Intelligent mobility management in emerging vehicular networks

车辆网络是一种新兴的网络，它可以将车辆彼此连接起来，并将它们与道路基础设施连接起来。管理此类网络中的移动性提出了我们希望在本提案中解决的几个挑战。这些网络使得实现不同类型的应用成为可能：例如，避免碰撞和道路工作的安全应用，自动驾驶辅助的实时应用，管理交通和建议绕路的智能交通系统（ITS）应用，高速公路自动收费等舒适应用，连接到在线多媒体内容等。为了高效运行，车辆对计算能力和几乎连续连接的需求不断增加。我们在此提出的研究计划侧重于新兴车辆网络中的智能移动管理。它涵盖了移动网络、人工智能、机器学习、数据科学和数学优化等广泛领域，以解决车辆造成的道路拥堵和交通堵塞问题。本研究计划的主要目标是设计智能方法、算法和工具，通过交通拥堵预测、检测和控制，有效地实时管理车辆的移动性。从长远来看，该研究计划旨在为下一代基于机器学习的自组织车辆网络提出一种移动性管理架构，并赋予其预测交通拥堵或减少交通拥堵的能力。为了实现这一目标，我们提出了一种基于大数据的四个v特征的方法：体积（Volume）、多样性（Variety）、准确性（Veracity）和可变性（Variability）。我们的长期架构将允许从网络生成的数据和用户中收集、提取和理解相关信息，从而推断出战略行动和决策，以检测、控制和预测交通拥堵，并为道路使用者提供更好的服务质量和更好的体验。无论是从科学研究的角度还是从自动驾驶汽车的商业扩张的角度来看，新兴车辆网络中的移动管理都是一个非常重要的问题。它提出了许多限制这些网络采用的挑战。事实上，车辆网络将使各种应用成为可能：道路安全应用、高效运输应用、监控应用。

Vehicular networks are emerging networks that make it possible to connect vehicles to one another and to link them with road infrastructures. Managing mobility in such networks raises several challenges we wish to address in this proposal. These networks make it possible to implement different types of applications: safety applications to avoid collisions and road work for example, real-time applications for automated driving assistance, intelligent transport system (ITS) applications to manage traffic and suggest detours amongst others, comfort applications such as automatic toll payments on motorways, connection to online multimedia content, etc. To function efficiently, the vehicle has an increased need for computing power as well as a nearly-continuous connection. The research program we hereby present focuses on intelligent mobility management in emerging vehicular networks. It covers areas as broad as mobile networks, artificial intelligence, machine learning, data science and mathematical optimization, in order to address road congestion created by vehicles and causing traffic jams. The main objective of this research program is to design intelligent methods, algorithms and tools to efficiently manage vehicular mobility in real time through traffic congestion prediction, detection and control. In the longer term, this research program aims at proposing a mobility management architecture for next-generation ad-hoc vehicular networks based on machine learning and endowed with the ability to predict traffic congestion or to reduce it. To achieve this objective, we propose an approach based on the four Vs that characterize big data: Volume, Variety, Veracity and Variability. Our long-term architecture will allow to collect, extract and understand relevant information from network generated data and users in order to infer strategic actions and decisions to detect, control and predict traffic congestion and provide better quality of service as well as an improved experience to road users. Mobility management in emerging vehicular networks is a problem of great importance both from the point of view of scientific research and the commercial expansion in autonomous vehicles. It raises many challenges that condition the adoption of these networks. In fact, vehicular networks will make possible various applications: road safety applications, efficient transport applications, surveillance applications.