1
Challenges Involving Ad Hoc Communications on the Road

1.1. Introduction

With the arrival of two key innovations that will revolutionize communications in future transportation systems, we felt it was the perfect time to present a summary of our research work, which for the most part has been collaboratively achieved over nearly a decade.

The first innovation is the fully autonomous vehicle equipped with a hyperconnected on-board computer and whose operation on the roads is scheduled for the near future. Initiated in the 1970s for guided public transport, in particular with the VAL metro in the region Hauts-de-France in France, the use of communications and autonomous vehicles has gradually expanded and increased in land transport. For example, in the 2000s, mobile communications (Global System for Mobile communications-Railway (GSM-R)) were introduced into the European Rail Traffic Management System (ERTMS), which made it possible to harmonize and streamline high-speed train traffic across Europe. Advances in automation and localization have enabled the Société nationale des chemins de fer français (SNCF) to forecast the autonomous freight train for the end of 2021 and to effectively start testing on rails in early 2022, including the autonomous regional passenger train. In addition, driver-assistance applications initially, offered to drivers through unconnected devices (navigation, hazard warning, etc.), have evolved into connected versions available on smartphones before being fully integrated into the now connected on-board computers of vehicles, including their cooperative versions. In parallel with this increasing connectivity, improvements in sensing systems and autonomous driving systems have made it possible to improve driving automation to the levels currently observed in some vehicle models. Following this perspective of full driving automation, safety considerations have been increased and, consequently, must be inevitably applied to the communication system that plays a crucial role in the operation of many of these applications for the vehicular context. Communication systems for the transport of the future will therefore need to clearly integrate both the autonomous vehicle and the ability to support all essential functionalities for the remote control of a vehicle, including that initially driven by a human (as for example in a distressful situation).

The second innovation is 5G technology, whose deployment in France is already effective at the time of writing. It completes the convergence of communications, in particular by integrating the Internet of Things (IoT) into the long list of devices already brought together by 4G through all-IP. In terms of access, 5G assimilates all modes of communication, especially those concerning vehicles, namely vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V), while providing very low latency and many other performances at levels largely adapted to vehicular communications. In terms of the organization of access and the core of the network, 5G incorporates most of the recent innovations, resulting in particular from the diversification of access (frequency extension and pooling), the softwarization of radio and network operations (Software-Defined Radio and Software-Defined Network, respectively) and the virtualization of network-related functions (Network Function Virtualization) through transferring most of the transfer operations of specific hardware devices to virtual machines, especially in the cloud. Combined with paradigms such as edge computing, fog computing and named data networking (NDN), 5G should enable much faster access to data as close as possible to its use, which represents a clear advantage for vehicular applications, most of which require very high responsiveness. This very wide variety of applications presuppose that 5G could soon become a mobile communication system for all smart city applications, and consequently for communications in land transport systems. Therefore, the communication systems for the transport of the future need to be rethought by taking into account the drastic changes that 5G will bring and whose major impact on the design and evaluation of these systems is foreseeable.

1.2. The railway experience in the vehicular networks of the future

In rail transport, most particularly, the question of the future communication system arises more urgently due to the obsolescence and the forecast end of the maintenance of GSM-R technology. The adoption of GSM-R as a telecommunications subsystem of the ERTMS had been validated following several years of measurement campaigns and on-site tests that could not be replicated today for cost reasons. Several research projects have proposed solutions for performing simulation-based evaluations of alternative communications technologies to GSM-R, taking into account elements of the railway context likely to affect communications (tracks, train movements, interchanged traffic, etc.) (Sondi et al. 2017). Similar solutions have been used to evaluate other communication technologies as potential candidates to replace GSM-R, including Wimax (Pinedo et al. 2015), LTE (Sniady and Soler 2014) and 5G (Ai et al. 2020).

Rail infrastructure provides the first model for large-scale deployment, compared to the underground, which often finds itself limited to the city scale, of a communication system that allows the use of fully autonomous vehicles in land transport in Europe in the short term. The harmonization already undertaken at the European level between the various national rail traffic management systems reinforces the idea that rail can be the ideal support model for the development of communication systems for all future land transport in Europe. It is a foundation that must nevertheless be observed through its successes, but also through its pitfalls, while considering its specificities to avoid biases in the conclusions to be drawn for future systems intended for all land transport. Just to mention a few of the most important aspects, we should consider the following:

• The dedicated nature of the railway infrastructure which currently circumscribes both the scope of investments, the investor and the beneficiary. It is a different matter when it comes to a future global infrastructure for all land transport. For example, even if it could have happened, financing the deployment of GSM-R base stations in a white zone today at a loss is already an issue if there is a safety issue for a few trains crossing it. It is not clear that in the future operators would deploy a similar infrastructure for motorists on roads in sparsely populated areas where it would not otherwise be profitable. The design of a comprehensive infrastructure will need to address this aspect.
• State hegemony over a railway operator that facilitates local decisions within each state and consequently harmonization at the European level. Each national historical operator actually generally imposes its technological choices at the level of its state. In addition, since the telecommunications infrastructure is a dedicated infrastructure, it also remains subject to management under the sole authority of the railway operator. It is thus easier to reach a consensus at European level with one decision-maker per state (the railway operator) rather than in a context where several actors with divergent interests have been involved. The challenge will have to be addressed differently for a global infrastructure including roads for cars, as it will involve a myriad of national land transport operators (rail, road, etc.), not to mention telecommunications operators also concerned by any infrastructure deployment in the public domain. Additionally, the latter can also comprise numerous service providers and a large variety of clients, ranging from government services to the general public, including companies of the transport sector, of which car manufacturers. This multiplicity of decision-making partners, combined with the pressure of the very varied expectations of the different beneficiaries with heterogeneous profiles, presupposes that the process of harmonizing technological and normative choices will be far more complex in the context of a global infrastructure for the land transport of the future.
• A relative de facto insensitivity to intermediate technological developments. Indeed, the colossal investments required to equip railway infrastructure on a European scale mean that, on the one hand, the technologies chosen have to be sufficiently mature to guarantee their sustainability, and, on the other hand, that they be need to be maintained for as long as possible over time to amortize costs. For example, it is interesting to observe that GSM-R is still being deployed on railway lines, although the end of the maintenance of the GSM technology has already been announced by manufacturers. Meanwhile, the general public has shifted through 3G, then 4G and now 5G. It will probably be necessary to wait for the generalization of this latest disruptive technology for the rail sector to start new investments in Europe to achieve a technological leap, while the respective implementation of 3G and 4G has been systematically accompanied by new applications for drivers through smartphones and on-board vehicle computers. A future global infrastructure for vehicular communications is likely to be more sensitive to technological developments. It will therefore have to be designed to allow for faster and less costly adaptation compared to what has been observed in the rail...