key: cord-0836036-xc8mvk05
authors: van Wee, Bert; Milakis, Dimitris; Thomopoulos, Nikolas
title: Overall synthesis and conclusions
date: 2020-07-14
journal: nan
DOI: 10.1016/bs.atpp.2020.06.001
sha: d35ffff3c44bae9d1f970825290e885d396e26ad
doc_id: 836036
cord_uid: xc8mvk05

Abstract This chapter first systematically summarizes the most important findings and policy implications of each of the chapters “Factors affecting traffic flow efficiency implications of connected and autonomous vehicles: A review and policy recommendations” by Narayanan et al.; “Automated bus systems in Europe: A systematic review of passenger experience and road user interaction” by Heikoop et al.; “Cyber security and its impact on CAV safety: Overview, policy needs and challenges” by Katrakazas et al.; “Cybersecurity certification and auditing of automotive industry” by Mateo Sanguino et al.; “The wider use of autonomous vehicles in non-commuting journeys” by Kimber et al.; “Policy implications of the potential carbon dioxide (CO2) emission and energy impacts of highly automated vehicles” by Annema; “Potential health and well-being implications of autonomous vehicles” by Singleton et al.; “Data protection in a GDPR era: An international comparison of implications for autonomous vehicles” by Costantini et al.; “Ethical issues concerning automated vehicles and their implications for transport” by Dogan et al.; “Governance cultures and sociotechnical imaginaries of self-driving vehicle technology: Comparative analysis of Finland, UK and Germany” by Mladenović et al.; “Wider implications of autonomous vessels for the maritime industry: Mapping the unprecedented challenges” by Ghaderi; “The potential for automation to transform urban deliveries: Drivers, barriers and policy priorities” by Paddeu and Parkhurst. Next it synthesizes the overall findings and policy implications, and discusses future avenues for policy making and research. A first conclusion is that the chapters make clear that the ranges in policy relevant implications of AVs, within the scope of each chapter/topic, are still relatively broad. Secondly we conclude that research that is conceptually rich is more valuable for policy making. Thirdly we hypothesize that context matters for the uptake, impacts, and specific system design characteristics of real world AV implementation. Fourth we conclude that research on the global south has been limited so far. Fifth we argue that AVs, shared vehicles and electric vehicles (EVs) might stimulate each other in a positive way, in all directions. Finally we conclude that AVs will have wider societal implications, such as in the area of land use, accessibility, social exclusion, governmental expenditures, the labor market, and the environment. The more indirect the effects of AVs are, the more difficult they are to understand. For policy making a first conclusion is that the issues of ethics, cyber security and data protection deserve way more attention than they currently get. We also conclude that future motorway network extensions might not be no-regret anymore, because of possible congestion reductions due to AVs, but also because of decreasing marginal values of time. Finally we argue that countries that introduce AVs later than other countries can learn a lot from the real world experiences elsewhere.

This chapter first systematically summarizes the most important findings and policy implications of each of the chapters "Factors affecting traffic flow efficiency implications of connected and autonomous vehicles: A review and policy recommendations" by Narayanan et al.; "Automated bus systems in Europe: A systematic review of passenger experience and road user interaction" by Heikoop et al.; "Cyber security and its impact on CAV safety: Overview, policy needs and challenges" by Katrakazas et al.; "Cybersecurity certification and auditing of automotive industry" by Mateo Sanguino et al.; "The wider use of autonomous vehicles in non-commuting journeys" by Kimber et al.; "Policy implications of the potential carbon dioxide (CO 2 ) emission and energy impacts of highly automated vehicles" by Annema; "Potential health and well-being implications of autonomous vehicles" by Singleton et al.; "Data protection in a GDPR era: An international comparison of implications for autonomous vehicles" by Costantini et al.; "Ethical issues concerning automated vehicles and their implications for transport" by Dogan et al.; "Governance cultures and sociotechnical imaginaries of self-driving vehicle technology: Comparative analysis of Finland, UK and Germany" by Mladenovi c et al.; "Wider implications of autonomous vessels for the maritime industry: Mapping the unprecedented challenges" by Ghaderi; "The potential for automation to transform urban deliveries: Drivers, barriers and policy priorities" by Paddeu and Parkhurst. Next it synthesizes the overall findings and policy implications, and discusses future avenues for policy making and research. Table 1 gives an overview of the results and policy implications per chapter. Table 1 shows that most chapters are based on systematic non-country specific literature reviews. In addition, all chapters do not only have clear Non-Specific The net energy and CO 2 emission balance for AVs seems, at its best, to be neutral, but is probably negative. However, the potentially accelerating role of AVs in relation to the uptake of electric vehicles could have the biggest impact on the CO 2 emissions per kilometer driven, but this accelerating role in the uptake of electric technology remains uncertain

Policies promoting alternative vehicle propulsion systems other than fossil fuel technologies seem most effective at the moment. Additional measures such as energy taxes, AV maximum speed limits as well as policies to curb the growth of VKTs could put in place to ensure further energy and emissions related benefits as soon as the adoption of electric AVs progresses 7

Health and well-being Deductive approach

Non-Specific AVs are likely to have overall positive impacts on some health and well-being aspects (safety, travel satisfaction, access to activities) and overall negative impacts on others (physical activity), while effects are more uncertain for other topics (urban built environments, air and noise pollution). An evolving systems approach to explore the multitude of potential impacts of AVs on health and well-being is proposed Policy measures should try to limit the possible negative effects of AVs on health and well-being, such as reduced physically-active travel (e.g., through improved infrastructure for cyclists and pedestrians) and increased vehicledistances traveled (e.g., focusing on provisions for active travel rather than placing high hopes on shared AVs which may also be associated with increased distances traveled) and urban sprawl (e.g., through compact and mixed-use neighborhoods) Milakis et al. (2017) or of the chapters in this book), we are very glad we now can conclude that the authors were able to review the literature on so many different aspects, and write a chapter that survived the review process.

Looking at the conclusions and policy implications of Table 1 a few general lessons can be learned. Starting with the research findings, the chapters make clear that the ranges in policy relevant implications of AVs, within the scope of each chapter / topic, are still relatively broad. A first reason for the wide range is the differences in the scope and demarcation of the studies reviewed. Take the impact of AVs on CO 2 emissions and road capacity as examples: it really matters if studies only include the impact of AVs on speed and distances between vehicles, or also include induced demand and changes in mode choice (from public transport and active modes to AVs).But even if the scope and demarcation is equal, estimates on impacts vary a lot, which is understandable because AVs at SAE levels 4 and 5 are not on the roads yet, so researchers cannot measure real world impacts, and have to rely on other methodologies to derive estimates. Research that is conceptually rich is more valuable for policy making. With "conceptually rich" we mean that it includes more relevant (clusters of ) variables, and both direct and indirect relationships between (clusters of ) variables. Policy makers need to be well informed about expected effects, including all factors and relationships influencing these effects. Future research aiming to estimate societally and policy relevant impacts of AVs therefore, preferably includes all dominant factors and interrelationships. Secondly, we hypothesize that context matters for the uptake, impacts, and specific system design characteristics of real world AV implementation. Take, for example, the more anti-government sentiments in the USA compared to most European countries. This could influence the support of the public for cooperative AV systems-such support probably is higher in European countries than in the USA. Another example: income levels influence the willingness to pay for travel time savings (e.g., Lam and Small, 2001) and probably also the willingness to pay for AVs: people with higher incomes are probably not only willing to pay more for travel time savings but also for AVs. So, in countries with lower incomes fewer people could be willing to pay for AVs. Such important concerns only started being addressed in 2019 via, e.g., the WISE-ACT multinational survey about AV user preferences and similar activities worldwide are essential to inform policy makers. The impacts of context factors on the design characteristics, implication options, and societally relevant effects therefore are a next promising research topic that deserves more attention. Third, and as a special case of topic 2, the chapters show that the focus in research on the global south has been limited so far. Especially because of the importance of context conclusions that hold for the global North do not necessarily apply for the global south (Thomopoulos and Nikitas, 2019) .

Extending the scope from AVs only to their position in the transport system, there is an increasing awareness that AVs, shared vehicles and electric vehicles (EVs) might stimulate each other in a positive way, in all directions, as conceptualized in Fig. 1 , at least as far as cars are concerned (e.g., Webb, 2019). Below we limit the discussion on these interactions to cars. Sharing can include both sharing cars as well as sharing rides. In the following discussion, we limit ourselves to sharing cars (or other small vehicles, like minibuses). AVs will make sharing vehicles more attractive and vice versa because AVs will likely be more expensive than conventional cars, and sharing reduces the (fixed) costs. And the combination of AVs and sharing will increase the market share of electric vehicles because people can use shared electric AVs for most of their trips at a lower cost because of the lower operation costs of such vehicles (Annema, 2020) and use internal combustion engine vehicles for only those trips for which the range limitations of EVs are problematic. EVs might stimulate sharing, at least as long as purchase costs of EVs (especially with a long range) are higher, again because sharing can decrease the fixed costs of cars. The links between AVs and EVs are a bit less obvious, but it could be that the status of cars becomes less important if people cannot drive their car anymore, so that for people who think AVs provide a lower status than a conventional car, that negative aspect of EVs might reduce. And vice versa: if people own an EV providing them less status, the barrier of being driven in an AV, and drive oneself might be reduced. In addition, speeding becomes impossible, while driving cycles and energy recovery is optimized in case of AVs (Annema, 2020) , so the strong range penalty for EVs is diminished. Fig. 1 also makes clear that the way AVs, EVs, and sharing interact, depends on policies (fiscal incentives, infrastructure provision, regulations for vehicle types), on activities of companies (developing AVs, EVs, and sharing services) and on users/consumers who buy or share and use vehicles. Many research challenges relate to the complex relationships as conceptualized in Fig. 1 .

Extending the scope to beyond the transport system, it is clear that AVs will have wider societal implications, such as in the area of land use, accessibility, social exclusion, governmental expenditures, the labor market, and the environment. The more indirect the effects of AVs are, the more difficult they are to understand. This because the impacts on the transport system need to be understood first, and these are already uncertain, and next the wider impacts need to be understood, adding more uncertainty. Yet, understanding the wider impacts of AVs on society is extremely relevant for policy making.

We now continue this chapter by focusing on policy making. Policy makers can guide the transport related developments and also these wider developments to some extent so that they work out in a societal desirable way. Although we cannot precisely predict which policies each government agency should implement in each geographical context and when, we briefly reflect subsequently on some of the roles of policy.

First, we argue that "wait and see" can be a good strategy for some possible implications of AVs, but not for all. We now reflect on what policy makers could or even should do in the short term (the coming few years), the medium term (up to 10 years) and the long term (>10 years). In the short term, the issues of cyber security and data protection deserve way more attention than they currently get, in case of cyber security at least of the automotive industry, probably also of public bodies, in case of data protection of public and next private bodies. Also ethical issues, particularly those focusing on ordinary situations, deserve attention before the (large scale) introduction of level 4 and 5 vehicles. Next, involving citizens in shaping the future of AVs deserves more attention than it currently gets, since this is the core of the demand side (Fig. 1) . Regardless of the topic, for reasons of efficiency and avoiding border issues, interstate or international collaboration with respect to policy making is to be preferred in many cases. Regarding the global maritime industry, international public bodies should by default take the lead in the process of future automation because vessels travel between countries and world regions, thus highlighting the need for even intercontinental policy making collaboration.

Extending the time horizon a bit further, say the next decade, an important lesson is that in developed countries with more or less complete motorway networks, future motorway network extensions might not be no-regret anymore. Many regions and countries face a diminishing population growth, if not: a decline in population. Car ownership levels seem to be not very far from saturation. If there are no (major) missing links, future extensions of motorways mainly aim to reduce congestion. But if car ownership does not increase substantially, and if more people can avoid the rush hours because they can work online as widely proven during the COVID-19 restrictions, the costs of motorway extensions may exceed the benefits. AVs could further strengthen this process, first of all because they may reduce congestion levels (despite induced demand) and secondly they will reduce peoples' willingness to pay for travel time reductions (Zhong et al., 2020) because of the possibilities for other categories of time use in AVs.

On the longer term, it is important to realize that the introduction of SAE level 4 and 5 vehicles will not happen at the same moment in time across the world and the transition to AVs will vary widely. So, countries that introduce AVs later than other countries can learn a lot from the real world experiences elsewhere. We expect research and business interest on the many effects of AVs to boom after their widespread real world introduction. Respecting the differences in context as addressed above, we do think that many lessons can be learned, in many areas, such as the process of decision making, acceptance and acceptability, the pros and cons of design alternatives, the societally relevant implications of AVs (accessibility, safety, security, the environment, land use, impacts on other modes), and the evaluation of policy alternatives (including all relevant effects, but also fairness issues). So, learning from real world experiences it is both a very important topic for research as well as for policy making and AV trials pave the way in this respect.

Policy implications of the potential carbon dioxide (CO2) emission and energy impacts of highly automated vehicles

The value of time and reliability: measurement from a value pricing experiment

Policy and society related implications of automated driving: a review of literature and directions for future research

Smart urban mobility futures, editorial

The future of transport: literature review and overview

Will autonomous vehicles change auto commuters' value of travel time?

We conclude that it is still unclear how autonomous vehicles will shape the future of transport systems, and what the policy relevant effects of AVs will be. But it is clear that the role of AVs in the vehicle fleets will have many policy relevant implications, some of which did not receive a lot of attention by policy makers yet. On the other hand, policy makers can also influence the uptake of AV. With this book, we review the AV literature on a wide range of topics. We hope it provides a source of inspiration for researchers, policy makers, and planners.