The quest for Autonomous Road Transport (ART)

The ontological complexity of the transportation system makes it challenging to accurately predict the effect of policies or measures. Therefore it is very difficult to design (socially) optimal solutions addressing inefficiencies inherent to the transportation system. The main origin of this complexity lies in its interaction with human behaviour. A clear example is Intelligent Transportation Systems (ITS). At the end of the 90s, they were expected to significantly contribute to increase road capacity by enabling a smarter demand management and avoid excessive infrastructure investments (as a response to economic and physical constraints). ITS, enabling continuous real-time information on the status of the transportation system, was expected to considerably reduce congestion by providing passengers and freight carriers information on the optimal route to follow. This proved to be not fully realizable since, unlike weather forecasts (which do not affect weather), traffic forecasts do affect traffic as users react to traffic information in a way that is, most of the time, not fully predictable. With further technological development and the possible deployment of autonomous vehicles in combination with electro-mobility, the realization of an “ideal” transportation system management is becoming more likely. Autonomous vehicles, can, at least partially, untwine the relation between the transportation system and human behaviour. Autonomous vehicles, some of them currently tested on European roads, are able “to perform all safety-critical functions for the entire trip, with the driver not expected to control the vehicle at any time. As this vehicle would control all functions from start to stop, including all parking functions, it could include unoccupied cars” ("U.S. DoT Releases Policy on Automated Vehicle Development". National Highway Traffic Safety Administration. 30 May 2013).

Autonomous vehicles are currently attracting the attention of researchers in many fields. The number of scientific papers published on this topic is growing exponentially (reaching more than 400 papers published in 2014) and the types of journals considering it worth of publications is also widening (reaching also the attention of Nature in 2015). The reason for this bubble is simple: autonomous vehicles might fundamentally reshape the rules of the road transportation system in the same way personal computers have changed our life (with current vehicle paradigms that in a few decades from now may give the same feeling of a typewriter today). Researchers and industry also expect significant synergies from autonomous driving and electro-mobility as possible enablers to an emission free individual mobility in Europe and see them as natural allies as they can technically cross-fertilize each other.

Apart from the further development of the on-board technologies and logics required to guarantee safe, efficient, and comfortable driving, new regulatory frameworks are necessary to allow these vehicles to smoothly flow over the existing road network (e.g. to address liability issues, the design of a new concept for driving licensing, etc.). Moreover, if these vehicles are connected to a central controller able to guide them on, e.g. by indicating the speed to keep, the path to follow and other decisions currently taken by drivers, then a step change in the optimization of the transportation system would be feasible. This would require evolving from the current “autonomous vehicle/driving” concept to the more complex Autonomous Road Transportation (ART) pursued with this proposal. An optimized ART could then be able to minimize the transportation system impacts on citizens and environment by setting the central controller to optimize a combination of travel time, travel costs, energy consumption, air pollution, collision risk etc.

Implementing an ART requires addressing a number of technical, political and social open issues, such as 1) who should be governing the central controller body and how it should optimize the transportation system (e.g. at which level? Urban, Regional, National? Which are the connected problems, also computationally?); 2) how the huge amount of data flows could and should be managed by the controller (and, e.g., how to deal with transmission problems/interruption? Are the vehicles provided with second by second information or only by predictive instructions that can be adapted during the trip? Which are the data privacy issues arising?); 3) which data are required by the vehicles to move on the transportation system in a fast, safe, reliable and efficient way (which devices are needed on any vehicle to explore the surrounding space in a reliable, fault-tolerant way? which data should be provided and maintained by the road transport authorities? Which requirements should they have?); 4) whether drivers would have in some cases the right and freedom to take his/her own decisions (for instance in relation to routing) or if, on the contrary, they would be basically users of a system owned by the society that can impose the rules of driving based on a collective optimal solution; 5) which are the priority areas for the ART (e.g. heavily congested and/or unsafe road areas that can be accessed only by autonomous vehicles); 6) how to manage a mix of autonomous and normal vehicles (at least during a transition period); 7) how people would perceive the possibility to buy an autonomous vehicle (e.g. how its efficiency and the reduced environmental impact would influence the choice).

All these issues will hardly hold a unique and simple solution. An in-depth analysis would therefore provide useful insights for a approaching the design of an autonomous system in a proper way. This will be extremely useful in the attempt of shaping the future legislations required to incorporate this new type of vehicles. There are a number of policy issues that need to be addressed in order to prepare Europe for a possible roll-out of ART. These include standardization, safety, security, data privacy and legal barriers such as liability. ART is a possible industry game-changer that could have huge impacts on the competitiveness of Europe and the creation or safeguarding of jobs in Europe. Delays in addressing pertinent policy questions as experienced in the relatively trivial case of a uniform EU emergency call number (112) to enable a basic feature as e-call, should be avoided. If the EU fails to be prepared for ART, it risks falling behind other regions as already experienced in the case of hybrid electric vehicles and electro-mobility.

In addition, assuming that a central controller has the role of optimizing a certain transportation system to minimize a combination of the overall travel time and costs, energy use and air pollution, risk of accidents, it will be necessary to have reliable, though inexpensive models able to evaluate the status of the transportation system, its short-term evolution, and the connected externalities in real-time. To this aim, traffic simulation models, fuel consumption and emissions models, pollutant dispersion models and collision risk models are necessary. Although many possible options in the different fields area available, an integrated solution going from traffic to pollutant concentration and risk of accidents does not exist yet.

Therefore the existing models need to be properly integrated and this is usually not a trivial task (with both modelling and software-related issues to be tackled). Research is therefore required to understand the need of the different models and to find a proper integrated solution. In addition, the estimation of CO2 and pollutant emissions from vehicles and their concentrations with sufficient accuracy (but also with sufficiently simple models) is an issue deserving research, especially in an era in which electric driving and other fuel savings technologies will considerably affect the vehicles' impact on the environment. It will require the development of a technology-based fuel consumption and emission model as well as a dispersion model able to take into account the land morphology.

Finally, if an efficient and robust modelling framework for the simulation of the transportation system and its externalities will be developed, the real-time optimization of the system itself and the possibility to detect critical traffic situations (peaks of pollutant concentrations, traffic situations with an increased risk of collision) are two topics that will also require research.

Real-time traffic optimization is indeed a topic that has considerably attracted the attention of researchers in the last decade. Many solutions have been proposed, all suffering for the computational requirements and the sub-optimality of the solution found.

Concerning the real-time identification of the black-spots for road safety, the main challenge is to identify the right proxy of road safety in traffic conditions. Many researches have been carried out in the last years, and the last proposals seem encouraging.