Wenchao Xu - Internet Access in Vehicular Networks

This Springer imprint is published by the registered company Springer Nature Switzerland AG

The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Modern vehicles generate a large amount of data for many emerging automotive applications, such as road safety, traffic management, autonomous driving, and intelligent transportation system. The Internet of vehicles are expected to enable effective acquiring, storage, transmission, and computing for such big data among vehicle users and thus can facilitate better perception of both internal and external vehicular environments as well as the status of drivers, passengers, and pedestrians. To efficiently connect mobile vehicles to the Internet and conduct massive information exchange among vehicle users and the transportation system, it is essential to evaluate the Internet access performance via both the vehicle-to-roadside (V2R) and vehicle-to-vehicle (V2V) paradigms. Besides, in order to support reliable and efficient Internet access for mobile vehicle users, it is very important to explore various spectrum resources rather than solely relying on cellular networks, such as unlicensed WiFi band, TV White Space, opportunistic V2V data path. To set up effective Internet connection, a practical signaling process between a roadside access station and vehicle users is necessary, e.g., various management frames need to be transferred to set up effective Internet connections. As vehicles move around, it is possible for them to set up opportunistic connections to share the data contents, bandwidth, and computing capacities, which can enable a variety of novel computing and communication paradigms that are beneficial to future automotive applications. In this monograph, we investigate the Internet access procedure and the corresponding analytical evaluation methods, as well as novel machine learning paradigms for reliable and robust Internet connectivity on the road. We first introduce the Internet access of vehicles and propose an analytical framework for modeling of Internet access performance via the roadside hotspots, considering the necessary Internet access procedure that comprised of association, authentication, and network configuration steps, where the access delay and throughput capacity are evaluated in drive-thru Internet scenario. We then explore the interworking of different V2X communication paradigms and study the opportunistic assistance from neighboring vehicles, which apply the V2V communication to conduct Internet data offloading upon the interworking of V2V and V2R communication, where the trade-off between the delay and throughput of the V2V assistance is analyzed. In addition, we take a close look at the wireless link management between the vehicle and Internet access stations. We investigate the V2X channel that is highly varying and thus difficult to accommodate proper modulation and coding scheme to satisfy various user quality-of-service (QoS). To deal with such issue, we apply big data analytics and show that the proposed data-driven and learning-based methods can greatly reduce the packet drop rate and thus improve the Internet access performance in terms of both access delay and transmission throughput. Several case studies are presented to examine the utility of the big vehicular data to enable the intelligent Internet access. Furthermore, to train the machine learning models among vehicle users in a distributed manner, we design efficient IoV protocols to boost the training process, including the rateless coding-based broadcasting scheme for intelligent model delivery that can enhance the process of collaborative learning among vehicles, whereby asynchronous federated learning can be conducted for mobile vehicles with high mobility and opportunistic inter-contacts. We hope that this monograph will provide inspiration and guidance on further research and development of the future Internet of vehicles.

We would like to thank Prof. Weihua Zhuang from the University of Waterloo (UW), Prof. Song Guo from The Hong Kong Polytechnic University, Prof. Nan Cheng from the Xidian University, and many UW BBCR members for their contributions in the contents of this monograph and great support to related research projects. Special thanks are also due to the staff at Springer Science+Business Media: Susan Lagerstrom-Fife, for her great help in the publication of the monograph.