Keywords

1 Introduction

Automated minibuses constitute an emerging technology that will transform public transportation in various ways. During the journey of this transformation from the conventional buses to the automated approach, a variety of safety, security and service quality needs are required to be addressed for achieving a successful transition. Sustainability and durability of the autonomous vision are also required for all stakeholders to engage acceptability. Traffic management, energy consumption, customer satisfaction are additional aspects have to be taken into account to accomplish sustainability. Mobility as a service (MaaS), which denotes mobility provided as a service by combining transportation services from several transportation providers, is deemed critical to change the future of transportation by enabling more personalised transportation services. Data privacy and relevant regulations have to be taken into account to facilitate this procedure.

In this chapter, a shared sustainability and durability target for the society and the related companies is described, along with the conditions to achieve these objectives. Next, the critical path for market introduction of safe automated minibuses is presented, while the quality and safety state of the art for automated minibuses is discussed. A self-learning automated transport system at the European level is also revealed, whereas the data privacy of incident analysis and lesson learned sharing is explained. Finally, the automated minibus safety and service quality levers are summarised.

2 A Shared Sustainability and Durability Target for the Society and for Companies

With the different functionalities tested on the demonstration sites, the AVENUE project confirmed that automated minibus solution is attractive in terms of urban mobility services and now technically accessible, as presented in the previous chapter.

This is very important for the public power, as this technology can also contribute to the urgently required urban ecologic transition, if carefully integrated into the existing urban mobility systems, as a green and flexible complement to the conventional public transportation infrastructure. The best example of a conventional infrastructure is railway, which is adapted to massive transportation but not flexible: this introduction of automated minibuses is now a clear target for the European Union.

The following chapters will discuss the environmental and social impacts, which are positive when replacing private cars by such solutions, thanks to new attractive services: no doubt about that, but preliminary conditions have to be fulfilled to avoid any future regression.

This introduction of automated minibuses has to be organised carefully by the European Union. If not, the transition towards a more flexible and ecologic transportation system will not be sustainable in terms of societal acceptance, and the created business will not be durable for private actors, investing in this technology, which has to be financially amortised in the future.

3 The Conditions to Make it a Sustainable and Durable Solution

With these two goals in mind, the new automated minibus services will have to be not only attractive but also to confirm this good feeling with efficiency, reliability and safety, to establish year after year the customer satisfaction.

These new services will also have to get confidence from the public authorities, including other concerns like traffic management, energy consumption and health protection, with the associated mediatic and cost impacts.

3.1 Traffic Management and Energy Consumption

On-demand and door-to-door services are crucial to get a flexible transportation service, which has to be inclusive for the elderly and persons with reduced mobility: these functionalities imply that automated minibuses will drive and park among other vehicles, including private cars, without any remarkable disturbance to the traffic of other vehicles, which does not mean any impact. If a conventional driver is autonomous in such situation, the vehicle has to be autonomous in the existing infrastructure and traffic: “a vehicle equipped with ADS aims to perform the entire dynamic driving task on a sustained basis within a defined operational design domain without driver involvement” (Standing General Order on Crash Reporting) (NHTSA, 2021, 2023).

In the future, the traffic performance will not be measured by the maximum speed available on a specific route. The average speed is more representative, but it will not be sufficient to measure the traffic performance: the target of traffic management will be to guarantee an acceptable travel time from A to B and so to limit the variability.

The responsibility associated to public transportation with automated driving will lead to a careful vehicle control, compared to private cars, with the hope of a quieter driving, without traffic jams, avoiding energy consumption and time wasting, with working loss, etc.

This quieter traffic could be slower, but the careful introduction of these automated minibuses has to avoid traffic disturbance and congestion, so the automated minibus will have to manage blocking situations, even when coming from driving rules’ strict application.

A simple example is the necessity to cross the white line to pass a private car which is not correctly parked, which implies a human solution to be used: in such a case, the US strategy for Automated Driving (NHTSA, 2020) is a driver mimicry based on vehicle artificial intelligence, which will be sometimes necessary, maybe thanks to the quick arbitration of a distant operator.

3.2 “Customer” Durable Satisfaction, Including Safety

The “customers” are service users but also other citizens, public transport authorities and more generally the different European states, with public concerns on emissions, energy economy, public health and cost.

Concerning the citizens who will pay in the future for this service (project demonstrators were free), they have explicit requirements in terms of comfort, punctuality, trustworthiness and resilience, which will have obviously to be satisfied.

In comparison, some other requirements are more difficult to manage as they are hidden, because they are implicit: it is generally quality and safety, with the associated conditions of technical reliability, cybersecurity, regulation compliance, insurance and juridical responsibility, which are behind quality and safety.

For these implicit requirements, the public power involvement is necessary: a roadmap has to be established from the realised demonstrations to the future durable customer-paid services.

To avoid future regressions, the roadmap has to anticipate the probable risks concerning traffic impact, service quality and road user safety, especially vulnerable users, and to organise the global improvement of these new services.

3.3 Safety Measurable Targets and Steps

In addition to the fact that it is an implicit requirement, there is a second difficulty to manage safety, which is the difficulty to measure it: the fatality target has to be close to zero and not measurable, as it is clearly unacceptable that a city transportation vehicle kills a citizen!

The minimal requirement level is to maintain the global safety level of existing transport services, with coaches or buses (EU, 2023a). This fatality level is not easily measurable but can be estimated globally, with the help of massive data collected in terms of person travelled kilometres, which is the principal function of public transportation, but also of the traffic system, where comparison can be made between different transport modes, including individual vehicles.

In fact, this minimal requirement is already very ambitious, as buses and coaches are very heavy in comparison with the planned minibuses and with the existing vehicles, which is a mechanical advantage to protect passengers in case of an accident.

As a consequence, it can be discerned that the passive safety performance of buses and coaches is much higher than cars, even with lower passive safety technology.

The future automated minibuses will be smaller, as the absence of a paid driver allows us to reduce the number of passenger places, with benefit to transportation flexibility: the targeted weight for the urban automated minibuses is close to the weight of the coming electric SUVs, i.e. 2500 kg.

Based on that, two recommendations have been made to the European Community:

  1. 1.

    The safety measurement has to be done at a vehicle level and based on injuries compared to running hours, associated later to travelled kilometres. One condition is that public transport operators will be obliged to declare any accident and more generally any incident putting at risk safety or quality, based on a regulatory list to be defined (door blocked, vandalism, aggression, forgotten luggage, etc.).

  2. 2.

    A realistic target should be officialised for market introduction, which is 10−5 injury/hour, for example, less than 1 accident during 1000 running hours in terms of active safety performance and less than 1 injury among 100 accidents for passive safety. After the continuous improvement process, at the end of the roadmap, the target should be at the estimated level of buses and coaches: 1 passenger injury +1 vulnerable injury for one million hours: not easy!

Required in the USA as identical to conventional vehicles which is demanding, the passive safety level should be high, as citizens will consider they have to be protected from external vehicles when installed in the minibus, and as pedestrians will refuse that the minibus front face would be less protective, or more aggressive, than a conventional car.

Today, conventional cars are equipped with a front bonnet over the thermal engine, which is very useful in case of pedestrian crashes. Without this bonnet constraint, these front faces will be vertical, to limit the occupied space and maximise place number, with high risk of pedestrian projection and crossing, even at low speed.

At low speed, the projection is not far: the pedestrian can fall just in front of the running automated minibus, limited in terms of braking deceleration because of standing passengers inside the minibus, which is also called the “tramway or trolley dilemma” (Exeter, 2020).

Passive safety innovations will have to be financed and developed for door-to-door transportation in pedestrian streets, and new principles have been explored, as an inflatable structure integrated in the vertical front face, giving a better kinematic for pedestrian crash with energy absorption.

For years, researchers have been waiting for regulation demand and industry financing, to protect pedestrians but also cyclists and scooterists in such situations, against a front face, relatively high, which is already the case for a child in front of an SUV.

4 The Critical Path for Market Introduction of Safe Automated Minibuses

More than active safety systems, the passive safety requirements are structuring for the vehicle architecture and have to be introduced at the early beginning of development planning.

Two different usages will have to be dissociated, as one of them is compatible with a conventional architecture:

  1. 1.

    Downtown usage with derogative access to pedestrian streets requires vertical faces with a specific architecture to protect vulnerable street users and to facilitate vehicle stop and go with persons with reduced mobility (PRM) and standing passengers. To allow such vehicle developments (low production volume, limited investments), passive safety requirements have to be adapted: hopefully, in that case, performances can be limited (speed, acceleration, braking, etc.).

  2. 2.

    On the contrary, the liaison usage between suburbs and cities at higher speed (or the rural usages) must reuse industrial vehicle platforms, with conventional passive safety requirements. For technical protection in case of crash, the best battery position is under the cockpit, with impact to the platform height and to the quick access, especially for elderly or for parents of young children: this reuse will reduce the critical path, with applications available in the near future.

To develop correctly and safely the door-to-door services in the town centres, we will need a specific architecture, a low robotised platform adapted to elderly and PRM access, accepting standing passengers, represented by our H2020 AVENUE demonstrators, but they did not have to be homologated in terms of vehicle safety.

The critical path is longer for such application, which is the most important societal target: the first commercial application will probably use new electric robotised platforms, already under development for duty applications, but these vehicles will not be so optimised as H2020 AVENUE demonstrators in terms of access and vulnerable protection.

Their lesson learned will be useful for the later introduction of specific vehicles for downtown transportation, everywhere at low speed, respecting new passive safety requirements, to be urgently defined and integrated in the vehicle certification process.

5 Quality and Safety State of the Art for Automated Minibuses

The AVENUE experimentations were too short to measure anything in terms of quality and safety, and only attractivity measurement and characterisation were planned and done.

In any case, these experimentations showed that four preliminary conditions were not met, where the state of the art has to be improved before market introduction:

  1. 1.

    The technical reliability is not able to guarantee a quasi-nominal behaviour, which has to respect the ISO 26262 (ISO, 2018) automotive standard, which is available and manageable.

  2. 2.

    This level of reliability has to be obtained with an acceptable maintenance level, which is not the case today for sensor cleaning, as an example. Another risk to be taken into account is the consequence of light material accidents if sensors are not well protected.

  3. 3.

    The robustness towards meteorology and environment factor components is not sufficient, especially with the lack of reliability of the connected infrastructure in the urban and complex environment.

  4. 4.

    The protection against cybersecurity attacks of connected vehicles, which has to be built very early in the software and hardware architectures, is distributed among many different actors: standards have to be established and their application controlled, since a connected automated vehicle can be turned into a weapon, inside the city.

When these four basics will be mastered or better satisfied in service, the guaranteed nominal behaviour of the minibus will have to be confronted to risky scenarios, coming from real situations and human behaviours.

These risky scenarios have to be identified as they are leading to incidents and sometimes to accidents: so in addition to the ISO26262 (ISO, 2018) standard, the safety of the intended functionality (SOTIF, ISO21448) (ISO, 2022) will have to be checked in front of real driving situations and real human behaviours.

To guarantee a positive result at the end of the development, these risky scenarios will have to be specified at the beginning, using a scenario library as built by the MOSAR project in France (SystemX, 2023b) or automated driving scenarios (ADScene) (SystemX, 2023a), so they will be taken into account in the design, which implies numeric simulation of these scenarios.

6 A Self-Learning Automated Transport System at European Level

As explained concerning safety targets, a continuous improvement process will be necessary, based on systematic declarations of incidents and scientific analysis at transportation system level (including human management), with anonymised feedback to all actors, at the European level.

The proposed improvement process will need the implication of the public transport authorities to demand a systematic incident declaration when giving agreements, to organise the confidential analysis, to share the lessons learned, to get the corrective actions and to adapt norms and regulations. This field experience will help to establish future norms or to improve regulations, in cooperation with technical representatives, ensuring technical feasibility.

The establishment of this improvement process under the control of a public authority will allow us to accept an intermediate level of safety for market introduction. The market introduction with paid services will begin the first quality measurements: for sure, corrective actions will be necessary and will benefit from the same improvement process, sharing lessons learned and best practices, to the benefit of all actors.

7 Data Privacy of Incident Analysis and Lesson Learned Sharing

The absence of a human driver introduces new risks in terms of vehicle safety and passenger behaviour: to manage these new risks and to avoid bad incidents which could lead to a rejection of the solution, it is necessary to add a specific process when an incident occurs, as depicted in Fig. 4.1.

Fig. 4.1
2 flow diagrams. Left, P T A for transportation management and service evaluation, normal service in the transportation system has O E M, outgoing infrastructure with no personal data, and incoming P T O with feedback. Right, incident analysis and public improvement by trusted experts. protected data goes out, and improvement comes in.

Data collection process in case of incident

Full size image

An incident will be defined as a situation that could compromise the safety or the service quality and lead later to a rejection and most probably to expensive corrective actions. It can be an accident but also a harsh braking, a difficulty to close the door, a vehicle blocked, vandalism facts and an aggression.

In case of an incident, the data collection includes personal data concerning passengers and other road users but also private information coming from technical products, the automated vehicle and its environment.

To establish robust lessons learned, we need all available data and relevant technical skills, which is common with the road accident analysis, for road safety improvement: the data privacy has to be protected, taking into account relevant regulations, such as General Data Protection Regulation (GDPR) (EU, 2023b; EUR-Lex, 2016).

Road safety methodologies can be reused to conciliate data availability with privacy, as presented in Fig. 4.2, but also to satisfy commonly legal treatment needs with data needs for scientific analysis:

  • Beyond the system daily management, incidents will be collected, and data will be protected and crypted with restrained access.

  • Data will be analysed by trusted experts in no-profit organisations: scenario coding, simulation and statistical analysis.

  • The technical feedback will be shared, as anonymised, without any personal data and technical secret.

Fig. 4.2
A flow diagram has arrows from vehicles to 3 elements, O E M, P T O, and P T A. In case of incident, protected package with personal data goes to P T A and then to trusted experts. The latter goes back to P T A with risky scenarios, then to P T O with service Q and As, and to infrastructure, then to vehicles with feedback.

Lesson learnt process for anonymisation and sharing

Full size image

Justified by safety and included in the AVENUE vision recommended to the European Community, this continuous improvement process will contribute to both safety and service quality improvement.

8 Automated Minibus Safety and Service Quality Levers

In addition to the societal benefits, the general target of the introduction of automated minibuses in the urban transportation system is to offer the same level of safety and service quality as conventional transportation, which is very ambitious for the explained technical reasons and dependent on many human factors not well known today.

This is very ambitious, and it cannot be reached without the anticipated implication of the public power to prepare our countries and towns to their future challenges.

Six levers will have to be used by the public transportation authority at the European level, but not only, as many agreements are delegated at the local level:

  1. 1.

    Continuous and general improvement process, based on known methodologies.

  2. 2.

    Systematic and limited incident data collection, feeding risky scenario database.

  3. 3.

    Functional safety and SOTIF ISO norms (26262 & 21448) (ISO, 2018, 2022), using this database.

  4. 4.

    Appropriate effort repartition between active and passive safety, with trade-offs.

  5. 5.

    Dissociation of urban needs from liaisons at higher speed and different vehicles.

  6. 6.

    Innovation for urban shuttle architecture, especially for vulnerable protection.

From now, the regulation work has to include active but also adequate passive safety for two different usage modes.

Secured by the public control on the applications, this coming “automated minibus step” is a valuable investment for general interest: it will not only contribute to the climatic and ecologic transition, but it will contribute to the future life in our towns.

It will also prepare the introduction of private automated cars in towns: it can be later with benefits from this experience or never because of unacceptable risks.

9 Conclusion

In the advancing of automated minibuses, sustainability and durability are required to enable wider adoption within Europe. Several safety, security and service quality needs are deemed also mandatory to be addressed in order to attain broad acceptability of the new services. In this chapter, the conditions to achieve these objectives are presented. The critical path for market introduction of safe automated minibuses is also discussed, along with quality and safety state of the art for automated minibuses. The data privacy of incident analysis and lesson learned sharing is introduced to enable further discussions about the evolution of the public transport system in Europe. Finally, the automated minibus safety and service quality levers are presented. This analysis envisions to adequately facilitate the adoption of automated minibuses by providing specific guidelines.