Cyber-physical Distributed Simulation of Intelligent Transportation Systems

点击翻译按钮获取中文摘要

Intelligent Transportation Systems (ITS) has been steadily growing in recent years due to the latest advancements in Information and Communications Technology (ICT). Vehicles have become smarter and embedded with sensors and processing units. Also, new communication paradigms have enabled connected vehicles, allowing the exchange of information, which significantly supports new trends in autonomous driving and transportation electrification. All these elements come along with incentives for developing Smart Cities, in which there is an inherent interaction between vehicles and mobile devices and the infrastructure, allowing to happen built-in context-awareness.Highly complex systems are built upon ITS, demanding more precise control and efficient management of resources, vehicles, systems, and the environment. Modelling and simulation move closely alongside such demands, becoming an essential tool for ITS training, emergency preparedness and response, estimation and prediction, and scenario analysis. However, the features and requirements of intelligent transportation have conditioned simulations to support high consistency and fidelity on reproducing actual environments. As a consequence, models for and of ITS become highly complex and detailed; the complexity grows as IoT-oriented sensed data sources are integrated into such simulation systems.Existing models and simulators have been explored to provide estimations and analyze traffic scenarios to support transportation systems on the traffic control. Macroscopic models have allowed predicting traffic load through abstraction of details and strong assumptions. Such assumptions might be unrealistic but effective for simplifying and speeding up processing towards timely optimizations. On the other hand, microscopic simulations enable exploratory analysis from traffic scenarios, which are generated based on a broad set of parameters. These simulations usually involve just traffic, movement of vehicles, and allow to observe complex virtual, synthetic environments extensively. This proposal aims at defining a robust research program towards the development of a distributed simulation platform for and of ITS. The platform contemplates the modelling of large-scale urban centres, where distributed agent-based simulations represent the individual movement and interaction of vehicles, pedestrians, and infrastructure, and traffic-flow models estimate the load and trends. The platform is intended to provide precise estimations from real or synthetically generated data and serve as a validation and evaluation tool for novel ITS solutions. The input of real-time data in the simulations requires a support system for constant communication/interaction with the real environments, providing high performance through Cloud-based resource management, load balancing, and data dissemination.