostly ineffective, since they completely neglect the inter-relationships existing among the multiple actors and infrastructural components making up large-scale, real-world, urban transport systems. They are unable to identify, anticipate and mitigate network vulnerabilities, and preventing cascading failures due to minor or hardly predictable major events. Additionally, such inter-relationships evolve very rapidly due to frequent changes in users’ demand and behaviors, network offer and external factors (e.g., weather and social events).
PROMENADE tackles the need for overcoming traditional static and mono-modal approaches in modelling and analyzing urban transport networks and their interactions.
To dominate such complexity, PROMENADE targets a novel holistic modelling framework based on multi-layer networks for capturing the complex and dynamic interactions existing among multiple transport modes, urban infrastructures (e.g., land, telecommunications, water system, power grid, etc.), and urban actors (i.e., network providers, users, planners and operators). The achievement of this first objective will enable a breakthrough with respect to state-of-the-art solutions that exploit simplistic, partial models and basic metrics to study small-scale and mono-modal urban transport networks.

As its second objective, PROMENADE aims at jointly mining multi-source, large-scale data on users’ mobility to retrieve more complete, dynamic and up-to-date information on the multi-modal travelers’ patterns in light of studying the resilience of multi-modal urban networks. Pervasive sensors, mobile devices and digital services are nowadays deployed with high spatial density and are frequently accessible in real-time. They are therefore capable of generating a large-scale, continuous stream of data that can be opportunistically leveraged for unlocking an accurate and dynamic knowledge of how, when, where and why people and goods move in the city. Such knowledge can thus be exploited to identify and anticipate critical situations as well as for making more informed decisions for performance improvement and robust reaction to unpredictable events. As an example, data from cooperative and autonomous vehicles allow for preventing shockwave formation and propagation, reducing accident probability, lowering energy consumption and CO2 emissions. Similarly, user-generated, large-scale data (e.g., mobile phone, social media and ticketing data) have recently gained great attention from researchers and practitioners as they offer low-cost alternatives to estimate urban dynamics, detect land use, identify vulnerabilities and efficiently manage communication resources.
However, mining such data remains very challenging, especially if heterogeneous sources must be jointly processed, in real-time and on large scale.
In PROMENADE, heterogeneous large-scale datasets on users’ mobility will be available and leveraged to accurately and dynamically reconstructing mobility patterns and indicators of traffic dynamics.
Based on the data and analytics above, PROMENADE also proposes to augment the multi-layer topological modelling of the urban network by projecting onto its layers (as time-varying attributes of edges and nodes) the reconstructed and dynamically evolving spatio-temporal demand features, as well as the supply characteristics of the analyzed sub-networks (e.g., road length and capacity, number of lanes). Such an informed and dynamic multi-layer graph can be therefore efficiently studied in order to evalu-ate the performance and resilience of the multi-modal transport network. Useful indicators of network vulnerability (e.g., centrality measures) and traffic performance (e.g., macroscopic fundamental dia-grams) will be dynamically estimated, at different scales, on top of the multi-layer, data-driven model-ling framework by relying on parallel solutions and big data technologies for efficient real-time computation. By achieving this objective, PROMENADE will overcome traditional approaches for mobili-ty pattern estimation, typically based on mono-source data with rapidly obsolescent results. Multi-source, real-time, big-data solutions will allow grasping more complete and large-scale information on user mobility. Moreover, by encoding and dynamically updating this information in the multi-layer model, a first-time data-driven, dynamic and systemic platform for multi-modal transport performance analysis will be realized.

As its third objective, PROMENADE targets the crucial need for novel metrics, approaches and strategies for the assessment and improvement of the resilience of multi-modal urban transport networks. Resilience is traditionally defined as the capability of a system to absorb hardly predictable shocks (i.e., robustness), to guarantee proper levels of quality of service in presence of disruptions (i.e., reactivity), and to allow for a rapid return to normal operation (i.e., recovery). Guaranteeing high levels of resilience in multi-modal transport is a fundamental challenge towards more sustainable and safer cities. A relevant example in that sense is the evacuation mandated by the local administration in Florida, a few days before the passage of hurricane Irma in Sept. 2017. The evacuation generated a major gridlock on the northern highways and tripled traffic on other major roadways, since Florida relies on only two highways going north. Another recent example is related to the unexpected cold wave that has hit Central Europe in February 2018 and completely paralyzed for days the urban transport infrastructure of both French and Italian capital cities: major and long-lasting traffic jams, train delays and bus lines suppressions have raged during and after the snow storms, with severe so-cial and economic consequences. State-of-the-art studies on transport resilience traditionally neglect multi-modality, real-time traffic dynamics, and users’ mobility patterns. Additionally, extreme weather events represent a fundamental category of risk that increasingly challenges transport systems and whose impact on multi-modal networks has still not been properly quantified from a resilience perspective. PROMENADE aims at jointly leveraging stress tests, coupled with data-driven simulation, for achieving a more accurate and dynamic characterisation of resilience in multi-modal net-works. Stress testing is a technique especially practiced in the financial industry and consists in designing critical scenarios for quantifying the capability of a system to absorb shocks, by estimating their risks for its users and developing contingency plans. PROMENADE will devote efforts to designing a set of relevant scenarios for the effective assessment of resilience for multi-modal networks, focusing on weather-related extreme events (e.g., heavy rains, flooding). Moreover, novel adaptive strategies for improving resilience will be proposed. The achievement of this third objective will enable a break-through in resilience engineering of multi-modal urban networks. Finally, all the metrics, solutions and analytics produced in the project will be integrated in the form of a customizable, extensible and freely accessible software platform aimed at anticipating and handling vulnerabilities and perturbations to improve the short-term adaptability of urban multi-modal networks.

Fig. 1. The 3 main research lines of PROMENADE

Fig. 2. Gantt Diagram