System Innovation in Passenger Transportation with Automated Minibuses in ITS: The Citizen-Centric Approach of AVENUE

Abstract

The first chapter explores three pathways for incorporating automated vehicles (AVs) into future mobility ecosystems: privately owned AVs, robotaxis and automated minibuses in a Mobility-as-a-Service (MaaS) and (later) in an Intelligent Transport System (ITS). The chapter emphasises that automated minibuses, when seamlessly integrated into a MaaS, could emerge as pivotal “game changer”, complementing and fostering other means of transport, in particular mass transport. Integrating in a next step AVs within an ITS could further make it possible to use mobility data and artificial intelligence (AI) to improve the transport system to a higher level of mobility evolution. By employing both fast and slow closed AI loops, it envisions a transport ecosystem that not only operates more efficiently and flexibly but is also consequently citizen-centric and addresses sustainability challenges. The successful implementation of this concept hinges on open data, Application Protocol Interface (APIs) and the potential of AI to create a self-learning transport system to serve the general interest. A sustainable and citizen-centric mobility is thus possible without a coercive (technology push) transport policy. Instead, it champions an approach that increases the transport options and enhances the provided mobility services (demand & attractiveness pull). Depending on the local governance, even privately owned or privately shared cars can be part of the model, justified, e.g. in areas where mobility infrastructure deficits loom large or when travel time is too high. This chapter therefore forms the basis for a vision to redesign our mobility ecosystem and marks the beginning of a disrupting system innovation, where integrated sustainability and citizen centricity reshape the nature of mobility.

1 Introduction

In this chapter, we will discuss and present the vision of AVENUE for a medium-term horizon (2030) of future public transportation. Our aim is to integrate automated minibuses into transportation systems available in a city. A central issue in this vision is the mobility needs of citizens which have to be satisfied in an optimal way: an abundant service offer portfolio with a high variety of private and public mobility modalities combined to one individualised intermodal trip. Automated minibuses play a central and critical role in this model: (1) as a feeder for the other means of transport, in particular mass transit, and (2) as a mobility complementor for the entire transport system in case of weak or not existing transport offerings (e.g. rural or remote areas), incomplete transport chains (e.g. for tangential connection) and disturbances of the transport system. Public as well as private transport operators (PTOs) are forming an enhanced public-private partnership (PPP) to utilise and synergise their multifaceted complementarity in this MaaS (Mobility-as-a-Service) vision. The application of advanced self-learning systems (e.g. based on human and/or artificial intelligence) could in a further stage let this visionary MaaS concept become a self-learning ITS (intelligent transport system). This disruptive transport system innovation could create best citizen-centric transportation with optimised and balanced private and public value. Further, it improves economic performance, supports environmental benefits and also increases acceptance by citizens, as this intelligent transport system is enabling smooth transition towards efficient, safer, nearly carbon-free, inclusive and sustainable transport in cities.

The AVENUE vision of the future mobility system builds upon insights out of the AVENUE sustainability assessment and the conceptual transition studies and recommendations (see Chap. 19). The following research questions have been guiding the analysis:

How can citizens’ mobility needs, stakeholders’ interests and general interest be optimally satisfied with AVs?

How can AVs and automated minibuses be integrated in a future transport system in the city in a meaningful and optimal way?

How should EU regulations evolve to support a reliable transport system that balances economic values of all stakeholders, regards privacy requirements and ensures safety and security issues and governance; general interest on social, economic and environmental level; freedom; and sovereignty (mobility data, technological sovereignty)?

These questions will be answered in the next sections of this chapter. The integration of automated minibuses into an intermodal and citizen-centric perspective is discussed in Sect. 18.2; the integration of the automated minibus within a MaaS system is designed in Sect. 18.3; and a deployment of automated minibuses integrated with artificial intelligence (AI) and intelligent transport systems (ITS) is proposed in Sect. 18.4. After a discussion in Sect. 18.5 about limitations of the vision, conclusions are presented in Sect. 18.6.

In terms of methodology, numerous interdisciplinary workshops have been organised with scientists, public transport operators, original equipment manufacturers (OEMs), industrial companies, public transport authorities and other stakeholders, both within and outside the project (see Chaps. 9 and 10). Various alternatives were developed and evaluated using a holistic approach. Experts from a wide range of fields, including automation engineers, IT specialists, economists, social scientists, psychologists, environmental scientists and lawyers, contributed to the discussions. The aim of this multidisciplinary dialogue was to ensure a comprehensive understanding of the interrelationships between different impact areas, potential rebound effects and conceivable futures. The results have been accepted by all project partners and the advisory board of AVENUE. A more comprehensive summary of this chapter has been published in 2023 by Fournier et al. (2023).

2 Integration of Automated Vehicles in Future Urban Transport Systems: From Product Innovation to Citizen- and Purpose-Centric Transport System Innovation

Automated vehicles can be integrated into future transport systems in different ways. Among others, three use cases are possible (based on Grisoni & Madelenat, 2021; Heineke et al., 2019; UITP, 2017):

• Private automated vehicles (AVs): Like the currently dominant choice of privately and individually owned cars, AVs can be integrated into an urban mobility system. An integration of AVs in such way would lead to an incremental improvement of the individually used vehicle technology from today’s driver-assisting systems to fully automated driving systems (Fraedrich et al., 2015; Deloitte, 2022).

• Robotaxis: The definition for robotaxis inhere follows the description as shared automated vehicles (Fagnant & Kockelman, 2018). This definition is quite general and does not distinguish between vehicle sharing (car-sharing) and service sharing (ride-sharing, ridepooling, ride-hailing etc.; see MOQO, 2020). Within the understanding of the AVENUE project, robotaxis mostly appear as driven by private operators with commercial interest. They are meant to be not or just poorly integrated with the public transportation and therefore stand consequently mostly in direct competition to it. They are further mainly used for single ridership although those capabilities would support ridepooling services. Examples for robotaxis are Waymo, Uber Pool or Moia in the German city Hanover (see, e.g. WEForum, 2020; May et al., 2020; Clewlow & Gouri, 2017; Merlin, 2017; Jaroudi et al., 2021; https://www.moia.io/de-DE/stadt).

• Automated minibuses in a MaaS system: Another pathway of integrating AVs into an urban transport system is the usage of automated minibuses as one optional transportation mode in a Mobility-as-a-Service (MaaS) system (see Fig. 18.1). The automated minibuses within the scope of this AVENUE project are defined as fully electric, automated vehicles without a driver [SAE Level 4 of automation; see SAE International (06-May-23)] and with a transportation capacity of up to 15 passengers. This vehicle can be combined with a means of either public (e.g. train, tramway, traditional or even automated fixed-scheduled bus line etc.) or private (e.g. bike-sharing, car-sharing etc.) transportation to a seamless journey.

Fig. 18.1

The integration of the automated minibus into a MaaS system

The integration into a MaaS is meant to pool multiple trips of users that request an on-demand, door-to-door service. Furthermore, the transportation service can cover mobility gaps in the public transportation network (Sochor et al., 2016; Shen et al., 2018). Aside from their capability to fill the gaps of public transportation networks, automated minibuses via MaaS enable the collaboration of private and public transport operators (PTOs) to provide one individual intermodal trip to citizens that optimises travel time, costs and/or environmental impacts (Sochor et al., 2018; Kamargianni et al., 2015; Vleugel & Bal, 2017). MaaS in this case provides citizens the freedom to select the optimised transportation modes for its travel depending on its user profile, persona and even external factors (such as weather, etc.) (see Fig. 18.2). Technically, MaaS systems rely on the commitment to a standardised application programming interface (API) for all participating transport operators to offer seamless intermodal trips to its users (Sochor et al., 2018; Kamargianni et al., 2015).

Fig. 18.2

Illustration of various citizen journeys as combinations of all means of transport offered to a citizen by the MaaS concept

These described three pathways each have advantages and disadvantages. In the development of privately owned AVs (1), the service can be seen as very convenient due to exclusive usage, achieving probably the highest degree of individuality and freedom. On the other hand, the costs of individual mobility would be very high for the users and not affordable for most citizens. A mass diffusion of this solution is accordingly not expected yet. It is also assumed to be one of the reasons why the automotive industry currently focuses on SAE Level 2+.¹ Nevertheless, in case of a wider diffusion of the privately owned fully automated vehicles, the convenience of these vehicles would encourage more vehicles on the road and lead to higher traffic volume and congestion costs (externalities). A weakening of public transport would be likely as well with growing market share (Fournier et al., 2020; Milakis et al., 2017).

It is also expected that robotaxis (2) will be widely accepted by citizen due to their affordability, availability and comfort, but they may not complement other means of transport in a synergistic way (Korbee et al., 2023). Robotaxis could in particular be in competition with and may substitute other means of transport, including cars, walking and public transport (WEF Forum, 2020). Individual unimodal mobility could be more convenient and strengthened. Simultaneously, the transportation system as a whole would become less efficient and sustainable, also resulting in increased external costs due to additional traffic and displaced public transport (Meyer et al., 2017; Niles, 2019; May et al., 2020; Clewlow & Mishra, 2017; Rayle et al., 2016; Merlin, 2017; Childress et al., 2015; WEForum, 2020). Market fragmentation and asymmetric competition between this market leader and other private and public mobility providers could also result in the emergence of a dominant robotaxi provider (Hassani, 2018; Niles, 2019). For example, if private transport platforms are not providing open mobility data in the same way as public transport providers, data asymmetries will arise and lock in the customer in the robotaxi ecosystem. As a result, a dominant market position for private transport providers and the associated “winner-takes-it-all” effect could appear and bear the risk of rising monopolies within the transport system (Hassani, 2018; Hoffmann, 2021; Cabral et al., 2021). Similar situations can be observed for other online services. There is currently an oligopoly dominated by a group of companies known as GAFAM (Google (Alphabet), Apple, Facebook (Meta), Amazon and Microsoft) or now MAAMA (Meta, Alphabet, Amazon, Microsoft and Apple) (The Economist, 2022; European Commission, 2021; Cabral et al., 2021; Toledano, 2020).

Integrating automated minibuses into a Mobility-as-a-Service system (3) could finally be the game changer: public and private transport could become individualised, by offering, e.g. first and last mile trips, as well as a critical mobility gap-filler for the feasibility of seamless citizen journeys. The concept of providing better transport services, improved efficiency and flexibility in the transport system and positive externalities in this system puts citizens at the centre (Coyle et al., 2020; Coyle & Diepeveen, 2021). Automated minibuses in MaaS create multifaceted value for the stakeholders involved, resulting in a win-win situation for all.

Automated minibuses in a MaaS thus seem to offer more benefits for the stakeholders (travellers, TPOs, cities) and for the general interest (Becker et al., 2020). The integration of automated minibuses into public transport in urban, suburban and rural areas and into a MaaS system (our vision) is highly beneficial and advantageous for its stakeholders, in particular for citizens and private and public mobility providers as well as cities and regions (administrations/government) in general. This integration is understood as follows:

A customer-/citizen-centred approach allows the user to orchestrate an individually selected combination of all different types of transport into a gap-less and reliable trip that meets the different preferences of all population groups. As such, automated minibuses in MaaS are seen as a transformative force in urban mobility, capable of providing personalised public transport that is diverse, accessible (including for PRM), inclusive and affordable. This could be an attractive and beneficial alternative to the private car transport. “The results of a representative survey among 1816 citizens (of which 1526 have privately-owned vehicles) in Lyon, Copenhagen, Luxembourg and Geneva confirm that 45% of car drivers are ‘willing’ (22%) or even ‘very willing’ (23%) to give-up using their own car to use automated minibuses to bridge the first and the last mile if this were available. If the service is on-demand and door-to-door, the acceptance could be even 52% (respectively 28% and 24%) in total” (Fournier et al., 2023, p. 3; Korbee et al., 2023).

For the transport operators, the “automated minibus in a MaaS concept provides a better exploitation of existing capacities and resources and a positive experience of fair competition” (Fournier et al., 2023, p. 3). In particular, automated minibuses are an important factor and a missing complementarity modality in making large mass transit systems such as rail, metro, bus etc. more attractive, more efficient and thus more profitable. By taking this perspective, all offered means or modalities of transport are potentially available for an individual trip design. The mobility choice and subsequently the modality choice of the citizen (the end user/customer as central “king/queen of the system”) are considered as right and definitive and not limited by the transportation portfolio of a single (private or public) transport provider (Fig. 18.2). The automated minibus integration into the urban mobility system contributes to avoid thereby the “winner-takes-it-all phenomenon”. On contrary, a win-win situation for all stakeholders in terms of value creation, fair value sharing, sustainable mobility and long-term purpose is achieved, as stands the Purpose Economy (Hurst, 2016; Business Roundtable, 2019).

For the cities (administrations), the designed concept would offer a significant optimisation of mobility excellence (efficiency, effectiveness, safety, quality of transport, user-friendliness etc.), serving the wider public interest and providing financially, environmentally and socially sustainable transport. The complementary use of public and private transport together, with automated minibuses, would therefore be able to make travel less costly, more competitive and more predictable and would support a higher level of resilience in the overall mobility system, decreasing external impacts of mobility (Alazzawi et al., 2018; Vleugel & Bal, 2017), and this will contribute to a new paradigm of sustainable mobility and to the Sustainable Urban Mobility Plan (SUMPs) that involve citizens and stakeholders in a goal-oriented approach that serves public interest. Moreover, it will increase the public acceptance through higher effectivity and flexibility in the whole system (Hurst, 2014; Nemoto et al. 2021; Korbee et al., 2023). Based on the results of the aforementioned survey, a transformation of the mobility paradigm could be realised without a coercive policy (push strategy). An improvement of the offer would change the behaviour (pull strategy).

Open data and interaction platforms and open APIs are prerequisites to enable access to all the means of transport, their interoperability and intermodality, to ensure balanced win-win situations among all stakeholders and thus a kind of democratisation of the automated minibus in a MaaS ecosystem (ERTICO – ITS Europe, 2019; Coyle et al., 2020). Open data and open APIs are further a condition to enable innovation and evolution of mobility concepts and develop the MaaS towards an intelligent transport system (ITS). Using data makes it easier to generate information and knowledge and provide the basis for decisions to improve the transport system and to generate positive externalities. The transport system could become a self-learning system with the support of artificial intelligence and humans (see details in chapter below). A “circulus virtuosis” (virtuous cycle/loop) and positive externalities could be created, to guarantee safety and to promote global improvement. To achieve this ITS vision, several improving loops have been identified which could be implemented and which make the transport system more resilient and future oriented (see Sect. 18.4).

The third path, “automated minibus in MaaS”, can thus provide the best public value and therefore represents the long-term vision of AVENUE. It is the main enabler of a holistic system innovation and is not only regarded as a classic product (e.g. automated minibuses) innovation. However, this is not to underestimate the importance of automated minibuses as a product innovation: combining automated minibuses and MaaS represents a disruptive game changer to take the entire transport system to a higher level of mobility evolution. This evolution will be described in the sections with AV in MaaS (Sect. 18.3) and the next progress with AV in ITS (Sect. 18.4).

3 Integration of Automated Minibuses in a MaaS System: A Citizen- and Purpose-Centric Approach

The concept of purpose economy refers to a new way in which people and organisations create value and define the principles for innovation and growth (Hurst, 2016). The value lies in establishing purpose and creating meaning value for employees and customers beyond their own benefits, but aiming personal and community development (Hurst, 2016).

The Purpose Economy creates purpose for people. It serves the critical need for people to develop themselves, be part of a community, and affect something greater than themselves. (Hurst, 2016)

The “citizen-centric automated minibus in MaaS” approach is in line with the purpose economy and targets a mobility system built by all and for all. The concept is focusing on the compilation of various transportation service modalities to one single seamless trip on demand and, according to the preferences of the citizen, including related services like trip planning, ticketing and others (European Commission, 2016).² Citizens have the choice to choose from various transportation options for their journey within or outside the city or even the country. To support them in their decision-making, the MaaS platform provides various functionalities such as simulations of time, distance, cost and CO2 footprint, as well as options for comparison. Once the citizen has made their decision, they are requested by the mobility integrator app to make the booking and place a binding order. The data from the selected trip option is then sent back to the MaaS platform and onwards to the private and/or public transport platform and transportation provider or possibly to other MaaS system platforms. “The purpose is to better satisfy citizen mobility needs and evolve towards a sustainable mobility system, create value for all stakeholders, contribute to sustainable societal transformations, and provide an answer to societal challenges (the so-called ‘purpose’)” (Fournier et al., 2023).

To achieve this, it is necessary to consider factors such as the availability and interoperability of both hardware and software devices, which are provided through standardised interfaces (APIs), as well as management and coordination software and services offered by an open ecosystem of service aggregators and other intermediaries. By combining digitalisation with a citizen-centric approach, the distinction between private and public transport operators will be blurred, leading to the introduction of coopetition (Haan et al., 2020). Coopetition is a new perspective that involves both competition and cooperation existing together, utilising complementary resources cooperatively (Liu et al., 2015). Additionally, democratised and federated governance of ecosystem stakeholders could be established, where beneficiaries are determined by their value-creating contributions to the defined purpose (Gassmann & Ferrandina, 2021; Schmück, 2022).³

Figure 18.3 depicts the integration of automated minibuses in a citizen-centric MaaS, with the citizens at the core. A variety of transport means is available for the citizens to choose from, as depicted in the middle circle in Fig. 18.3. The transport options are provided by various stakeholders (the actors named in the yellow blocks, at the outer circle of Fig. 18.3). Other governance scenarios to coordinate the different stakeholders are of course possible depending on market (private, public and coopetition) and data schemes (private or open data, open interfaces and protocols). These scenarios have been developed by UITP 2019a (UITP, 2017, 2019; Ertico ITS-Europe 2019; Capgemini, 2020). The scenario which fulfils best the needs of the customer, enables fair competition through avoiding the “winner-takes-it-all” phenomenon, enables positive externalities and fulfils sustainable and societal goals is the citizen-centric MaaS approach chosen by AVENUE. It defines a purpose, balancing the interest of the numerous stakeholders to serve the general interest.

Fig. 18.3

Automated minibuses integrated into a citizen-centric MaaS

Important for the success of the AVENUE vision is the full integration of automated minibuses into the described citizen centric MaaS system. To ensure a fair understanding of how automated minibuses could be integrated in the MaaS system, a typical booking process which integrates automated minibuses is set up with the following five steps:

These steps are detailed and explained below; each of the steps is visualised by a figure (Figs. 18.4–18.8). The step-by-step description starts with a current, conventional outline of the system, in which the citizen has to enquire about options and has to make the decisions themselves, building up to a fully integrated system. With each of the steps, more complexity (through the integration of stakeholders and additional platforms) is added to the MaaS system.

Step 1: The first step in a MaaS booking process consists of the citizen choosing the destination of its trip. All combinations of several means of transport, offered by both public and private operators, can be chosen. Public transport operators provide demand-responsive transport (DRT), which aims to optimise the match of transportation supply with transportation demand efficiently and uses automated minibuses, for the first and last mile and as mobility complementor, and conventional public transport (CPT) services, like bus, metro and train. Private transport operators provide transportation modes like taxi, ride-hailing like Uber, car-sharing, bike-sharing and other micro mobility devices like e-scooter.

Traditionally, the citizen is forced to select and schedule separately every modality from private or public transport operators or a combination of it on his own and singularly according to his trip planning from A to B and related modality compilation, as is visualised in Fig. 18.4. This also includes a multi-factor optimisation of the modality-portfolio sample regarding time, cost, travel requirements (e.g. luggage, wheelchairs), personal preferences and others, including inherent error-proneness.

Fig. 18.4

Vision of the mobility of the future: Step 1

Step 2: In the next step, the position and availability of an automated minibus is documented together with other public and private means of transport on the MaaS platform. This connective platform is visualised as the yellow bar in Fig. 18.5. To optimise the decision and selection process, these tasks are transferred from citizens to specialised intermediaries: a MaaS data service platform is installed. This data service platform is designed, managed and controlled by a MaaS data and service platform operator and governed by Public Transport Authorities (PTA) as regulation bodies, for instance, by extending the existing National Access Points (Art. 3 Delegated Regulation 2017/1926 of 31 May 2017). In this platform all necessary data about modality and transportation provider offerings and related metadata from private and public transport operators are integrated, managed and exchanged. In particular private transport operators should provide all of their journey-relevant data which are necessary to offer a seamless intermodal trip to the customer. Providing data avoids an information asymmetry with public transport operators which are per law (Art. 4 and 5 Delegated Regulation 2017/1926 of 31 May 2017) obliged to share static and eventually dynamic travel data. This information asymmetry could lead to the aforementioned “winner-takes-it-all phenomenon” and to a dominant position of the private MaaS provider. Private transport operators are already obliged to provide the data foreseen under Annex I of Delegated Regulation 2017/1926. In addition, the private transport operators might receive a reasonable and cost-based compensation for providing this data: “Any financial compensation shall be reasonable and proportionate to the legitimate costs incurred of providing and disseminating the relevant travel and traffic data” (Art. 8 (4) Delegated Regulation 2017/1926 in fine).

In the case of an integration of automated minibus services in this MaaS platform, information about current location of the automated minibus and the status of occupation by passengers should be made accessible. However, this should not include personal data or only depersonalised data in accordance to privacy and security regulations.⁴

For this reason, only the status and telemetry of the data are provided to the private or public transport operators. To make the data available to all users in a standardised and quality assured form,⁵ an open application programming interface (API) is required, as well as commonly agreed standards. These APIs⁶ enable authorised applications to communicate, use one another’s functions and exploit data sets provided by other applications or databases (Matthes & Bondel, n.d.). They make it in fact possible for the stakeholder to communicate with other stakeholders in order to provide a uniformed offer to the MaaS user (citizen). APIs are considered to be the “connective tissue of the cloud”; they are essential to integrate the transport system. However, the urban mobility ecosystem is still very fragmented (Bestmile, 2020). To overcome this barrier, initiatives envision setting standards of open APIs to enable mobility providers to integrate services. Examples of such an initiative are the projects MyCorridor, MaaS4EU and IMOVE. These projects foresee the use of a “common language” to designing a transport service API, comprehending “the use of communication protocol and data format to security standards, basic methods and service calls, responses and general behaviour of an API” (MaaS Aliance, 2019). Another initiative, the Information Technology for Public Transport (ITxPT), focuses on open standards and procedures for integrated Information and Technology Systems for public transport (Rogg, 2021). Therefore, open interfaces, protocols and standards are key factors for MaaS ecosystems.⁷

Consequently, an open standardised API⁸ concept is initially required for a functioning MaaS platform. It has to connect the vehicle modality and its providers, the transport operators and the exchange platform and its platform operator. Additionally, they are utilised by further relevant parties of the ecosystem, like other MaaS systems (e.g. when trips are overreaching MaaS systems of other cities), or regulatory tasks of PTAs and city infrastructure management. For the automated minibus, its MaaS relevant information is transferred to the open data MaaS platform to be exchanged between the public transport operator (PTO) on the one side and the “MaaS data and service platform operator” on the other side. In addition to the exchange of data, open APIs ease the management of automated minibuses from different OEMs. This allows a flexible use of the individual vehicles as well as the entire fleet. A functioning system results in lower costs of operation as less vehicles are needed.

An important reason for an open API for the “MaaS data and service platform operator” is that in case a Public Transport Authority (PTA) desires to redesign the “MaaS data and service platform” itself, there is no barrier due to dependencies with the data model of the original equipment manufacturer (OEM).

Fig. 18.5

Vision of the mobility of the future: Step 2

In Step 3, the availability of other means of transport of other MaaS in a particular country or other countries of the EU is documented in the MaaS platform. The MaaS platform provides further data to stakeholders (city, PTA, PTO etc.) which can be used to improve the transport system (city planning, traffic, incidents, accidents, fleet management, vehicles safety). This is visualised in Fig. 18.6 by adding yellow arrows connecting the MaaS platform to the private operators, other MaaS systems and other stakeholders.

Fig. 18.6

Vision of the mobility of the future: Step 3

Step 4 integrates MaaS mobility service aggregators, as intermediary organisations between public and private transport operators and the citizen to provide the existing choice for the requested passenger journey. This new layer is introduced by local, national or supra-national private or public mobility aggregators and even by PTAs themselves. PTAs regulate, delegate and supervise this market. The mobility service aggregators are designed to analyse and process citizen profiles (in an anonymous manner, under Art. 4 (4) GDPR) and trip requests as well as general data about transport providers, vehicles and services and specific data about current positions and status of vehicle usage from public and private transport operators. Hence, this additional layer integrates information from both sides: trip request or transportation demand, and transportation offering services or vehicles, in combination with further information from other MaaS ecosystems and city infrastructure.

All relevant data are utilised for compilation of alternative trip options, as displayed in Figs. 18.2 and 18.7. Alternative trip options are selected based on the trip request, including individual preferences, but have to be protected by adequate privacy and security concepts and measures. These concepts must be completed/developed and deployed. Especially in the case of disruptions, critical scenarios, and accidents, these incidents will be analysed and evaluated using concepts and tools of self-learning systems and artificial intelligence (AI).

By applying algorithms of artificial intelligence (AI) to a MaaS ecosystem, the infrastructure and the value chain of transport companies can be significantly improved according to the needs of the citizens. Furthermore, with the support of AI all kinds of incidents, risks or problems within the MaaS ecosystem can be identified, tracked, traced and analysed in order to define and conduct strategies and measures for detecting, preventing and solving them.

AI applications comprise multiple functionalities for supporting success critical tasks, such as finding patterns and new insights, making predictions, interpreting unstructured data and interacting with the physical environment, their machines and humans (Scherk et al., 2017). One of the central topics of AI are self-learning algorithms of systems (e.g. based on neuronal networks) with the goal of finding independent solutions to new and unknown problems (Scherk et al., 2017).

Fig. 18.7

Vision of the mobility of the future: Step 4

To summarise, all potential trip-relevant information from the MaaS stakeholders has to be available, and the quality of related data has to be assured (i.e. static and dynamic real-time data, data validity, data privacy and security). This information should be provided to the MaaS mobility service aggregators, who are able to aggregate all transportation needs or requests from citizens on the one hand and all transportation modality offerings (including automated minibuses) on the other hand and match them in an intelligent way (supported by AI algorithms) according to citizen preferences (first priority, e.g. time efficiency, cost-efficiency, carbon footprint) and offering goals (second priority, e.g. route efficiency, vehicle utilisation) in order to subsequently provide the information of alternative transportation options to each citizen.

Fig. 18.8

Vision of the mobility of the future: Step 5

In Step 5, individual citizens can select their preferred transportation option for his trip and can place a booking and ordering with payment. The automated minibus is planned and reserved for the citizen. The mission is provided to the vehicle, which requires a return flow of the data (indicated by the blue dotted arrows in Fig. 18.8). The citizen has the choice to select between different transportation options for his trip from A to B in or outside of the city or even outside of the country. Hereby various valuable functionalities (like simulations about time or distance or cost or CO2 footprint, together with option comparisons) are provided to the citizen. After making this decision, the citizen is requested by the ticketing app to make the booking and place the binding order. After ordering, the data from the selected trip option is sent back to the MaaS data and service platform and hereafter to the private and/or public transport platform and transportation provider or possibly to other Maas system platforms.

As mentioned above, automated minibuses in MaaS could be used to generate information and knowledge and provide in this way the basis for human or AI-based decisions to enable a self-learning resilient transport system. This will be the endeavour of Sect. 18.4.

4 Automated Minibuses Integrated in Intelligent Transport Systems: A System Innovation Approach for a Resilient Self-Learning Citizen- and Purpose-Centric City Transport

The concept of “citizen-centric intelligent transport system” combines the intelligent transportation system with the citizen-centric approach and sustainability. Intelligence in this context is provided by artificial intelligence (AI) technologies. An ITS aims to provide services relating to different modes of transport and traffic management, enabling users to be better informed and make safer, more coordinated and “smarter” use of transport networks. They include advanced telematics and hybrid communications including IP-based communications as well as ad hoc direct communication between vehicles (V2V) and between vehicles and public and private infrastructure (V2X) (ETSI, 2021). Another keystone refers to a self-learning system with data-driven approach and AI for the development of intelligent transport system for smart cities and sustainable mobility (Iyer, 2021; UNESCO, 2021). AI and digital transformation could support to identify common patterns of mobility and quantify the crucial factors affecting the efficiency of the whole system (Lucca, 2022).

Figure 18.11 depicts the citizen-centric intelligent transport system. At the core are the citizens as it follows a citizen-centric approach. A variety of transport means is available for the citizens to choose from; these are depicted in the middle circle in Fig. 18.11 in a similar way to Fig. 18.5. The transport options are provided by various stakeholders (the actors named in the yellow blocks, at the middle circle of Fig. 18.11). In addition to Fig. 18.5, data are used to generate information and knowledge about and within the transport system. An intelligent transport system can learn and improve through experience gained by diverse data; this procedure is visualised by seven AI loops in the outer circle. Highly relevant is that those positive externalities only work in closed loops. This means that collected data are not enough since there are still decisions that need to be taken to close the loop (e.g. to improve the transport system). This is the case when automated vehicles have “learned” from data how to avoid an obstacle or a critical scenario in the city and the improved drive algorithm and eventual related hardware have been uploaded (fast closed loop). This can happen as well when a city chooses to change the traffic rules on places with a high level of safety problems (slow closed loop).

On-demand and door-to-door data, compliant with the General Data Protection Regulation (GDPR), could be delivered with unprecedented precision in quality and quantity. In fact, this data could be used to generate information, including decisions and knowledge about the transport system, enabling continuous improvement of the urban mobility system through state of the art, self-learning human and/or AI loops and controls. This is visualised by the loops in the outer circle of Fig. 18.11, which show the loops in detail. The loops in the outer circle of Fig. 18.11 illustrate this as they are designed around stakeholders to implement lessons learned to improve locally and globally. Data and information are generated by all components of the transport system such as the automated minibus, infrastructure, platform, applications and especially the user. These generated data add in return value to the other stakeholders which closes the loop and results in an optimisation of the overall transport system (ERTICO – ITS Europe, 2019).

The functional context of an intelligent transport system is systematically described by the following step-by-step illustrations.

Depicting the seven relevant loops (as indicated in Fig. 18.9):

• Loop 1: OEM centred (within building blocks (BB) (Fig. 18.10)

• Loop 2: PTA, city–PTO centred (between BB) (Fig. 18.11)

• Loop 3: PTO–OEM centred (between BB) (Fig. 18.12)

• Loop 4: PTA, city OEM centred (between BB) (Fig. 18.13)

• Loop 5: PTA, city centred (within BB) (Fig. 18.14)

• Loop 6: PTA–MaaS mobility service aggregator centred (between BB) (Fig. 18.15)

• Loop 7: MaaS mobility service aggregator–citizen centred (between BB) (Fig. 18.15)

Fig. 18.9

Citizen-centric intelligent transport system (ITS)

Fig. 18.10

Automated minibus integrated in ITS: Loop 1

Fig. 18.11

Automated minibus integrated in ITS: Loop 2

Fig. 18.12

Automated minibus integrated in ITS: Loop 3

Fig. 18.13

Automated minibus integrated in ITS: Loop 4

Fig. 18.14

Automated minibus integrated in ITS: Loop 5

Fig. 18.15

Automated minibus integrated in ITS: Loops 6 and 7

Each of the loops will be described and explained below. The descriptions are accompanied by visualisations (0, Fig. 18.15) of the seven self-learning loops to achieve a complete self-learning ITS ecosystem. Each loop builds upon the previous system. The coding within the graphical representations is defined in the following way:

Arrows have different meanings: Dotted arrows indicate the data flow; plain arrows indicate the information flow and the related decisions. Yellow arrows indicate the outward flow of data or information, blue arrows indicate the return flow and the clothing of the loop; a data exchange loop is represented by the combination of outward and return flow. The determined loops can be fast (because, e.g. just one stakeholder or one building block like an OEM is involved), medium or slow (when a consent between stakeholder (building blocks) or laws/vehicle specification/vehicle certification must be developed and published).

Loop 1, as depicted in Fig. 18.10, is an OEM-centred fast loop. Three data flows and one information flow are emitted by the automated minibus to the surrounding systems. The closed data loop on the left in Fig. 18.10 illustrates the data flow between the automated minibus and the road infrastructure and environment, generating both sensor data from the automated minibus’ on-board and roadside systems. Characteristic for the second data stream is the long-time data storage in a virtual black box similar to those used in air transport. The automated minibus uses the backend systems to store permanently real-time data about recorded incidents and accidents. Access to this real-time data is also available for the PTAs (Public Transport Authorities), who then can optimise service quality and safety (Loop 2) for cities, PTOs (Public Transport Operator) and OEMs (Original Equipment Manufacturer). A further beneficial data flow is the direct connection to the PTOs. This linkage allows the PTO to derive route and service optimisations based on the real-time data about incidents or accidents. Further recommendations to optimise the road infrastructure, traffic rules or change of transport means become possible based on this data flow. Data of the vehicle can finally be used to monitor and optimise the charging of the vehicle. Maintenance costs of the vehicle can be improved as well to move from a curative (in case of breakdown) to a preventive (plan is better than cure) and later to a predictive (forecast breakdowns and predict a malfunction) maintenance.

Loop 2—PTA, city–PTO centred: The second loop, as depicted in Fig. 18.11, is PTA, city–PTO centred.

To ensure a trustworthy accessibility of the generated data for PTA, an intermediary scientific service provider (trusted AV third party) shall be established. This trusted third party gets certified and operates on behalf of the PTA as non-biased, neutral and objectively acting institution. As neutral institution it could analyse, process and forward all data about any incidents and accidents to the PTA and propose improvements but also support for juridical treatment when necessary. They could also proof the AI used by the OEM for their vehicles and transmit this information to the PTA and provide recommendation for improvement measures for regulation.

Based on the PTO-generated data, the PTO is as well able to provide learnings and suggestions for technology or regulatory improvements to the PTA. Subsequently, the data cycle of permanent information between the PTO and PTA to drive optimisations on the regulations for service quality and safety is captured in Loop 2 and also with the connection to OEM (Loop 4) (see in Fig. 18.13).

Loop 3—PTO–OEM centred: The third loop, (see Fig. 18.12) is PTO–OEM centred, and further completes the intelligent transport system.

Critical automated minibus scenarios are transferred from the OEM to the PTO empowered by AI algorithms. Subsequently, the data is then returned back from the PTO to the OEM. This provides the OEM with information on adjustments to the transport specification such as capacity, range, availability and persona profiles, adjustments to the vehicle specification and new and passive safety requirements. This enables the OEM to analyse the data and provide an optimised utilisations of automated minibuses by taking into account the new road safety requirements.

Loop 4—PTA, city–OEM centred: The fourth loop, depicted in Fig. 18.13, is PTA–OEM centred. Based on mobility data from the trusted AV third party, PTO and the PTA, PTAs can create and improve OEM certification requirements and (type) approval requirements and define automated driving regulations (e.g. vehicle architecture, AI algorithms etc.). For example, new active and passive safety requirements can be introduced from accident analysis and lessons learned coming from the third party.

Loop 5—PTA, city centred: The fifth loop, depicted in Fig. 18.14, PTA, City centred. This data cycle shows a slow closed loop of PTAs receiving information and data provided by all stakeholders and using them as a basis for improving urban infrastructure and traffic management, updating, monitoring and mobility (re)planning.

Loops 6 and 7 complete the ITS by adding both the citizens and the MaaS data and service platform. The two-way information sharing between the PTAs and the MaaS mobility service aggregator is key:

The data provided by the mobility aggregator and the MaaS data and service platform provider on one hand informs the PTAs anonymously about the mobility demands and patterns of the citizen (attitudes and behaviours).

In return, the PTAs are in a better position to enhance the mobility system and the mobility infrastructure (Loop 5) aimed at inducing positive and reducing negative externalities of mobility. Eventually, the data collected by the mobility aggregator will allow the provision of customised unimodal or intermodal mobility alternatives to optimise the customer’s choice according to their profile, persona or even the weather. Trip duration, expense and environmental effects provide a foundation for decision-making. To ensure a positive mobility experience, trust and satisfaction, safe travel and acceptance, these last loops are crucial.

Through the seven loops, the integration of automated minibuses in the future transport system could thus be beneficial for all the stakeholders, lower negative externalities, provide positive externalities and satisfy better the citizen needs and their acceptance. All the means of transport can be offered and combined to a seamless offer which is nearly as individual, safe and attractive as an individual car. With the integration of automated minibuses into an intelligent transport system, the whole transportation system could further be improved continuously. The transportation system would become more efficient but at the same time more resilient and flexible as failures could be determine in a predictive way and in case of failure in the transport system, automated minibuses could bridge the mobility gaps. Achieving these antinomic economic goals (efficiency and flexibility) into one transport system could make the transport system ambidextrous (Raisch & Birkinshaw, 2008).

Various upcoming digitalisation technologies like digital twins of automated (smart) vehicles as transportation objects, digital twins of (smart) citizens, digital twins of (smart) infrastructures and even digital twins of all kinds of stakeholders (PTOs, intermediaries, PTAs etc.) and mobility subsystems are enabling the governance, simulation, monitoring, tracing, analysis and optimisation of the entire mobility ITS/MaaS ecosystem. The AI-based self-learning loops of the mobility ecosystem could be ready for design and implementation into a future citizen-centric “Mobility Metaverse” aiming to optimise the physical/real “Mobility Universe” as a digital twin.

As the herein described concept is combining two disruptive and technologically complex innovations, there are also limitations that need to be reflected: on the one hand, there is the automated minibus itself as a product innovation and on the other hand MaaS/ITS as a mobility system innovation. The automated minibus depends on technical capabilities (perception and ability to determine, speed, level of automation etc.), the automated minibus environment (mapping and modification of routes, e.g.) and human involvement (for disabled access or the necessity of on-board safety driver, e.g.) that can be questioned even if fast improvements are expected via AI deep learning. These include the highly sensitive issues of user privacy and data and personal security, as well as the ethical issues that will always accompany automated driving (such as the trolley problem) (Fagnant & Kockelman, 2015; Geisslinger et al., 2021). The MaaS innovation is further driving a wider social, economic and environmental “ecosystem” innovation and will be impacted by an appropriate governance approach focusing on the general interest.

The main risks and limitations of the automated minibus in MaaS concept are addressed in the following section.

5 Discussion and Limitations

The aforementioned advantages of the AVENUE vision 2030 can only be achieved through the described disruptive product and system innovation. A prerequisite for achieving the AVENUE vision 2030 is an alteration of the current mobility paradigm. This paradigm is based on cheap fossil fuel energy, high CO2 exhausts, individual mobility (product orientation) and a linear economy (Fournier, et al., 2018).⁹ Altering this mobility paradigm requires social transformations in addition to technical innovations (such as automated minibuses) to accomplish a socio-technical transformation of the mobility system (Geels, 2011). A socio-technical transformation requires alterations in society, in business (ecosystem collaboration & market places, governance & management, vehicle/fleet demand & offering side) and in particular for stakeholders such as passengers, transport operators and related companies, technology provider, governance etc. For one, this entails an integration of stakeholders in designing, implementing and evaluating sustainable solutions for the mobility sector. As an optimum solution, all involved stakeholders should cooperate and agree upon a common purpose-centric strategy. This strategy should evolve towards a sustainable mobility system and, like mentioned above, create value to all stakeholders including citizen, public and private PTOs, OEM, city etc. and contribute to the transformation of our society in providing a no coercive answer to societal challenges.¹⁰

To fulfil this overarching purpose of serving general safety and sustainability interest, there is a strong need for stakeholders to harmonise different interests and cooperate with each other in an innovative coopetition mode. Known difficulties are that stakeholders will make choices based on knowledge available to them with the aim of maximising their individual gain. The combined outcome of the individual strategies could result in a situation that was not intended and not desired by any of the involved stakeholders [see, e.g. the prisoner’s dilemma from game theory by Tucker (1950/1983) in Morrow].

To establish a mobility system as shown in the AVENUE vision 2030, some stakeholders could still try to promote individual mobility with private cars or take advantage of the “winner-takes-it-all” phenomenon to capture the value creation to the detriment of other stakeholders or even take control of the entire mobility system by taking control over a significant mass of captured system relevant data (private MaaS) of the mobility ecosystem and selling it to the system’s customers (such as city governments) and public transport authorities (PTAs). It is therefore crucial to coordinate the actors involved and to synchronise operations both vertically (between the actors along the different modes of transport of the journey) and horizontally (among actors operating the same modality, such as different automated minibuses) to ensure seamless intermodality and interoperability of the transport system, similar to a supply chain (Giusti et al., 2019).

For a successful socio-technical transformation of the mobility sector, this process needs to be managed, to avoid and overcome divergent interests and resistance. Research has shown that 70% of transformations fail, and part of the solution for the path to transformation includes moving fast, “shared vision and stretching aspirations” by leaders and a “more holistic and expansive approach to transformation” (McKinsey, 2020). In addition, cornerstones of public sector transformation relate to performance and organisational health (McKinsey, 2021). In the following, we have identified transformation challenges at the technical, environmental, social, economic and governance dimensions. These dimensions are part of a holistic and expansive approach to transformation, and they can influence each other through the interdependence of success factors and barriers.

Next, we present in more detail the success factors and barriers that need to be addressed in order to successfully implement the AVENUE vision 2030. For each of the dimensions, we subdivide into success factors and obstacles for the automated driving system and separately for the automated minibus integration in an ITS/MaaS system.

5.1 Technical Success Factors and Obstacles

In order to achieve the desired vision, there are several technical obstacles and success factors that need to be addressed and solved. The deployment of automated minibuses in all environments, both in urban and rural, requires the technology to be able to handle all safety-related scenarios and to be effective, intelligent, connected and fully implemented within the ITS infrastructure. The technical obstacles and success factors related to the automated driving ecosystem concern automated technologies, such as integral sensor systems and technologies. The automated driving ecosystem can be defined in the following categories, with reference to Table 10.3 (Chap. 10), which is specific to the on-demand and door-to-door solution developed by AVENUE.

Automated shuttle capabilities: In terms of active safety, LIDAR and camera technologies that detect and avoid obstacles as an integrated part of the decision-making engine will be key. Machine learning algorithms and AI technologies will enable continuous development of driving capabilities and vehicle positioning in traffic patterns and infrastructure. Perception and judgement, which are also related to AI technologies, will allow the vehicle to build up and constantly update a library of objects and obstacles, so that it knows the difference between a car, a bicycle, a truck, a pedestrian and so on. Radar technologies are needed to perform emergency braking and detect large objects. The processing power of automated minibuses s and AVs is developing rapidly around the world, and technologies need to be shared to achieve the highest possible processing power—and therefore both speeds and braking algorithms and other safety features—ultimately making AVs safer and more reliant than human drivers. Accident analysis will also lead to passive safety requirements, which have to be anticipated, as they are structural to the vehicle platform. Conventional crash tests are not applicable to existing automated minibuses, and a compromise will need to be found quickly.

Automated shuttle environment: The main factors here are mapping and commissioning of new routes and updates of already existing routes. Current automated minibuses can yet only drive on maps recorded by the vehicle manufacturer, which is time-consuming and costly to record. These recorded maps are not shared with other manufacturers. In the future, these “map routes” will need to be shared and constantly updated across Europe to allow automated minibuses to drive anywhere. Similar strategies are followed in Japan and South Korea. The processing power of automated minibuses and AVs needs to be developed at a rapid pace in order to achieve speeds, braking algorithms and other safety-related features. Such elements are important to ensure that AVs are safer and more reliable than regular drivers (as already described in the prior paragraph). All weather-sensing systems should be able to operate in all conditions: this robustness is fundamental to vehicle reliability, a key requirement for service quality and safety. Currently, heavy rain and snow still cause sensors to malfunction and lead to many operational problems. In these cases GNSS makes it possible to position the vehicle precisely. The mobility infrastructure of cities and other application areas (e.g. roads, transport facilities and services, roadside sensors, charging facilities, communication technologies, future: AI and digital twin-based traffic management) also plays a crucial role in the implementation of automated minibuses and should be considered in local development planning. Finally, cybersecurity is critical as laws and standards need to be further developed to include automated minibuses.

Human involvement: The key factor here is the need for a safety operator on board of automated minibuses. Human interaction with automated minibuses will not be eliminated, but the transition from on-board safety drivers to remote supervisors and local technicians need to be developed and tested. Currently, automated minibuses can only operate with a safety driver on board and, in many scenarios, with a remote supervisor at the same time. This is making the current implementations very costly. Finally, access for disabled people is critical, as human intervention is still required. Human-machine interaction and interface technologies play an increasingly important and critical role for passengers, supervisors and other stakeholders in operations and service applications.

With regard to automated minibuses in a MaaS/ITS ecosystem, as previously mentioned in Sect. 18.2 of this document, open data, open platforms and open APIs are key factors to ensure a consistent citizen-centric approach and a single-point-of-contact applications for travellers in ITS. This openness is a prerequisite for multisided access to stakeholders, in particular passengers, and for integrating all modes of transport into a seamless journey. Interoperability and intermodality can improve performance and flexibility of the system, enable positive mobility externalities and reduce negative externalities. Open data, open platforms and open APIs are therefore a prerequisite for developing the benefits of automated minibuses in MaaS. The described AI loops are the precondition to enable an ITS with its benefits, but they require a collective effort for the common good. This process is challenging to manage as it involves several private economic interests of the related ecosystem.

These advantages however raise the risk of unauthorised tracking and traceability of passengers (privacy and safety of passenger data) and thus the risk of multiple undesired monitoring possibilities of mobility. In the public discussion on automated minibuses in MaaS, open data could lead to a major limitation and therefore needs to be technically secured (Meijer et al., 2014). With the definition of the open standards and regulations for data (e.g. GDPR) and APIs, as well as specific software-based technologies (ledger technologies such as blockchain), strong economic interests are further impacted as an ecosystem is open to all. A closed ecosystem could enable companies to capture and privatise the value of data in particular the positive externalities associated with it. The technical definition of the communication protocol, data format and interfaces has thus to be managed by the PTA and probably at European level to avoid security issues and “the winner-takes-it-all phenomenon” (see Chap. 2). The last point is important, as in an extreme case, the creation of private MaaS based on automated vehicles could limit the profitability of running mobility businesses but even severely limit the influence of the PTA in creating own rules for the urban mobility ecosystems.

5.2 Economic Success Factors and Obstacles

The results of AVENUE indicate that the business with automated minibuses is not yet profitable within the current automated driving ecosystem due to high driving-related personnel costs. This is mainly because safety drivers are still required to be inside the vehicle, and remote supervisors are currently legally forbidden or restricted in most European countries. The profitability of using automated minibuses in the fleet of PTOs is heavily impacted, making it a main restriction for usage in urban, suburban or rural areas. Furthermore, the high acquisition costs of the vehicles, elevated costs of feasibility studies and legal authorisations, infrastructure works and high annual depreciation are also seen as constraints (see Chap. 12; Antonialli et al., 2021).

The non-integration of automated minibuses with other means of transport within a MaaS reduces further the attractivity for passengers in terms of time and usability as several apps have to be used. Currently, there is no customer-centric approach that integrates automated minibuses with other means of transportation.

In both cases the technology has to be improved (see above), and governance (see below) has to be adapted. Regulatory sandboxes for self-driving vehicles could facilitate the testing and implementation of new technologies, leading to economies of scale and accordingly lower prices for wider adoption (BMWi, 2019). The limited acceptance and integration of automated minibuses’ into mobility networks by private and public PTOs, fleet owners and city/regional governments as B2B customers is due to technology readiness and operational business risks, as well as low subsidies for this innovative technology.

Integrating automated minibuses in MaaS/ITS is in fact one of the several possible future scenarios where the expected benefits are the highest due to the integration of societal goals. The governance discussed later will determine the benefits for the stakeholders and the general interest. Accordingly, it is expected that the stakeholders which will be negatively affected by serving the general interest could be inhibitors of the required mobility transformation. The leading stakeholders of the current mobility paradigm, such as the automotive industry and its employees, may be inhibitors and unwilling to adapt to the new constraints like it could be observed with environmental constraints (particulate matter) or electric vehicles. This is understandable as a real alternative to the spatial and temporal flexibility of cars could be provided and could satisfy the needs of the citizen (see Chap. 15): individual mobility with private cars could be more easily substituted by automated minibuses in MaaS/ITS, with the consequence of losing the shaping of the mobility value chain. The PTOs and PTAs will have to further adapt to the new digital ecosystem and customer-centric approach. Cities and regional governments, which would have to transform not only their operational mobility concepts but also their strategic business ecosystem (including open data, platforms, APIs, towards a new IT concept) from a “traditional” type of vehicle and (government) business ecosystem to a balanced, federated or democratised governance and operational IT concept with automated minibuses, could also face multilateral obstacles regarding the cost and effort of transforming a “running system” and the often notorious lack of budget as well as the difficulty of a necessary paradigm (mindset) change of public authorities that are often risk-averse (non-entrepreneurial). However, even this kind of disruptive (game changing) transformation of the business ecosystem can be guided by individually adapted and designed approaches (e.g. incremental transformation steps, “think big, but start small and scale fast”).

Finally, all stakeholders tend to maintain their current habits: passengers, members of organisations or the organisation itself try to reduce the psychological costs of change.

Automated minibuses in ITS will last but not least require investments in digitalisation, hardware, software, training etc. In particular, as cited in the Digital Economy and Society Index (DESI) 2021 (European Commission, 2022), the availability of employees with digital skills could in particular slow down the digitalisation and dematerialisation of mobility. The comparatively slow digital transformation of companies in many EU countries is linked to “a lack of employees with advanced digital skills” (European Commission, 2022). For example, by 2020, 55% of companies will report difficulties in recruiting IT specialists (European Commission, 2022).

5.3 Environmental Success Factors and Obstacles

The automated driving ecosystem of AVENUE shows a low environmental impact (see Chap. 13). A Life Cycle Assessment (LCA) of the automated minibus deployed within AVENUE indicates that its automated technologies account for less than 5% of the total energy used. In a near-future use case, 59% of the automated minibus impact comes from the use phase, while 39% comes from the component manufacturing phase (Huber et al., 2022). In addition, the automated minibus’ environmental credentials depend on many factors such as occupancy, vehicle speed, mileage and lifetime.

The analysis of the energy demand of automated driving technology and the connectivity-related demand shows that, on the one hand, predictive, adaptive and information sharing through vehicle communication with infrastructure and other vehicles improves driving performance (e.g. braking performance) and, consequently, energy consumption. On the other hand, a highly connected vehicle means more data processing inside and outside the vehicle, which may outweigh the sustainability of V2X. Overall, the energy-saving potential of predictive driving features is likely to exceed the energy used for data transmission.

The AVENUE environmental impact assessment reveals that if automated minibuses are well utilised in terms of mileage and are used regularly by passengers, they will have major environmental advantages over single-occupancy vehicles (Huber et al., 2022). The automated minibuses are seen as complementary vehicle in public transport and could increase the availability, flexibility, efficiency, effectiveness and reliability of regional public transport (Huber et al., 2022).

More relevant is whether and how automated minibuses could be integrated into a MaaS/ITS ecosystem. Depending on the governance scenario chosen, automated minibuses can be used in a separate ecosystem competing with current transport modes (laissez-faire strategy). In this strategy, for example, the convenience of the robotaxi and the customer centricity of the robotaxi provider could substitute public transport, private cars, bicycles and even walking. This is expected to generate additional traffic in the city (rebound effects), with associated external costs such as congestion, CO2 emissions, additional space etc. The proposed automated minibus in a MaaS/ITS, on the contrary, aims to satisfy the best needs of the citizens and the general interest in the leveraging of positive externalities and the reduction of negative externalities. Positive externalities like cost-efficiencies gained from increased usage (Mohring effect), strengthening network effects, enhanced accessibility etc. can be generated additionally through the use of data, intermodality and interoperability.

5.4 Social Success Factors and Obstacles

An important finding of the AVENUE social impact assessment (see Chap. 15) is that users and potential users demonstrate a positive attitude and a receptive (goodwill) attitude towards the automated minibus. There is therefore potential to convince those who are not yet refusing, but who are open-minded, by means of targeted communication campaigns, particularly on social media. We defined five target groups of potential users (unreserved goodwill; sceptical goodwill; undecided; critical reservation; unconvinced refusal), which differ not only in their perception of perceived benefits and concerns but also in their level of knowledge, preferred transport system, willingness to use and pay for travel by automated minibus and sources of information used to form their attitudes. The results show that high levels of goodwill among potential users (e.g. car drivers) and high levels of satisfaction among users of public transport translate into high levels of willingness to use (again). In general, users were very satisfied with their experience, and most were willing to use the automated minibus again. In Copenhagen 76% of users are very willing to use the automated minibus again, and in Sion 55% of respondents are willing or very willing to use the automated minibus again.

The most important factors for social acceptance are the (perceived) need for improvement of the current situation and whether the proposed alternative service meets this need for improvement. Fears of a lack of safety or security are currently less important for social acceptance. In addition, real experience with the automated minibus has a generally positive effect on trust in the system. Therefore, an important success factor for the social acceptance of automated minibuses is to allow citizens to use and experience the advantages of automated minibuses.

Some of the critical points encompass the highly sensitive issues of user data handling and data security, as well as the ethical questions that constantly accompany automated driving (such as the trolley problem, which addresses the ethical dilemmas of the trolley’s role in the car’s life, and the question of whether or not the trolley’s role is to act as a safety device for the driver) (Fagnant & Kockelman, 2015; Geisslinger et al., 2021). In addition, the results of a representative survey of citizens (n = 1816) indicated perceived concerns about the use of automated minibuses. The main concerns are related to the interaction of the automated minibus in unforeseen situations and with other road users (motorised and non-motorised), liability in the event of an accident, security issues and the risks of hacking and misuse of the software (Korbee et al., 2023).

In addition to this, the safety operators also mention that they are observing a high level of goodwill towards the innovative service. In the opinion of the safety operators, the users are highly satisfied, especially when subjective aspects such as the good atmosphere are taken into account, not least because of the lower number of passengers.

Based on their observations, safety and accessibility are qualities that are also evaluated as being satisfactory. In the viewpoint of the safety perspective, the automated minibuses meet the needs and requirements of users and contribute to a positive user experience.

A risk for the target groups of enthusiasts, uncritical goodwill and sceptical goodwill is that they may be disappointed if they recognise actual performance in terms of speed and flexibility. It is very important to increase both the speed and the flexibility of use through an on-demand service, or at least improved temporal and area flexibility (nearly as flexible as private cars) compared to existing public transport services, in order to ensure that the high level of goodwill actually leads to a high level of acceptance of the new systems. This positive perception is a good point, but it cannot be taken for granted: the perception can change in the event of an accident, for example.

Beyond the current AVENUE approach, integrating automated minibuses into a MaaS/ITS system aims to best satisfy passenger needs and increase social acceptance. We could therefore expect automated minibuses to be a real “game changer” in the future, making public and private transport more personalised and providing a real alternative to individual privately owned vehicles as it has the potential to increase flexibility for users and choice for passengers while serving the public interest. The results of a representative survey of 1816 citizens (of whom 1526 have privately owned vehicles) in Lyon, Copenhagen, Luxembourg and Geneva confirm that 45% of car drivers would be “willing” (22%) or even “very willing” (23%) to give up using their own car in order to use automated minibuses for the first and last mile, if it were available. If the service is on demand and door to door, acceptance could be even higher (Korbee et al., 2023). Of course, this needs to be explored in future research if these measured attitudes can be changed in behaviour.

Another alternative could be the robotaxi which could satisfy best individual mobility needs without a change in transport mode but at the same time would reduce and privatise positive externalities and increase negative externalities through additional traffic and space. This choice could be better accepted than automated minibuses in MaaS in ITS but without serving the general interest. The choice between a non-integrated robotaxi in a MaaS and automated minibuses in a MaaS/ITS therefore needs to be discussed, organised and formalised in a future governance. Automated minibuses in an ITS/MaaS is of course the better alternative for the city and its citizens, but this requires a transformation towards a new social contract for mobility to avoid social frustration and crisis (Shafik, 2021; Rousseau, 1762). Special attention should be paid to digital illiteracy to avoid excluding parts of the population from the transformation process.

5.5 Governance Success Factors and Obstacles

The above limitations are all influenced by the chosen governance. Thus the governance is a key issue to enable the technical ecosystem and the integration of automated minibuses in MaaS or ITS.

For the automated minibus and its vehicle-related ecosystem, the development of regulations for the automated driving ecosystem encompasses amendments of the Vienna Convention on Road traffic and the adoption of a new legal instrument on the use of automated vehicles in traffic at the international level.¹¹ At regional and national level, it should be possible to approve the operation of automated minibuses without strict requirements for the safety operator (in a first stage) and with a remote supervisor only (in a second stage),¹² as the costs and time for type approval are a key issue for the diffusion of automated minibuses. In Member States that have not yet adopted AV-related legislation, regulatory sandboxes¹³ for self-driving vehicles, as mentioned above, could ease the disruption process and provide EU stakeholders with similar competitive conditions to other more liberal parts of the world to test and trial new technologies and facilitate their diffusion. Moreover, a legislative database at European level which brings up to date and real-time information about the fast-growing automated vehicles legislation (e.g. type approval, commercial, cybersecurity of vehicle, definitions, infrastructure and connected vehicles, insurance and liability, licensing and registration, operation on public roads, safety operator requirements etc.) on local, state and regional regulations is missing. This could simplify the planning of all the stakeholders like vehicle manufacturers, deployers of automated minibuses, transport operators, states, municipalities etc. and encourage the diffusion of the automated minibus technologies and the improvement of the transport system. The USA is a good example of how such a database could be managed (NCSL, 2022).

At the city or local authority level, it will be necessary to design, implement and monitor individual technical (vehicle, IT) certification and licensing schemes for automated minibuses and their integration into MaaS according to relevant regulations or standards (similar to TÜV in Germany) as well as related training courses for operational personnel (experts).

City or local authorities are also in the best position to understand the local mobility needs and where automated minibuses could provide new services and/or improve the existing services (as feeders).

Governance is therefore crucial for balancing these interests, although it is possible that the creation of closed ecosystems supported by the users who are individually highly satisfied by customer-centric affordable robotaxi could weaken the power of the PTA and the ability to balance interests.

Implementing AV for mobility is also a technical and innovative approach that believes in progress and science, using dematerialised consumption to save resources for a sustainable development (mobility) path. It will have to overcome the growing scepticism generated by debates on concepts such as “degrowth” or “frugality”, which promote a change in consumer behaviour through sacrifice in order to save energy and resources (sufficiency strategy).

For automated minibuses integrated into a MaaS or ITS ecosystem, public institutions have a crucial role to play in developing an interoperable, standardised and connected data landscape across different sectors, including the mobility sector in particular, through specific regulatory frameworks. The benefits of interoperable, standardised and connected data include improvements and new offerings in public services, an increased government efficiency, data-driven policymaking and the value of open data (Domeyer et al., 2021). The regional and national data strategies are also important: these address data protection and privacy requirements. Some of the challenges to interoperable and connected data stem from the fact that public and private data remain dispersed, not digitally accessible and not interoperable (referring to obstacles that prevent the combination and joint processing of data) (Domeyer et al., 2021). Currently, vehicle manufacturers can control who has access to which vehicle data, creating an undesirable “gatekeeper” position that will probably only be resolved partially by the cross-sector Data Act (European Commission, 2022).

To address these issues, European and national strategies for integrated data management are needed, as well as infrastructure and the setting of technical standards, as “fast and automated data exchange is only possible through harmonized data formats and standards” (Domeyer et al., 2021). The “European Data Strategy” aims to create a single market for data, and one of the main goals is to create safer and cleaner transport systems (European Commission, 2020c). For automated minibuses in MaaS or ITS, open APIs and open data are accordingly a prerequisite and a key to achieving interoperability and coordination of all stakeholders in the mobility system. As a result, seamless trips to meet the needs and acceptance of citizens become possible.

At the same time, this raises the difficult question and debate for regulation: who benefits from the use of data? This is clear for data collected in the event of an accident, as it will help to understand the causes and build anonymised lessons learned, to the benefit of all mobility stakeholders (see Figure Loop 1, Fig. 18.10), as feedback is shared. For other data, while individuals may have an interest in not sharing their personal data for privacy, safety or security reasons, private companies are interested in collecting these data to create new profitable business models (“winner takes it all”, advertising or other forms of value capture) but also to better meet customer needs. The use of data is further necessary to allow OEMs to optimise automated minibuses and make the drive more reliable and safer or to realise the loops described in Sect. 18.4. PTOs could use the data to optimise the use of the fleet and the PTA to serve the general interest. A balance therefore needs to be found and managed between the interests of the individual data provider (the mobility user and its stakeholders, GDPR) and the legitimate interests of the community in using the data to serve the general interest. These discussions should take place at European, national and local levels. This will require a less administrative corporate culture on the part of PTAs in order to manage the transformation and, in particular, to find the right balance between the numerous stakeholders towards the unity of society for sustainable mobility (purpose). A decentralised governance model which unifies the mobility network and supports a federated, democratic, decentralised mobility platform using distributed ledger technologies (DLT, so-called Web3) to “integrate stakeholders and different technologies in a non-discriminatory and equal way” (Schmück et al., 2021) could accordingly be chosen. As a result, the beneficiary of the system would be directly the data subject rather than an intermediary party that owns the mobility platform and the customer touchpoint (Schmück, et al., 2021). As an alternative, other federation-based governance strategies could be used, where so-called substantive platforms are used to serve the general interest between stakeholders. The aim is to integrate economic dimensions to satisfy (mobility) needs on the one hand and political issues (through collective deliberation) to preserve the commons and enable purpose within a community on the other. Linking all stakeholders allows defining the needs, rules and structures of the ecosystem without intermediaries. The substantive platforms can vary and depend on the valuation chosen, which can be based on (classical) markets, on off-markets (based on donation, reciprocity and/or redistribution) or on hybrid forms combining market and off-market valuation. These new models are still exploratory. These respect GDPR by design (ethics by design).¹⁴

In terms of data protection and data privacy, the GDPR is an asset for personal data protection, safety and security, but it brings complex and challenging implementation in the field of connected and automated vehicles, as well as obstacle for AI (see the aforementioned loops within this chapter).

From a legal perspective, there are many challenges ahead, including the following. With regard to privacy and personal data protection, certain key concepts (e.g. “personal data”, “identifiable natural person”, “anonymous information” and “processing”) are defined and could be interpreted in such a way that the rules of the GDPR would apply, to the extent that this would require additional investment or hinder or slow down the development of the processing activities described in previous sections above (Sects. 18.3 and 18.4).

With respect to the data flows between the different stakeholders described in Sect. 18.4, the distinction between “controller” and “processor” is complex, taking into account the concepts of joint controllers, controllers in common and sub-processors.¹⁵

The “controller” is any person or entity that determines the purposes and means of the processing of personal data, and the “processor” only processes personal data on behalf of the controller. Hence, it may be difficult to identify the role of each actor intervening in this context their underlying obligations.

Only technically collected data (i.e. non-personal data) or personal data anonymised by the “GDPR filter” (described in Fig. 18.8 above) or by another technique will be provided to the MaaS data and service platform operator. This data may become personal data or even sensitive data due to the transformative impact of big data analytics and thus trigger the application of the GDPR and further privacy and data protection laws.¹⁶

The potentially high volume of data (personal and non-personal) collected by automated minibuses can be assimilated to big data analytics, challenging certain core assumptions of EU data protection law, such as data minimisation and purpose limitation.¹⁷

The scope of the e-Privacy Directive (lex specialis to the GDPR) is much broader and applies not only to personal data but also to all information regardless of its nature. Consequently, this framework requires an assessment of which processing activities may fall within the scope of the e-Privacy Directive and, subsequently, what specific requirements the e-Privacy Directive imposes on stakeholders in relation to these processing activities. Furthermore, where personal data are involved, the interaction between the GDPR and the e-Privacy Directive needs to be closely examined.¹⁸

The (current) lack of standards for open APIs could lead to coordination issues, and it should be ensured that Member States apply the same standards to avoid the creation of technical barriers.

Finally, it is important to note that there are different governing approaches from public authorities for MaaS development (see Table 18.1). The aforementioned economic and environmental scenarios show that the governance determines if the rules serve the general interest and promote sustainable mobility and accordingly the technical, social, economic and environmental impacts. A governing by authority on one side which puts forward the general interest can be imagined as well as a governing by “laissez-faire” (e.g. voluntary lack of regulation or minimalistic regulation) on the other side where few private companies could take advantage from the “winner-takes-it-all phenomenon” of the mobility ecosystem and capture the expected benefits for private purposes. In the case of “laissez-faire”, potential competitors would be excluded (vendor lock-out strategy) and could not find a viable business model anymore. Vendor could also be complementary to the private ecosystem (vendor lock-in strategy) when an alternative market is not viable anymore and the competitor has to become a complement of the vendor (proprietary ecosystem owner). Even a threatening of Public Transport Authorities might be imagined when the private terms and conditions of the ecosystem of a private MaaS substitute to the public laws of the PTA. This has already been observed in other European markets. The lack of speed of adapting the rules of the PTA to the disruptive changes in mobility could hence generate the same problem that the EU had to experience before introducing the DMA (Digital Market Act). The aim of the DMA is to ensure the proper functioning of the internal market, by promoting effective competition in digital markets, in particular a contestable and fair online platform environment. More specifically, the DMA’s objectives are (1) to address market failures to ensure contestable and competitive digital markets for increased innovation and consumer choice, (2) to address gatekeepers’ unfair conduct and (3) to enhance coherence and legal certainty to preserve the internal market (European Commission, 2020d). As the DMA is not likely to apply to the mobility sector yet, an anticipatory approach for the local, national and European mobility markets which takes the specificities of the transport sector into account through a dedicated “mobility data act” is most likely needed and would accordingly be recommendable as well.

Table 18.1 Expected actions from public authorities in the development of MaaS and the different governing approaches by Audouin and Finger (2019)

6 Conclusion

For all initial questions about “how to integrate AV in the city transport system to serve general interests”, the AVENUE vision could provide a convincing conceptional solution approach and a pragmatic transformation concept based on the hypothesis that automated minibuses integrated into intermodal transport and MaaS or better ITS can be a promising game changer in urban mobility. The automated minibus, deployed on demand and door to door, will provide more mobility choices and flexibility for all citizens, including better accessibility for people with reduced mobility (PRM) and potentially better acceptance of AVs based on positive experiences.

As argued above (Sect. 18.3), a coopetition governance scenario and open data schemes (open data, open platforms, open interfaces and protocols) are further key factors to guarantee fair competition between public and private mobility providers, avoiding dominant position and “the-winner-takes-it-all” effects. In addition, AVs coupled to ITS and AI are expected to make the transport systems more reliable, safe, efficient and flexible (concept called ambidexterity), and thus the antinomic goals—incremental and disruptive innovation—are combined. As a result, the transport system becomes, through ITS and automated minibuses, citizen-centric, inclusive and sustainable, enabling positive externalities (Mohring effect, network effects, enhanced accessibility etc.) and lowering negative externalities (e.g. congestion costs, space, use of energy and materials). The citizen-centric approach could thus become purpose-centric, serving the general interest to the best for all stakeholders. This vision is a key for acceptance and thus coherent with the SUMP concept of a holistic approach and the requirement of the EU Green Deal (European Commission, 2020a), EU Sustainable and Smart Mobility Strategy (European Commission, 2020b, 2021) and the European Data Strategy (European Commission, 2020c).

Through the vision concept, the EU could gain on sovereignty with automated minibuses in ITS and become the worldwide mobility system leader in purpose-centric—this means citizen-centric, responsible, independent and sustainable for all the stakeholders—mobility which respects human and individual rights (data privacy and security, GDPR), by aligning and utilising future (technical and business) mobility product and system innovations for this purpose. This could be an alternative to the path of development chosen in other continents.

Of course, this approach is a socio-technical and innovative approach which believes in progress and science and uses dematerialised consumption to save resources for a sustainable development (mobility) path (McAfee, 2019). It will have to convince the growing scepticism which arouses in discussions around concepts such as “degrowth” or “frugality” which promotes a paradigmatic change in consumer behaviour through renunciation to save resources. This so called sufficiency strategy asks people to consume consciously and since around 13% of consumer spending by European citizens is spent on mobility, this issue is an important factor (Statista, 2020). Our vision shows a feasible and promising path of how dematerialisation of mobility through automated minibuses and ITS could save energy and resources without sacrificing individual mobility needs and comfort (Linz, 2004; Mauch et al., 2001).

This vision will be the basis to structure and deduct our transition goals, recommendation and transition roadmap to design the future (public) transportation service. The Horizon Europe project ULTIMO (Safe, Resilient Transport and Smart Mobility services for passengers and goods) implements parts of the concept from 2022. The concept has been further presented at Transportation Research Area Conference 2022 (TRA, Lisbon) (Fournier et al., 2023) and at the High-Level Dialogue on Connected and Automated Driving organized in June 2024 in Ghent by the Belgian Presidency of the European Council (Flemish Ministry of Mobility and Public Works 2024).