1 Introduction

Driverless vehicles are no longer something just anticipated by visionaries, confined to research labs and a technology of the future. In fact, many new vehicles on the road are already being equipped with automation level 2 functions (overview of automation levels see Table 1). Most recently, Mercedes (2022) received a lot of attention, as it is now the first car manufacturer in the world to offer an approved Level 3 system and is thus also liable for accidents during automated driving mode. Furthermore, we are all familiar with the automated and autonomous vehicles (here jointly referred to as AVs) used for Google Street View (Reuters 2020) or the autopilot system developed by Tesla (2022) that has already made headlines in every respect. Both these examples are already driving on the road, wherever the legal framework permits. In contrast to extensive discussions on the environmental effects, legal issues and safety aspects of this technology, the questions of accessibility and inclusivity have so far only been dealt with in an insufficient manner.

A study conducted by Neumann et al. (2003) found that 48.1% of people with disabilities in Germany would travel more frequently if there were more barrier-free options. While some barriers have been removed since then, people with disabilities continue to travel significantly less than those without disabilities and experience notably more travel difficulties with any type of trip (Clery et al. 2017).

And yet, the access to society and transport for people with disabilities receives a high level legal obligation of countries. The United Nations Convention on the Rights of Persons with Disability (UN CRPD) was adopted in 2006 and came into force in 2008. The UN CRPD does not explicitly mention the right to access AVs for people with disabilities but includes articles that cover the obligation to provide access for disabled users, on equal basis with others, to transportation and technologiesFootnote 1. The EU and its Member States have ratified the Convention, and therefore undertaken such obligations. On EU level, the Articles 25 and 26 of the Charter of Fundamental Rights of the European Union (GRC 2012) include related issues, i.e. the rights of the elderly and integration of persons with disabilities and their rights to lead an independent life whereas access to transport is a precondition.Footnote 2 The existence of both of these articles emphasises the importance of including vulnerable groups in society and facilitating their autonomy and independence.

How this can be supported with the help of AVs will be outlined in this chapter. In doing so, we draw not only on current scientific literature, but also on the findings of the research project HADRIAN, which will be briefly presented below.

The project Holistic Approach for Driver Role Integration and Automation Allocation for European Mobility Needs (HADRIAN)Footnote 3 aims to shape automated driving from a holistic, user-centred perspective, starting the development process with specific mobility scenarios concerning individual users as well as mobility needs and constraints. This much more technical project offers us insights regarding the interfaces for human-technology interaction specifically for older people, which are at least partially transferable to other vulnerable groups as well. In addition, a co-determination approach was also part of the methodology in this project.

But before focusing on the potentials and the risks of AVs for citizens from vulnerable groups in mobility, the basics of automated driving are addressed in Sect. 2. In addition to an explanation of the automation levels, we approach the following questions: Where do we stand technologically and will we all soon be sitting passively in self-driving cars and buses? Sect. 3, after defining the vulnerable groups addressed in this chapter, presents the potential benefits, as well as new barriers, that the use of AVs in our mobility system can bring to these vulnerable groups. Section 4 will show how the previously mentioned hurdles can be overcome. In this respect, the potential of the use of universal design will be considered in particular. A good practice example of automation in vehicle development provides Sect. 5. The HADRIAN project will be used to illustrate how automation levels 2 and 3 can facilitate the activities of the vulnerable elderly group in practice. Section 6 contains the conclusions of the previously identified aspects.

2 Automated and Autonomous Driving – A Glance at the State of the Art

2.1 Level Classification of Automated Driving

The terms automated and autonomous are often interchanged in non-scientific representations and discussions, although they do not mean the same thing. In technical and scientific literature, different levels of automated driving are mentioned. Depending on the defining instance, the number of levels and subtleties in the respective descriptions vary. In order to give an introduction to the basics of autonomous driving, we will use the latest standards of SAE International (2021). According to them, the levels of automation can be defined as follows:

Table 1. Level of automation in vehicles (SAE International, 2021)

According to this classification, levels 0 to 2 of automation are actually the major fraction of vehicles already on the road today. Vehicles with technology assisting the driving experience and contributing to increase road security such as ABS and ESP functions required by law, or assisted parking and acceleration tools, are indeed classified as level 1 and level 2 automated vehicles, respectively. Although not explicitly mentioned in this case, following the classification presented in Table 1, autonomous vehicles are those belonging to level 5, i.e. those that are in fact driverless vehicles.

2.2 Current Discussions on AVs Barriers - Why Do not Cars Already Drive Autonomously?

While AVs can transport both people and goods, the notable AVs currently in fast expansion and with potential for deployment in near future are robot-taxis, bus shuttles or similar forms of public transportation. Currently there are already level 3 public transport shuttles driving on small and controlled sections of roads in different cities, never without the accompaniment of a human driving assistant and mostly on pre-programed routes. Martinez-Diaz and Soriguera (2018) provide an overview of AV technology and challenges that are still an issue.

Although the technology of automated driving functions has grown rapidly in recent years, current forecasts of market availability are no longer as optimistic as those from a few years ago. Some manufacturers have already presented prototypes for level 4 and 5 cars, trucks (Volvo 2022), bus shuttles or even buses (CAVForth 2022). However, most expert predictions do not foresee a generalized availability of high or fully automated vehicles on the market and on-road fleet in the near term. Roos and Siegmann 2020, state in a technology roadmap based on numerous expert interviews that since the technological and also regulatory leap from SAE level 3 to high level of vehicle automation (SAE level 4) is significantly higher than from level 2 to 3, highly automated driving will only be possible between 2040 to 2050. The current obstacles are numerous in quantity and kind. The challenges can be divided into technical and non-technical ones, inspired by the 5-layer model on automated driving (Eckstein 2016). Besides technical aspects (layer 1) the non-technical challenges (layers 2–5) human factors, economic, legal and societal aspects play a major role. These different layers are interlinked with each other and cannot be regarded strictly separately.

Starting with the technical issues, recent Research and Development (R&D) and demonstration projects like HEAT (2020) or STIMULATE (2021) have shown that there are still several technical limitations to be improved. Being able to deal with deviations from the programmed route, e.g. due to road works, dealing with unpredictable road obstacles or being able to drive in all-weather conditions are just a few of them. Liu et al. (2020) state, that there is still a remarkable gap between current state of the art of computing and communication systems and the expected robust system for level 4 and 5 autonomous driving.

Furthermore, implementation of AVs will bring new cyber risks with it, such as hardware and software failures and potential hacking attacks (Litman 2022), which are directly related to societal factors and people’s concerns regarding their safety, privacy and data protection (Fagnant and Kockelman 2015).

From an economic point of view the co-existence of automated and non-automated vehicles on the road is a challenging scenario. Not only do the vehicles have to be equipped with expensive sensor technology, algorithms and chip technologies (which, in addition have recently become difficult to purchase) but the infrastructure also requires expensive recognition and communication technology. Especially in the complex urban environment, the cost factor plays a significant limiting role for the implementation of autonomous vehicles and public transport services.

Very challenging legal aspects still to be solved are security and liability in case of accidents. AVs are ultimately just machines that will be programmed by humans and all decisions made on the road will be based on pre-defined guidelines. Two equally programmed machines would unlikely be involved in the same accident, but when the roads are shared with other users such as cyclists and pedestrians, the scenario becomes more complex, with many unpredictable variables. Currently, there is no societal consensus on how such ethical and moral as well as liability guidelines should be enforced on the vehicles and how accidents should be addressed. It is in any case vital to ensure that the roads can be safely shared among different users. In spite of Germany having already published its ethical guidelines (BMVI 2017), these are not acknowledged across borders. To further complicate matters, machines are susceptible to functioning errors or could even potentially fall victims of cyber-attacks, which would undermine the road safety even if it were well defined from the start. However, there is much progress to develop or update and implement regulations addressing security as well as liability of AVs (UNECE 2020).

One considerable societal impact that needs to be addressed is in fact common to many sectors profiting from digitalization. Replacement of humans by different machines and/or robots will deem some professions obsolete leading to an increase of unemployment. In the case of autonomous vehicles, the redundancy of drivers and other functions in the transport sector might hinder social acceptability of these vehicles. On the other hand, there is a considerable shortage of skilled workers, not only in the transport sector, so that the elimination of the need for a driver could be minimized when addressed with suitable capacity building opportunities.

As mentioned before, the different layers of automated driving are interlinked and need to be addressed in a holistic approach. AVs are attributed with some advantages, such as improved and more energy-efficient driving and consequently the reduction of congestion and road mortalities (see e.g. Krail et al. 2019). However, it is important to emphasize that not all scenarios are as positive regarding these benefits and that AVs can potentially lead to an increase in car-use due to low occupancy rates and travelled distance, and therefore, an increase in energy use (e.g. Acheampong et al. 2021). Therefore, the implementation of a policy framework of governance measures is necessary to ensure that the environmental potentials are exploited and the corresponding risks are minimized. To give an example, without efficient road pricing it may be cheaper for users to have their vehicles to continue driving around the block instead of parking which would add to traffic congestion.

Last but not least, an important factor for the adoption of AVs will be the acceptability by the population in general, which is affected by aspects of all layers such as attractiveness, cost and trust in new technology and appropriate regulatory frameworks, and also the named burden on the environment. Penmetsa et al. (2019) presents a summary of studies focused on public perceptions of AVs. Because of the identified challenges and often for purely psychological reasons, many people simply cannot yet imagine being able to sit back, relax, work or even sleep while being driven and completely relinquishing the driver’s role to the car. This resistance to change may become the main limiting factor once technology has advanced sufficiently. So, while there is a major technology development taking place, society as a whole may not be ready to have vehicles above automation level 4 on the road just yet. The same applies to other road users, who potentially do not feel safe in the presence of driverless vehicles, although the human driver has significantly more sources of error, such as mistaking the accelerator and brake, falling asleep at the wheel or driving drunk. In addition, interaction with other road users has already become much safer thanks to driver assistance systems. One example is the turn-off assistant for trucks, which warns the driver of pedestrians and cyclist in blind spots (Weinrich 2017). Hence, for a successful deployment of full AVs, it is important that the general population is introduced to the use of vehicles where no human will be in control.

Against the backdrop of these multifaceted discussions in science, politics and among the public, and although the societal factor is giveen more and more priority, aspects regarding inclusivity and access to mobility for vulnerable groups are hardly considered. Before we go into this in more detail, however, we should define which vulnerable groups we want to take a closer look at here.

3 How AVs Can Enhance the Mobility of Vulnerable Groups

3.1 Narrowing Down the Definition of Vulnerable Groups

There is currently no clear and established definition of vulnerable groups used uniformly by international organizations and authorities. While vulnerable groups are often considered as those at risk of poverty and social exclusion or with some type of disability, a more comprehensive definition was proposed in the European Recast Reception Conditions Directive (2013). Art. 21 of the directive defines vulnerable persons as:

(...) minors, unaccompanied minors, disabled people, elderly people, pregnant women, single parents with minor children, victims of trafficking in human beings, persons with serious illnesses, persons with mental disorders and persons who have been subjected to torture, rape or other serious forms of psychological, physical or sexual violence, such as victims of female genital mutilation (...).

While the Inter-agency Network for Education in Emergencies also includes this definition, they provide a more comprehensive definition by stating: “Vulnerable groups are physically, mentally, or socially disadvantaged persons who may be unable to meet their basic needs and may therefore require specific assistance. Persons exposed to and/or displaced by conflict or natural hazard may also be considered vulnerable.” (UNHCR 2006).

In addition, there are several pieces of EU legislation on passengers’ rights in the field of transport including the rights of passengers with reduced mobility and disability. These legislations, e.g. the EU regulation on rail passengers’ rights and obligations (European Parliament 2021) gave a human rights based definition on persons with disabilities and those with reduced mobility saying that:

A ‘person with disabilities’ and a ‘person with reduced mobility’ mean any person who has a permanent or temporary physical, mental, intellectual or sensory impairment which, in interaction with various barriers, may hinder his or her full and effective use of transport on an equal basis with other passengers or whose mobility when using transport is reduced due to age;

This definition is significant as it draws the attention to the fact that disability results from the interaction between persons with impairments and various physical, information-communication etc. barriers. In terms of autonomous cars, it means that if persons with disabilities are considered when designing AVs, they could be able to use these cars on an equal basis with other passengers.

These definitions are in line with the vulnerable groups identified by the INDIMO project (Kedmi-Shahar et al. 2020) which include the groups in the table below with their corresponding specific needs.

Table 2. Vulnerable user groups, their share in European society and examples of their specific needs (Eurostat 1 2022, Eurostat 2 2022, Eurostat 3 2022, Eurostat 4 2022, Eurostat 5 2022, EU-KOM 2020, EU-KOM 2021, ENAR 2019, EBU 2022, Kedmi-Shahar et al. 2020)

3.2 Meeting Gaps in the Mobility System - Benefits AVs Can Bring to Users and Vulnerable-To-Exclude Groups

Vaa (2003) shows that due to certain physical and cognitive impairments (e.g. visual, neurological, hearing, medical condition, mental workload, and easy distraction) driving capabilities might reduce with age. Furthermore, elderly people generally seem to have little self-assessment capabilities of their own driving abilities. A paper by Horswill et al. (2011) shows that elderly drivers have poor insight into their own hazard perception capabilities. With one fifth of the European population being over 65 (see Table 2), certain impairments can lead to severe safety limitations for the driver and other road users.

The HADRIAN project shows, even vehicles with SAE level 2 and 3 assistance systems can already take over essential driving tasks. That can enable certain vulnerable groups, like elderly people, to fulfil their own mobility needs, stay independent and at the same time are safe participants in road traffic. At the same time, persons who are too young to hold a driver’s license could also benefit from an introduction of such systems: as the operation of a vehicle becomes easier, the legal age to use it may also be lowered. Furthermore, the detection of people suddenly stepping onto the road and the associated automatic braking contribute to improved overall safety, in particular of pedestrians with cognitive impairments and sensory limitations.

One to two SAE-levels higher, self-driving cars can enable independent mobility for people who temporarily or permanently cannot drive a car at all (elderly, people with intellectual or visual disabilities, epilepsy) or can only drive vehicles with major and expensive adaptations (severely physically disabled). Their dependency on these vehicles is exacerbated if their residency location, e.g. a rural area, does not provide public transport, or if that available is not accessible. Therefore, autonomous cars or buses may be important means of overcoming mobility barriers for those who currently face difficulties in driving or using cars.

During a workshop hosted by the Budapest Association of Persons with Physical Disabilities (MBE) representatives of organisations of persons with different disabilities have explicitly validated the above-mentioned potential of AVs and how they could transform their life by providing them with independent travelling and thus enabling them to actively take part in society (Földes 2017). A survey conducted by Földes et al. (2019) showed furthermore, that the importance of the presence of staff for future users of AVs is not an essential issue for disabled people. The respondents, from which 6% declared themselves to be mobility or visually impaired, did not identify the presence of staff as important, attributing only 1.6 on a scale from not important (1) to very important (3). Additionally, this was also rated as only slightly more important by mobility and visually impaired respondents (1.8).

In summary, that means, by eliminating the necessity of a human driver, people who cannot drive (fully) on their own will be able to move around more independently. This can take place in the private transport sector, for example, through the use of automated driving functions in private vehicles or also in the public sector through the increased offer of autonomous buses or taxis in rural areas, which currently only run a few times a day due to the personnel costs for drivers. In addition, the above mentioned shortage of skilled bus drivers also plays an increasingly important role for public transport operators.

Additionally, autonomous taxis and buses are expected to be a lower-cost transportation option, making them more affordable for non-drivers (Litman 2022). Thus, thanks to the introduction of AVs a multitude of users will become more mobile and able to reach more places (Rojas-Rueda et al. 2017). Meanwhile, they will be less dependent on assistance from other people in their everyday lives and can engage in society more freely. In the future, people could either own private vehicles that are able to drive autonomously or there could be shared vehicles that provide an on-demand service.

Consequently, people living in peri-urban or rural areas will be able to give up their personal cars. Furthermore, car-sharing services can become more flexible as users will no longer be required to pick up a car in a designated area. Instead, users will be picked up directly from their current location.

3.3 Potential Barriers and Risks of AVs for Vulnerable Groups

For all of their advantages, AVs also come with some disadvantages, particularly for certain vulnerable user groups. These groups are especially underprivileged in situations where people could be excluded due to financial means or other access limitations. Furthermore, there could be potential issues in the operation of automated assistance functions, in particular barriers related to the Human Machine Interface (HMI). In this context, people with little or no digital skills would not necessarily be able to use such services. Moreover, although it will likely still take years if not decades until vehicles are driving completely independently on our roads, there may be risks due to the remaining human drivers who tend to act more unpredictably.

In addition, AVs do not only have to interact with other drivers, but they must also avoid hitting moving and non-moving objects, including pedestrians which could be elderly people or people with low vision or reduced mobility who may not act the same or as fast as an able-bodied person would. Since AVs often already drive electrically, they can also be considered relevant for visually impaired people in terms of the associated quiet driving. Elderly people or people with visual and hearing disabilities and even pedestrians who use headphones are at risk of being hit by electric and hybrid cars due to how quiet they are at low speeds. Therefore, these cars must be fitted with an Acoustic Vehicle Alerting System (AVAS), a low-speed alerting sound to keep pedestrians safe. In the event of an accident, details regarding liability are still undefined (Fagnant and Kockelman 2015).

An additional aspect is the underlying infrastructure. In this regard, dramatic change is needed, especially regarding refuelling/charging and maintenance stations which need to be accessible for everyone and parking spots that must offer sufficient space for any user, including wheelchair users who may require ramps. So far, these aspects were not taken into account in city planning, resulting in an access barrier for certain vulnerable groups.

Moreover, there above mentioned the lack of trust in the new technology is particularly high among elderly people who are not used to technology taking over a task they used to undertake themselves. Diepold et al. (2017) found out that about 75% of elderly drivers are not willing to ride with automated vehicles due to uncertainty and distrust in the technology.

In order to address these challenges and avoid these barriers, it is important to identify the requirements of all users at an early stage. Then, the potentials mentioned above can be exploited.

4 How to Overcome the Barriers

4.1 Identifying Potential Requirements of Vulnerable Users When Using Autonomous Cars

The user of an AV must give information to the vehicle by the HMI on the one hand and receive information from it on the other. According to Földes (2019), for the information input, interfaces should be simple and accessible and not rely on one single sensory channel. It is important that control alternatives to vision, auditory, speech and tactile elements is provided, as, for example, visually impaired users would need both tactile interfaces and voice-controlled systems, but other disabilities might not allow for the use of systems that are exclusively voice-controlled. Furthermore, passengers with intellectual disabilities and autism are more reliant on support to navigate them from one place to another. However, for some people a system that offers the possibility for supervision and tracking the journey through video cameras and GPS can help their caregivers.

When AVs potentially share information with the passenger, it will happen either in terms of vehicle maintenance notifications or regarding the vehicle route (such as, current location, progress of ride, potential deviations, etc.). How differently this flow of information can be perceived is shown by a study by Kim et al. (2012) that investigated the impact of multimodal, in-vehicle navigation systems for different aged drivers. They found that, while young drivers benefit from multi-modal navigation systems, older people are oftentimes overwhelmed because of their already high workload and issues concerning selective attention. Avoiding excessive information is advantageous not only to elderly but also to people with intellectual disabilities and/or with autism. This can for example make use of symbols. It is also important to keep in mind that some noises and excessive information can be quite disturbing for people with autism. Moreover, it is essential to implement different sensory channels. Visually impaired users, for example, need both audible and/or braille format, and would also benefit from large fonts, contrasting colours and appropriate illumination. These features would likewise serve elderly passengers and those with hearing disabilities that cannot rely only on audible information systems. Additionally, for these passengers, a proper illumination is crucial for lip reading.

The above mentioned mistrust of technology can also be addressed through HMI. An interview study by Li et al. (2019) focuses on the general design of an age-friendly highly automated vehicle. As mentioned, some people, especially elderly, find it difficult to fully relinquish control. They require information and a monitoring system, to be able to control the behaviour and the decision making of the automated vehicle. Additionally, the takeover requests of an automated vehicle should be adjustable and explanatory. The driving style should be imitative and corrective, such that it imitates the standard driving style of the driver and corrects bad and dangerous driving behaviour at the same time.

The main requirements of passengers with physical limitations, like elderly or people with disabilities, such as wheelchair users, are much more oriented towards the main body and internal arrangement of the vehicle that needs to be accessible and barrier-free (i.e., have wide and step-free doors, easy to use door handles). Therefore, as with current accessible vehicles, adjustable floor to accommodate to wheelchairs, or a ramp/ lift system is needed, as well as adequate space for manoeuvring the wheelchair. Furthermore, to ensure a smooth continuation of the passenger’s trip, the AVs should navigate to accessible disembarking locations where, for example tactile pavement and audible traffic lights exist, or barrier-free space away from traffic.

This list is not exhaustive and could potentially be extended through consultations with the heterogenic groups of persons with disabilities. But what is clear from the results and examples shown is that the requirements of different vulnerable groups for AVs vary greatly depending on the individual user’s impairment.

4.2 Making Use of Universal Design

In 1997, a working group of architects, product designers, engineers and environmental design researchers from the North Carolina State University developed the 7 Principles of Universal Design (NCSU 1997):

  1. 1.

    Equitable Use;

  2. 2.

    Flexibility in Use;

  3. 3.

    Simple and Intuitive Use;

  4. 4.

    Perceptible Information;

  5. 5.

    Tolerance for Error;

  6. 6.

    Low Physical Effort;

  7. 7.

    Size and Space for Approach and Use.

The purpose of the principles is to guide the design of environments, products and communications. According to the Center for Universal Design in NCSU (1997), the principles “may be applied to evaluate existing designs, guide the design process and educate both designers and consumers about the characteristics of more usable products and environments.”

The term universal design is often used interchangeably as design-for-all, which includes the respect of human diversity. Instead of removing barriers, the method focuses on prevention. Products, services, systems and the surrounding environment is designed in such a way that the final product of the design is usable and accessible for the widest possible range of people regardless of their age, gender or capabilities. As a result, more users can be reached and production costs can be reduced by sharing them among a larger market. Moreover, products can adapt to the changing needs through our lifetime without costly and burdensome alterations due to their flexibility.

These cost benefits of accessibility have been supported by surveys, e.g. the European Commission published its Impact Assessment accompanying the document Proposal for European Accessibility Act in 2015 (EU-KOM. 2015). Annex 2 of the document contains the results of the Stakeholder consultations where companies were asked to provide information about how accessibility is considered when providing goods and services, and estimates of the costs and benefits of accessible goods and services. The great majority of the 180 respondents were micro, small and medium enterprises. The respondents generally regarded the extra costs of accessibility to be relatively low, at less than 5% of production costs. 55% of companies that provide accessible goods and services have increased their clientele as a result of improving the accessibility of their goods and services, and 39% have experienced increases in their financial benefits for this reason.

The UN CRPD (2006) mentioned in the beginning brought the definition of universal design to an international binding legal regulation. On page 4, Article 2 of the UN CRPD includes the following definition of universal design: “Universal design means the design of products, environments, programmes and services to be usable by all people, to the greatest extent possible, without the need for adaptation or specialized design. Universal design shall not exclude assistive devices for particular groups of persons with disabilities where this is needed.”

That means, for any design to be inclusive, the built environment, products, services, information and communication technology should work well for everybody. That relates to the transport infrastructure and vehicles, as well as, the transport service. Accessibility and usability are being shifted from being optional to mandatory and it is based on the universal design for all approach.

As a further step, in 2019, a new CEN standard EN 17161:2019 entitled ‘Design for All - Accessibility following a Design for All approach in products, goods and services - Extending the range of users’ was published (CEN 2019). The document helps an organisation to meet their statutory and regulatory requirements regarding accessibility of its goods and services. It promotes accessibility following a Design for All approach in mainstream products, goods and services and interoperability of these with assistive technologies. However, this document does not provide technical design specifications and does not imply uniformity in design or functionality of products, goods and services. The standard is a tool to mainstream a universal design approach throughout the internal process of manufacturers and service providers, which would result in more accessible goods and services.

For improving the access of vulnerable groups to AVs and also the interaction of such persons with AVs in road traffic, the application of Universal Design in the development of the corresponding HMI is a key element. However, Dey et al. (2020) found this is where a major gap still exists. Most existing HMIs for the interaction between human road users and AVs use a single modality (i.e. lights or sounds) and would thus not address road users with special needs, such as people with vision or hearing impairments. Same can be stated for HMIs for the interaction between passengers and vehicles.

But all guidelines are of little help as long as the very differentiated needs identified above are not considered in the development. Therefore, products like AVs which are moving on public roads and consequently have to meet the requirements of a wide range of people including such with disabilities cannot be created without close cooperation between designers and end-users. This is where the concept of co-design has to come in. ‘Co-design is an approach to the discovery, definition and design of products, services and environment that invites the end-users into the design process as active participants. The co-design approach leverages a combination of methods and tools to gain deep insights about people’s experience, latent needs, dreams and aspirations (Sanders 2021).

And furthermore, article 4 (f) UN CRPD states: ‘General obligation of the UN CRPD stipulates that States Parties shall undertake or promote research and development of universally designed goods, services, equipment and facilities, which should require the minimum possible adaptation and the least cost to meet the specific needs of a person with disabilities, to promote their availability and use, and to promote universal design in the development of standards and guidelines.’

5 Good Practice Example: HADRIAN Shows a User Centred Development Approach to Enable Diverse Mobility Needs

A practical example of how user needs should be researched in order to shape the development of automated vehicle functions is the HADRIAN project. Within the project, user-centred mobility solutions for AVs were developed for SAE levels 2 and 3, taking into account different users and their individual (mobility) needs.

The HADRIAN project primarily aims to shape automated driving from a holistic, user-centred perspective by addressing three main pillars: First, a fluid Human Machine Interface (f-HMI) helps drivers and users to appropriately interact with the vehicle. Second, through integrating the AV with road infrastructure the AV is made more predictable and available. Thirdly, an onboard tutoring application teaches the driver to develop safe AV usage skills over time and improve the safety. These innovations practically extend the SAE automated driving levels (SAE ADL). Specific mobility scenarios concerning individual user as well as mobility needs and constraints, have been developed. These mobility scenarios are an important basis for the development of system functions, simulations and tests during the HADRIAN projects. Specific personae have been conceived which all will benefit from the automated driving functions in their specific user context. The imaginative description of their specific driving tasks significantly help system developers designing the technical applications in a user-friendly way according to their anticipated needs.

The HADRIAN partners identified elderly drivers as one important user group that can benefit from the fluid HMI functions. Hence, one of the HADRIAN uses cases describes Harold, a 78 year-old man living in the suburbs of Paris, which start to encounter some difficulties driving his car. Based on Harold, three potential mobility scenarios, including potential obstacles on the way (e.g. difficult intersection, highway entry), have been designed and used as a guideline for the development of the specific HADRIAN f-HMI components.

The first scenario describes a visit of his daughter who lives in the countryside, where Harold gets an adaptive information assistant (sensing system), giving situational information about the state of the environment and depending on the driver’s Fit-to-Drive (F2D) value. It presents an extension to partial driving automation (SAE ADL 2). Harold always stays in the loop of driving but is supported through an adaptive assistant. Only if safety cannot be ensured, a planned emergency stop is executed.

In a second scenario, Harold goes on vacation to an unknown place at the sea. This is an extension to conditional driving automation (SAE ADL 3). Assisted driving takes over when Harold seems not to be able to complete the driving process and actively suggests transitioning from manual to automated driving level (SAE ADL 3).

Finally, Harold wants to visit his doctor in the city. During this trip he will be provided with a “Guarding Angel Protection”. This is an extension of SAE ADLs 3–4 (conditional/ high) driving automation. A Guarding Angel functionality provides a self-enabling safety mechanism against accidents and keeps the vehicle within safe operational boundaries while allowing Harold to actively manoeuver the vehicle within those boundaries. Emphasis is on supporting Harold where necessary without overwhelming him, offering him information or supporting takeover functions as needed and asked for with high transparency of actions (Fig. 1).

Fig. 1.
figure 1

© 2019 Hadrian.

Three Scenarios of Harold. a daughter visit in the countryside, b vacation at the sea, c doctor visit

The above described mobility scenarios led to detailed application descriptions (DAD) as design basis for the development of the specific HADRIAN f-HMI components, including “Increased Automated Driving Predictability”, “Human-Centred Fluid HMI”, “Adaptive Seating Orientation”, “Visual Head-UP Display”, “Haptic Feedback on Steering Wheel”, “Ambient Light Indicators”, “Fluid Interface Design”, “Tutoring System”. Hereby a task analysis investigated the implications of using the HADRIAN innovations (see Fig. 2) within the mobility scenarios, leading to initial requirements for vehicles, road infrastructure and the driver/user for the HADRIAN operational concept.

Fig. 2.
figure 2

© 2019 Hadrian.

Relation of main technical and procedural HADRIAN innovations, modes of automation and the corresponding mobility scenarios

Figure 2 also shows which HADRIAN Innovations (indicated with HI# at the top) support the Harold mobility scenarios (indicated with H# at the bottom of the figure). In the first mobility scenario (H1), Harold is supported by an awareness assistant (HI1), active driver monitoring & fluid interventions (HI5) and adaptive tutoring (HI6), marked though the lines connecting the HADRIAN Innovations and the Modes of Automation. Those innovations serve as a manual driving aid for elderlies. In the second mobility scenario (H2) support will be realized by providing minimum guaranteed time for human driver to transition from automated driving to manual driving (HI3), guaranteeing minimum duration of automated driving at level 3/3+before the trip (HI4), as well as HI5 and HI6. During the third mobility scenario (H3) HI5, HI6 and the Guarding Angel safety protector (HI7) support Harold.

The DAD are the basis for legal and ethical considerations for the HADRIAN operational concepts as well as considerations for driver information needs, knowledge and skills. The special needs of elderly drivers and the corresponding DAD have been discussed and verified in focus group discussions with invited elderly people. The participants (65+years) were introduced to Harold, his main driving challenges and four to six scenario segments with specific driving situations. Detailed description and presentation of the scenarios allow a realistic understanding of the situations, which might come up in connection with automated driving. Following the video presentation, they were asked to rank the importance of given ethical values “privacy”, “autonomy of the driver”, “safety”, “security”, “vehicle performance” and “costs” in relation to the different driving scenes. This specific user group discussion provided basic information for the HMI design in the vehicle. User-centred design principles could be developed. Additionally, the potential risks concerning the driver vehicle interactions are evaluated with respect to the DADs, and tests and simulations are operated in alignment with the different mobility scenarios.

The identification of mobility needs of specific user groups and the translation into detailed requirements for the vehicles and, in this case, the HMIs is a recommendable approach to the start of the technical development of new functions. Road infrastructure, vehicles and drivers themselves will benefit from greater acceptance within that user group. The automated driving innovations will be relevant to the users and also support inclusivity. Therefore, this approach will lead to a successful implementation of automated driving, also for wider and more versatile user groups.

Specifically, the Harold mobility scenarios provide an example for the use of vehicle automation on levels 2 and 3 for the benefit and inclusion of elderly people. Certain impairments of elderly people and based on that specific mobility needs, mobility challenges and driving requirements are also relevant for other vulnerable user groups (e.g. people with physical disabilities, novice drivers). By designing the “diving process”, a more detailed picture of the actual needs for certain user groups can be drawn and user specific HMI components can be develop or adjusted. HADRIAN shows how to approach this with view on different and special user needs. This approach can easily be transferred to other vulnerable user groups. Furthermore, the focus on inclusive solutions for automated driving systems and the HMI will be beneficial for all vulnerable user groups. Ongoing discussions on legal and ethical issues and safety concerns of automated driving can only be profitable.

6 Conclusion

One important function of transport systems is to provide accessible mobility for all. Legal requirements both in the EU and in its member states have grown to address the needs of everybody in new investments and legal frameworks have been established to ensure non-discrimination and equality. Increased life expectancy and ageing of the population requires countries to establish policies, which enable the elderly and people with disabilities to live independently and be active members of society for as long as possible. However, the ongoing transformation of the mobility sector has not always been user-oriented, neglecting at times the needs of different people. To ensure real inclusiveness, it is important to consider the needs of every citizen when launching new commodities in society.

The implementation of AVs forms an opportunity to empower vulnerable citizens, namely people with disabilities and elderly people in fulfilling their mobility needs. This technology development will help people with disabilities and age-related impairments to live more independently, with the additional benefit of improving their participation possibilities in education and the labor market, and in general having an as active life as they want.

On the other hand, like many (digital) technological developments, there are still several hurdles to overcome to fully address the needs of vulnerable groups during the deployment of automated systems and vehicles. Universal design as an essential tool to accessibility must play a major role in the development methods of manufacturers. By following this design method, AVs will be usable and accessible for as many user groups as possible. Therefore, knowledge and good practices of Universal Design shall be promoted among AV and HMI developers from an early stage on.

To ensure that the HMI used in AVs, transportation infrastructure, and the vehicles themselves are usable and accessible to all, potential users, especially people with disabilities, must be meaningfully included in the design process throughout the development process until the final product is presented. Co-creation and participative initiatives are critical methods that offer this possibility, bringing together industry experts and developers, researchers and users to work together and develop better products. This has in fact been demonstrated in projects such as HADRIAN where the user needs have been the main focus of the work and were well integrated into the development process. Members of vulnerable groups have established organizations from grassroots to national and even European and global level. The involvement of these organizations’ representatives in the development and design of autonomous vehicles will lead to more accessible and usable solutions.