Navigating autonomous demand responsive transport: stakeholder perspectives on deployment and adoption challenges

ABSTRACT

As cities implement intelligent systems and public transportation users seek flexibility, autonomous demand-responsive transport (ADRT) will play an important role in a region’s broader public transportation system. We have seen this in cities where trials have been conducted. Great efforts and investments are being made in the industrial and academic sectors to make autonomous driving a reality. This study investigates the challenges and opportunities of adopting ADRT in South East Queensland, Australia. This is critical as ADRT can help be an alternative way to provide public transport, especially for first and last-mile service. The adapted methodology is an exploratory qualitative study through interviews with transport experts to provide comprehensive insights regarding the ADRT service concept, potential customer groups in terms of the supply and the demand side, potential use cases, followed by an explanation of the ADRT deployment and adoption challenges in terms of the supply side and the demand side. Accordingly, we offer recommendations mainly contributing to mindset change, education and exposure, subsidies and incentives, liability and insurance that can support policymakers, transport planners, and engineers in making effective policy choices and developing successful transportation systems, while also increasing public recognition and adoption of ADRT technology.

KEYWORDS:

• Autonomous demand responsive transport
• pubic transport acceptance
• adoption challenges
• autonomous mobility
• autonomous vehicle
• smart mobility
• transport policy

1. Introduction and background

The emergence of electric and autonomous vehicles is the biggest disruptor facing the Australian transport industry today. To turn this challenge into a greater opportunity, the industry could focus on embracing and integrating these technologies into their existing operations, developing new infrastructure and charging stations for electric vehicles, investing in the research and development of autonomous vehicles, and reducing their carbon footprint (Faisal et al. 2021). Additionally, embracing the sharing economy model could be an opportunity for the industry to revolutionise the way people move around cities, positioning themselves as leaders in the global movement towards more sustainable and efficient transportation systems (Golbabaei et al. 2023).

Government initiatives to provide more customer-focused services have elevated Demand-responsive Transit (DRT) from a specialty mode to a widely accepted mode of public transportation (Hensher 2017; Butler, Yigitcanlar, and Paz 2021a). This mode provides door-to-door or stop-to-stop travel in response to passenger requests using vehicles that vary in size and operate on flexible routes and schedules. Despite its benefits, DRT has faced challenges as private and public providers have had to depend on subsidies to cover their costs, largely due to low demand (Currie and Fournier 2020). The need for extra resources to schedule services on-demand and process requests, which are now made through technology, has added to costs during a time when low participation is a major concern. This, combined with the inflexible structure of many bus contracts, has led to DRT facing significant institutional challenges, resulting in some governments phasing out or prohibiting its provision (Perera, Ho, and Hensher 2020).

Governments are committed to encouraging the adoption of eco-friendly transportation solutions that offer a more convenient and efficient experience for urban dwellers. With cities becoming more densely populated and the overuse of personal vehicles contributing to traffic and environmental problems, the arrival of Autonomous Demand Responsive Transport (ADRT) is a desired development. Leading-edge technology allows ADRT providers to deliver a seamless experience, and the business and academic communities are investing heavily in making autonomous driving a reality. ADRT also has the potential to significantly alleviate traffic congestion, especially in the form of autonomous shuttle buses (Palmer, Dessouky, and Abdelmaguid 2004; Arbib and Seba 2017; Golbabaei et al. 2020; 2021). If cost savings and the ease of not having to find parking spaces can convince people to switch, ADRTs have the chance to replace conventional cars. Some reports predict replacement rates between 5-30% (Bierstedt et al. 2014; Davidson and Spinoulas 2016; Butler, Yigitcanlar and Paz, 2020). However, the continued use of public transportation, with policies aimed at reducing the use of personal vehicles, is assumed to play a big role in determining the success of this transition (Webb, Wilson, and Kularatne 2019).

There have been over 30 automated vehicle trials in Australia to date, in every state and territory. Based on National Transport Commission (2020), the majority of trials have involved low-speed automated shuttle buses operating on set routes in various parts of Australia including coastal areas in the South East Queensland (SEQ) region, such as Karragarra Island, Cleveland, and Main Beach. SEQ, an Australian metropolitan area centred on the capital city of Brisbane (see Figure 1), has a land area of 35,248 km² and a population of 3,817,573 million (2021). The per capita Gross State Product of Queensland is AU$71,037 (US$53,280) (Yigitcanlar et al., 2022). Despite their advancement, public confidence in these leading-edge technologies is still in its early stages (Haboucha, Ishaq, and Shiftan 2015; Roche-Cerasi 2019; Mouratidis and Cobeña Serrano 2021). This presents a critical obstacle that must be overcome to ensure the success of the project, as the level of public trust plays a direct role in the shuttle’s ridership and efficiency (Nordhoff et al. 2018; Chen et al. 2020; Papadima et al. 2020). The transportation industry can assess public support to make well-informed decisions on policy and management, such as creating pricing or subsidy models. The fully automated and renewables powered ADRT can then be tailored to meet public attitudes and preferences (Golbabaei et al. 2022; Wang, Pei, and Fu 2022).

Figure 1. The South East Queensland local government area (LGA).

While the role of government in shaping autonomous transport is widely recognised (Porter 2018), a comprehensive understanding of how it impacts the field is lacking. Governments have several options to promote the next stage of automated transport. Additionally, there is a chance to evaluate Australia’s preparedness for the widespread use of these vehicles by considering multiple factors, such as trials, regulations, infrastructure, and public perception. Previous studies on the topic have primarily focused on regulation (Shladover and Nowakowski 2019; Mordue, Yeung, and Wu 2020), ignoring other methods of governance. The literature also tends to concentrate on how to react to the technology rather than exploring proactive efforts commonly seen in other innovation areas (Salmenkaita and Salo 2002). Governments are crucial in guiding technological advancements to prevent market failures and negative consequences, particularly concerning disruptive ADRT technology (Moon and Bretschneider 1997; Stone, Legacy, and Curtis 2018).

Their role in fostering innovation in transportation includes both direct methods like R&D incentives and indirect ones like regulations. The timing of such interventions is pivotal for their effectiveness (Docherty 2018; Freemark, Hudson, and Zhao 2019). Balancing the need to protect public safety and encourage autonomous transport innovation can be a challenge for governments (Shladover and Nowakowski 2019). To promote the adoption of electric vehicles, governments can implement measures like emissions standards, charger safety regulations, and standardising vehicle requirements (Steinhilber, Wells, and Thankappan 2013). However, government approaches to ADRT innovation will differ globally, depending on the country’s economic, technological, social, and political context (Li et al. 2019). ADRT technologies will require changes in planning, policy coordination, infrastructure investments, and revenue collection at all levels of government (Freemark, Hudson, and Zhao 2019; Mladenovic 2019; Manivasakan et al. 2021). Sperling, van der Meer, and Pike (2018) recommend that national governments focus on vehicle design, while local governments focus on vehicle use.

The majority of literature on autonomous transport policy focuses on adapting current and future regulations to accommodate ADRT technology. There is a lot of interest in how safety, cyber security, privacy, licensing, and liability rules will need to be revised (Claybrook and Kildare 2018; Freemark, Hudson, and Zhao 2019; Lee and Hess 2020). Shladover and Nowakowski (2019) note that while road vehicle regulations have been stable, autonomous technology disrupts this by removing the human driver from decision-making. These regulatory challenges, influenced by ethical and value-based interpretations, hinder the technology's advancement (Claybrook and Kildare 2018; Mordue, Yeung, and Wu 2020). Furthermore, strict governmental oversight is slowing down the technology's commercialisation (Sperling, van der Meer, and Pike 2018). Hence, it is crucial to understand not only how ADRTs will affect regulations but also how these regulations will impact ADRT development.

Thus, to gain an in-depth exploratory understanding of enablers and hurdles in adopting ADRT in SEQ, expert interviews are appropriate to drive additional early adoption measures owing to their proficiency in this area (Grimsley and Meehan 2007; Wu 2011). As per Herrenkind et al. (2019), qualitative approaches could provide insights into the factors that influence whether a new product or service is accepted or rejected. Qualitative methodologies enable us to discover opportunities and barriers to the deployment of innovation, particularly in the market launch phase (Quiring 2006). This approach does not simply extract information from collected data but also treats the data as a living entity, leading to solid and trustworthy discoveries (Holloway and Todres 2003; Sandelowski and Barroso 2006).

Similar methods were employed in the latest adoption studies in the transport field to create a solid foundation for acceptance criteria (e.g. Amann 2017). Mars, Arroyo, and Ruiz (2016) examined travel behaviour studies that employed qualitative analysis, while Simons et al. (2014) utilised grounded theory to uncover the forces that shape the travel mode choices of students and working young adults. Qualitative studies have also delved into the factors that influence the selection of specific modes of transportation, such as bicycles (Fishman, Washington, and Haworth 2012; Sherwin, Chatterjee, and Jain 2014), personal vehicles (Gardner and Abraham 2007; Nguyen-Phuoc et al. 2018), and public transit (Beirão and Cabral 2007). Additionally, in an earlier study, travel self-containment in Australian master planned estates were analysed (Yigitcanlar et al. 2007).

This exploratory study aims to investigate feasible ADRT service models and potential customer groups as well as deployment and adoption challenges and mitigations to establish this concept from transport stakeholders’ perspectives. For our qualitative study, we designed a semi-structured questionnaire, to address three following research questions:

1. What do transport experts believe the feasible ADRT service models would be?

2. Who do transport experts believe the potential customer groups of ADRT would be?

3. What do transport experts believe the ADRT development and adoption challenges are?

4. How could transport decision-makers mitigate the ADRT deployment and adoption challenges?

Following this introduction, and review of relevant policy studies, Section 2 explains the adapted qualitative survey and analysis method. Section 3 details the findings regarding the ADRT service concept, potential customer groups, and potential use cases, followed by an explanation of the ADRT deployment and adoption challenges and the mitigation strategies. Section 4 discusses the main takeaways and exemplifies the policy recommendations and limitations and future research potentials. Section 5 presents the concluding remarks.

2. Methodology

The following sections describe the data collection procedure of the interview sessions.

2.1. Questionnaire design

There are three categories of interview approaches: (1) structured, unstructured, and semi-structured interviews. Structured interviews use a predetermined set of questions; while the unstructured type is conducted spontaneously, with no pre-determined questions. Semi-structured forms are intertwined between these two categories and are the most used type in qualitative research (Robson 2002).

The semi-structured interview incorporates predefined open-ended questions, which enable interviewees to convey their perspectives in their own words and give further detail about the main topics of the conversation (Robson 2002; Smith, Flowers, and Larkin 2009). Associated sub-questions help to maintain the consistency of the structure of the study and data acquisition process, hence contributing to the accuracy of the information attained (Saunders, Lewis, and Thornhill 2007). The semi-structured interview is considered to be the most appropriate method for exploratory research. According to Seidman (2013), semi-structured interviews enable us to uncover and acquire additional subjective findings. However, spontaneous additions were acceptable to capture more comprehensive and expressive responses where applicable (Liu, Nikitas, and Parkinson 2020).

The questionnaire guide is summarised in the following domains. For each domain, several questions were posed to cover the main aspects of that domain (Butler, Yigitcanlar and Paz, 2021b).

1. ADRT user adoption perspectives

   • Who do you believe would be the potential customer groups for autonomous demand-responsive transit?

   • Do you believe that uptake of autonomous demand-responsive transit would vary between potential customer groups, and if so, how?

2. ADRT deployment challenges

   • Do you believe that the needs of ADRT transit would vary between potential customer groups, and if so, how and why?

   • What characteristics of autonomous demand-responsive transit do you believe are important for potential customers to accept this mode?

   • What characteristics do you believe are important to other road and transport system users to accept autonomous demand-responsive transit in the fleet?

3. ADRT planning recommendations

   • What do you believe are the most significant barriers to the success of autonomous demand-responsive transit, and how do you believe each barrier could be mitigated?

   • Do you believe that incentives would be important to encourage potential customer groups to use autonomous demand-responsive transit, and if so, in what form/s?

2.2. Data collection

The sample size in a qualitative study can vary, but for semi-structured interviews, a small sample can suffice (Qu and Dumay 2011). Typically, most of the interviews per study fall between 11 and 20 (Marshall et al. 2013). Saturation, which is the point where all questions have been fully explored and no new ideas or themes emerge (Braun and Clarke 2013), usually occurs after 12–24 interviews (Guest, Bunce, and Johnson 2006; Hennink, Kaiser, and Marconi 2017). The data collection for the study was halted after 26 interviews (with two extra interviews) to ensure a diversity of viewpoints. According to Guest, Bunce, and Johnson (2006), most findings in qualitative research – over 70% – can be gained from the first six participants. By the time 12 interviews have been conducted, saturation is reached and 92% of the findings have been captured.

The first interview was the pilot Interview, which uncovered a few small deficiencies that were later reviewed and modified (Hilgarter and Granig 2020). All participants were interviewed individually to prevent the creation of any bias. At the beginning of each interview, the interviewer introduced a summary of the research and the purpose of the interview. The interviews were carried out in April/May 2022, with an average duration of 40 min. The interviews were audio-recorded and transcribed afterwards for further analysis. The general flow of the interview process is illustrated in Figure 2. Approaching appropriate interviewees is crucial for ensuring the credibility of the gathered information. According to Bolger and Wright (2011), multiple-perspective interviews help prevent one-sided assessments and give consistent insights and a broader spectrum of details.

Figure 2. The general flow of the semi-structured interviews.

As per Eisenhardt (1989), the current study employed a purposeful sampling technique to select the participants rather than random sampling (Esmaeilpoorarabi et al., 2018). According to Liu, Nikitas, and Parkinson (2020, 70), ‘in-depth interviews with members of the scientific, and industry elites provide valuable insights that although could be critical to the exploration of a research topic, may not be obvious to the general public. This is because information on how elites perceive situations and make key decisions provides a unique perspective that often cannot be obtained through other data collection methods’. Hence, the target participants for this part of the study consisted of elites who occupy management and senior positions in various transport sectors including state transport authority, consulting, peak body, service providers, and academia who were engaged with Australia’s industry environment. Table 1 outlines the key characteristics of the elite interview participants.

Table 1. Expert interviewee details.

No.	Affiliation	Sector
1	Department of Transport and Main Roads (TMR), Policy Director / Academic Partner	state transport authority
2	TMR, Executive Director (MaaS Program Management Office)	state transport authority
3	QUT, Deputy Director	academia
4	ADVI-ARRB, Director	consulting
5	ITS Australia, CEO	peak body
6	ARRB, National Discipline Leader	consulting
7	ITS Australia, FvaT Industry Reference Group	peak body
8	TMR, Principal Business Analyst	state transport authority
9	Griffith University, Professor	academia
10	NZTA, Technology advisor	state transport authority
11	iMOVE, Education Manager	academia
12	Department of Infrastructure, Director of Passenger Transport	state transport authority
13	Monash University, Director of Public Transport Research Group	academia
14	APAC, Managing Director	service provider
15	PATREC, Research Fellow	academia
16	TMR, TransLink Devision, Principal Network Planner	state transport authority
17	Technical principal transport planner	consulting
18	Griffith University / Liftango	academia
19	The University of Sydney, Senior Research Fellow	academia
20	University of Western Australia, Professor	academia
21	Deloitte, Transport & Infrastructure Partner	consulting
22	RACQ, Transport Planning and Infrastructure Advisor	service provider
23	TMR, Infrastructure Management and Delivery, Chief Engineer	state transport authority
24	University of South Australia, Professor	academia
25	LaTrobe University, Professor	academia
26	The George Institute, Professor	academia

We recruited participants by sending email or LinkedIn messages with interview invitations to the target sample. We also used a degree of snowball sampling (Biernacki and Waldorf 1981), whereby participants who were directly recruited referred colleagues who would be interested in participating in the research interview. No monetary participation bonuses were offered for recruiting purposes. In total 40 experts were selected and then officially invited to an online interview, of whom 26 replied and participated in the interview. Prior to their participation, the interviewees were told of the interview settings; they were advised and agreed that the sessions would be audio-recorded and transcribed, but the data would be anonymous (the requirements of the ethics approval from the university human research ethics committee; UHREC reference number: 2000000747) and used exclusively for the research purpose.

2.3. Data analysis

We followed Braun and Clarke’s (2006) impulsive thematic analysis technique, which aims to provide a dynamic foundation as opposed to the rigid structure used by the typical codebook approach. Table 2 illustrates the six phases of our thematic analysis approach.

Table 2. Phases of thematic analysis (Braun and Clarke 2006).

Thematic analysis was used in prior transportation research (Gössling, Cohen, and Hares 2016; Hafner, Walker, and Verplanken 2017; Alyavina, Nikitas, and Njoya 2020; Liu, Nikitas, and Parkinson 2020) and proved to be a ‘sophisticated qualitative tool that allows for conducting research in a precise, consistent and exhaustive manner through recording, systematising, and disclosing the methods of analysis and the study results with enough detail to enable the reader to determine the credibility and validity of the process’ (Nowell et al. 2017; Liu, Nikitas, and Parkinson 2020, 72). According to Vaismoradi, Turunen, and Bondas (2013), qualitative descriptive methods like thematic analysis are better suited for research that calls for less interpretation, as opposed to the grounded theory which demands a more in-depth level of interpretation. Given the aim of this study, and the capability to uncover sufficient information with a minimal level of interpretation, thematic analysis was selected as the research method to identify, analyze, and present the overarching patterns (themes) found within the data (Braun and Clarke 2006; Nikitas, Avineri, and Parkhurst 2018; 2019).

Thematic analysis in literature can be approached in two ways: the inductive and deductive methods. The deductive approach follows a pre-existing theory as a roadmap (Bengtsson 2016; Lune and Berg 2017), while the inductive approach allows analysing the contents with an open mind to uncover unanticipated topics and perspectives from interviews with expert participants that answer the research question. It also enables appreciation of the formal and informal dynamics involved in the policymaking process (Yigitcanlar & Dur, 2013). As suggested by Glaser and Strauss (1967) and Liu, Nikitas, and Parkinson (2020), the interviews were conducted, transcribed, and analysed through a manual process that involved multiple readings and annotating of the interview transcripts. Thematic analysis was chosen as the preferred method of analysis due to its versatility and the ability to uncover rich, complex accounts of the data. The freedom provided by the theoretical nature of thematic analysis allowed for an in-depth examination of the data, providing a detailed and nuanced understanding of the subject matter (Braun and Clarke 2006).

As recommended by Pettigrew, Cronin, and Norman (2019c, p.3), data coding is completed by a single coder due to ‘the little prior work in this area, the highly exploratory nature of the study, and the resulting emergent nature of the coding hierarchy’. Erlingsson and Brysiewicz (2017) explain a code as ‘a label; a name that most exactly describes what this condensed meaning unit is about. Usually, one or two words long.’ In a similar vein to Pettigrew, Fritschi and Norman (2018) and Dichabeng, Merat, and Markkula (2021), the coding hierarchy covers deductive notions (factors extracted from the literature on AV adoption) and inductive notions (topics raised by experts). As a kind of member checking, emergent interpretations are discussed with the following interviewees to evaluate their utility (Wallendorf and Belk 1989). The emerging themes were reviewed and reached consensus on the final interpretation.

Through the process of identifying themes, we note that some themes may have overlapping aspects, and a few statements may represent more than one theme. This, however, is not an issue as ‘the themes and the way these relate to each other do not have to be smoothed out or ignored but instead retain the tensions and inconsistencies within and across data’ (Braun and Clarke 2006; Liu, Nikitas, and Parkinson 2020, 73). It is important to emphasise that the themes were derived based on the importance of the interviewee’’ perspectives, not just the number of participants mentioning a theme (i.e. theme frequency). The frequency does not necessarily indicate the validity of the theme (Yigitcanlar & Bulu, 2015).

3. Results and findings

This qualitative study revealed five main themes from the thematic analysis including:

1. ADRT feasible service models

2. ADRT potential customer groups

3. ADRT deployment and adoption challenges

4. ADRT deployment and adoption policy recommendations

The findings of the study are presented in the following sections, which are organised according to the themes identified. Each theme is thoroughly described with its sub-themes and quoted examples from the participants to provide a clearer understanding of the concepts. The themes, while distinct, may have overlapping dimensions, which are reported as sub-themes in the scope of this study. To convey the thematic analysis effectively and objectively, selected relevant quotations are used to provide evidence for the themes presented. This method of presenting the findings through quotes is one of the most concrete ways to deliver the analysis (Nikitas, Wang, and Knamiller 2019; Liu, Nikitas, and Parkinson 2020).

3.1. Feasible service models

Across all stakeholder groups, there was agreement that there are several scenarios for offering ADRT service in SEQ i.e. the first mile/last mile, access/egress mode from transport hubs or localised on-demand transit on a more commercial note.

Generally, it is foreseen that the first/last mile integration is going to be more attractive to an on-demand service that feeds into a transit trunk line. So, the needs will be quite different and the service models that support their technology will also be quite different. This is probably about distinguishing different technology for every single customer growth segment. It will also indicate employment types, destination choice and location of origin. These matters will depend upon the service provider and the technology provider to accommodate:

• The needs around trip facilities that need to be accommodated, i.e.

  • − If any disabled persons may require additional services.

  • − If people are travelling in a group.

  • − If there is luggage.

• The needs around the systems and technologies that need to be supporting this actual service, i.e.

  • − Selection criteria that might make people feel more comfortable. Depending on the technology and the volume of subscribers, we can consider more tailored offerings for people to build their confidence.

  • − Understanding who else might be in the vehicle with the passenger.

  • − Being able to choose a vehicle with passengers of the same gender.

• The needs around trips, i.e.

  • − If they have multiple stops because they are setting down multiple passengers, reliability is important regarding when they are going to leave and when they are going to arrive.

  • − There will be matters that might need to be tailored in the future because the timeliness of the trip, the cost of the trip, and the reliability and convenience of that will be the drivers of uptake.

Many experts agreed that the vehicle concept requires consideration of the trade-off between costs, feasibility, demand, and government subsidy. So the real factor in driving service design is the distribution of origins and destinations of passengers, as evidenced in the following quote:

Let the private industry service the mass market for on-demand ADRT and then I’ll let the right agencies deal with areas that don’t stack up from a commercial player to operate or, as is the case, some agencies might have both and then they just subsidize or incentivize the road the transport provider to service those other lower volume routes with the money that they earn from the high-volume routes. These are all existing models for running public transport that I think cannot be adopted. (Consultant)

Of note is that dysfunctions in the network will drive the behaviour of people using the network that is offered. The challenge for adoption is to enable a system that is competitive with customers’ private vehicles in terms of seamlessness and benefit.

3.2. Potential customer groups

The interviewees considered the potential customer groups from both the supply and demand sides.

3.2.1. The supply-side

It is foreseen that the customer groups will be state government, local government, and public transit operators to expand their transit networks in low-density suburbs. There is a huge growth potential for these types of services.

In Queensland, public transport services are currently contract-based, and tenders are awarded by the state government (Translink Division of Queensland Department of Transport and Main Roads). So, the state government will be very interested in how they can increase the efficiency of these services and run some trials in SEQ with providers. This may see current bus operators, fleet operators, and new providers entering the market.

3.2.2. The demand-side

In terms of the demographic groups, as expressed by some experts, one notion exists that the customer group could be all citizens, as a full range of market opportunities exists for other groups including all age profiles (most people between the ages of 8- 80 should be able to use these vehicles) and all genders, particularly to free up their time. For example, it would be ideal if a customer could board a level 5 autonomous shuttle that is physically able to move in the network, they could perform other tasks quite comfortably, such as writing or typing or making phone calls. It would also fulfil the purposes of different customer groups including partygoers. Supply-side could be influenced by variations that would depend on the range of choices available, and more importantly, system management.

Despite reforming the Disability Standards for Accessible Public Transport (Vintila, 1996), not all transport systems have been retrofitted, meaning that equal access remains unavailable for everyone. It is still a necessity of the future value proposition of autonomous vehicles to be an all-inclusive transport mode and be able to provide mobility for everyone, whether they are elderly, have a disability, or live in a regional area.

Transport disadvantage and social exclusion are among the major problems many urban communities are facing today (Kamruzzaman, Hine, and Yigitcanlar 2015; Yigitcanlar et al. 2019). In this case, as foreseen by the majority of experts, the early adopters will probably be transport disadvantaged groups including:

• Elderly people (senior citizens) and retirees, who do not drive and therefore have lost the ability to be independent travellers.

• Disabled people (with different types of disabilities) who are unable to travel on their own.

• Peri-urban/rural dwellers who don’t have regular bus services or taxi services in fringe communities that are currently not well connected.

• Younger members of the community who are probably more technology enthusiasts, however, children are dependent on the decision of their parents, so it is needed to gain the trust of their families.

• People who do not have multiple cars in their household and who do not have car-based options.

• People from lower socioeconomic backgrounds, if the price is satisfactory and for those who are interested in a novel mobility solution.

In terms of the use case, based on the infrastructure requirements around aVs, most respondents viewed it as a precinct-specific solution for people. They noted that as it is not a speedy form of transport, it would not be suited to long distances. It would be suitable for short-distance activities in low-speed environments with low volumes of traffic, including:

• Nursing homes, retirement communities, and aged care facilities that are well suited to assisting people who need to move at slow speeds.

• Hospitals that are seeking to reduce missed appointments, as it is often quite difficult to either reach the venue and/or find a parking space, particularly where customers may be sick or injured.

• Airports where people want to limit their time on the ground side of the venue.

• University campuses, business parks, central business districts or potentially the athlete villages.

Generally, from a safety standpoint, the operation of ADRT services is more feasible in locations where there is more control over the vehicle’s interactions. These areas are the most appropriate locations to begin, and they are where most of this work has been done globally. However, it will make progress with time and more ADRT use cases would be developed in SEQ. As exemplified by one participant:

In our review of the failure rates of DRT systems in the last three or four decades, the autonomous shuttle trials emphasize technology beyond practice. Since their capacity and speeds are not currently at all within any bounds of a reasonable quality of service, they can only really be deployed in very special situations where fixed-route services are too expensive. Higher capacity systems with lower marginal and operating costs compared to autonomous vehicles and capacity could work at the moment. (Academia)

Overall, there are different customer segments today, which will have different views of ADRT, however in time and as customers become more familiar and comfortable with new technologies, it is feasible that uptake will increase.

3.3. Deployment and adoption challenges

During the interview, the experts were asked to provide their views regarding the challenges and barriers to ADRT development and adoption, particularly in SEQ. The major overarching challenges discussed by the interviewees emerged in four core themes comprising: technical, financial, regulatory, and behavioural challenges (see Figure 3).

Figure 3. Overarching challenges in deployment/adoption of ADRT.

3.3.1. Technical challenges

The technical challenges were categorised into 3 subcategories: technology, management system, and infrastructure as depicted in Figure 4.

Figure 4. Technical challenges in deployment/adoption of ADRT.

3.3.1.1. The technology

A recurring view was that technical development is not fully mature yet. Accordingly, there is a limited number of use cases and practical applications. The current technology is unable to accommodate all possible situations. For instance, one participant highlighted:

The vehicles still struggle with rain, and snow, and have problems identifying potholes at times. There are all sorts of little things that are still being ironed out, but we are much closer than we’ve been at any time in history. The safety and security issue of vulnerable passengers is something that needs to be resolved again. (Academia)

Several industry-wide challenges have not been ameliorated, and more work is needed. The procurement of technology is a particular issue it presents. Australian DRT service trial in Shellharbour, NSW in the early 1990s failed because of technical issues stemming from poor planning and excessive reliance on unproven technology. This led to dissatisfaction among operators and a loss of enthusiasm, eventually resulting in the return to traditional service (Perera, Ho, and Hensher 2020). The followings are the main technology challenges mentioned by the interviewees:

• – Unmature simulation environments/ commercial models

• – Shake-out of players, particularly after COVID

• – Has taken longer than what was anticipated

• – Not able to build it on the scale

• – Not being an attractive market for foreign direct investment

• – Safety and security issues of vulnerable passengers

• – Struggling with severe weather and topography

3.3.1.2. The management system

During the discussions, the interviewees indicated that along with the technology, it is necessary to manage how to deploy ADRT systems because they need to interact between users, transport infrastructure and other vehicles. For instance, one participant from academia highlighted:

At the moment, most trials are pretty much like the entertainment park. They’re just road based with very slow speeds, and everything is controlled and none of them is on demand. But this does not mean that an algorithm cannot slightly change. (Academia)

The Dial-a-Bus programme in Adelaide, South Australia was a flexible ‘many-to-many’ service without a set schedule and ultimately failed due to a lack of passenger demand and practicality. Although the elimination of spatial and temporal restrictions on passenger pickup and drop-off enhanced flexibility, it reduced the effectiveness of public transportation due to decreasing the capability of combining trips. In scenarios of low demand, a more rigid service, such as a pre-booked option or fixed-time operations, is more appropriate. High flexibility levels could be maintained just in case of ample travel demand (Perera, Ho, and Hensher 2020). Furthermore, many system choices and protocols exist. The main management challenges extracted are:

• – Global providers are remote from the context where they are providing their apps

• – Global providers do not understand local issues or are not concerned with the local policy

• – Private sector managers of MaaS raise the price because of their business model

3.3.1.3. The infrastructure

The need for a clear picture of what infrastructure is required, whether on the roadside, mobile coverage or connectivity, was strongly emphasised, with some participants worried that supporting infrastructure has not been fully addressed, particularly at set-down/ pick-up locations. For example, one participant claimed that

The autonomous shuttles cannot run everywhere by themselves and there should be a hub and spoke for high-volume public transport access into urban form areas. Because the autonomous shuttles of any type won’t be too congested to meet the needs that your customers want. (State transport authority)

The followings are the main infrastructure challenges mentioned by the interviewees:

• – Supporting infrastructure has not been fully addressed, mainly at set-down/ pick-up locations

• – Nobody wants to invest in infrastructure as they do not know what the requirements are, which increases the cost of services

• – Shuttle can not deal with heated climate, so running the air con would need more recharging

3.3.2. Regulatory challenges

The regulatory side is evolving as technology is evolving. A key challenge that has already been heavily discussed in the field and was mentioned in the interviews surrounds very few jurisdictions that have an agreed future regulatory framework. In some jurisdictions, some frameworks are not adapted to current technology. Accordingly, many respondents highlighted the importance of having consistent legislation and operational regulatory framework in one form or another. As indicated by one comment from a service provider

At the moment there are no two frameworks that are the same throughout the world, and Australia is more advanced than the rest of the world on this point as the National Transport Commission made the proposition to the federal government that was accepted last year, and that will kick in 2026. But there are some jurisdictions from which remains extremely complicated to get your vehicle certified amalgamated or to get a permit to operate on public roads. (Service provider)

The main regulatory challenges as presented in Figure 5 are:

• – Frameworks not adapted to current technology

• – Getting trial permits non-compatible due to current legislation

• – Shuttles’ low-speed allowance prevents operation in public

• – Risk of ADRT becomes very income/status orientated

• – Incentivize innovation without compromising public safety

• – Caution in short-run makes technology deployment take longer

Figure 5. Technical challenges in deployment/adoption of ADRT.

3.3.3. Financial challenges

The financial challenges were categorised into 2 subcategories: price value, and business model evolution as highlighted in Figure 5.

3.3.3.1. The price value

The uncertainty amongst experts regarding the feasibility of ADRT services concerning the cost–benefit of their deployment is evident in some reported opinions

if you are using this DRT to fill the transport need that has not been met, then I think if it is somewhat fit for purpose, it’s going to be held. If you are trying to make it a commercially successful service in an already crowded transit space, for example in Sydney, that’s going to be much more challenging because the economics of it have to stack up, and I think the last few years have shown that the economics of anything autonomous is much harder to make the case for than people originally. (Academia)

According to Perera, Ho, and Hensher (2020), for ADRT services driven by public policy, having an operator who is investing in the success of the scheme is essential. The discontinuity of most prior DRT trials has often been attributed to a lack of operator enthusiasm. Conversely, commercially led DRT services may face opposition from local authorities. Hence, the success of ADRT programmes often depends on a productive partnership between multiple stakeholders. Overall, the following are the main challenges regarding the price value mentioned by the interviewees:

• – Making commercially successful service in an already crowded transit space is challenging as economics must stack up

• – Economics of anything autonomous is much harder to make case for initially

• – Shortcomings & high cost of components, & supervisor

• – Cost per passenger too high compared to traditional transit

• – Cost of subsidy per passenger too hard to justify

• – Removing driver cost-saving but needs to have a human operator/ steward

3.3.3.2. Business model evolution

The same issue translates into concerns that might occur for the evolution of a sustainable business model that can deliver profits to the service provider and benefits to the users and comply with government and community requirements (Perveen, Kamruzzaman, and Yigitcanlar 2017). It is challenging to find a profitable or economical business model, as another state department recognised,

Are people going to pay more for an autonomous vehicle if it’s fundamentally no better than a human? That’s the other big barrier, and maybe it won’t make a profit, but maybe councils and governments will choose to subsidize because it gives more equity of access to them. It’s a variance. (Peak body)

The Dial-a-Bus trial in Milton Keynes, UK, ultimately failed due to fares that were set too low. After being scaled down to only operate during off-peak hours, the cost of running the programme was higher than initially budgeted. Finally, there was a lack of political commitment to maintain the programme, leading to its termination (Perera, Ho, and Hensher 2020). The main challenges extracted regarding the business model evolution are:

• – Develop a sustainable economic business model

• – Lots of providers want to operate in SEQ because of commercial value

• – Are people going to pay more for ADRT if not better than humans?

• – Vehicle/fleet size, high costs of resources to service frequency

• – Technology companies do not provide services to socioeconomically disadvantaged people who need them

• – No large-scale trials to know the cost of having a completely autonomous system in the network and its feasibility

Hence, service providers must develop a viable business model that impacts the regulatory and operational framework and infrastructure readiness.

3.3.4. Behavioural challenges

User adoption of ADRTs was stated in form of behavioural challenges during the interview. As claimed by one participant from academia ‘it is about knowing the markets, as different people have different requirements’ (Academia). All respondents considered how generic ADRT services are and whether they can cater to personalisation and expressed that tailoring the experience would be challenging. Accordingly, the success/failure of the services may be a consequence of (see Figure 6):

1. Service user expectations: How do ADRT services meet the customer expectations around the level of service quality, ease of use, safety, and privacy?

2. Road user expectations: How do other road users interact with ADRT services and what are their expectations for safety and equity?

Figure 6. Behavioural challenges in deployment/adoption of ADRT.

3.3.4.1. Service user expectations

• Service quality

All interviewees highlighted the need to provide good service quality. They stressed that passengers are very time-sensitive and time conscious. Ride quality, acceleration, deceleration, number of turns, and vehicle stops will be crucial to whether the technology is accepted. All participants referred to various aspects of service quality in one form or another as below (see Table 3):

• – Reliability: Many respondents pointed out that it is important if potential users can rely on being able to book a trip and the timeframes involved, which affects how users would need to manage their time around travel. A recurring view was about the extra cost associated with extra travel time associated with deviations to collect multiple passengers during a revenue trip, which is the unplanned time from the users’ perspective. Reliable pick-up and travel time estimations would provide a level of trust in the system.

• – Accessibility: Coverage is important concerning the proximity of vehicle pick-up and set-down locations to potential users’ origins and destinations. Near door-to-door convenience is desirable. Another point concerning accessibility that has been addressed in the literature and was mentioned in the interviews was how the user will get ADRT service. For example, they may book service via an app, and it may arrive at a certain time to enhance the ease of use in that entire value chain from waking up in the morning, then to deciding to go somewhere. The user will need to learn how to book service, reach the boarding location, wait for the required time, board the service, travel on the service, alight the service, and then reach their final destination. Many respondents discussed that accessibility is not just in terms of the convenience of location; emphasis is placed on the need for catering for the various mobility needs of our diverse community. According to an interviewee, in terms of accessibility needs one in five Queenslanders has a disability, so the business providing those shared services must meet regulations that ensure the protection of those users.

• – Speed: Some respondents expressed that the biggest impediment to shuttles so far has been the speed with which they can travel within the traffic if they are on a dedicated lane. There will be some infrastructure in addition to running the service. However, this additional infrastructure cost could deliver some benefits in times for people to use it.

• – Comfort: Some experts underlined the importance of considering the demographics in providing a comfortable service.

• – Status: Several respondents stated that there is a certain latency in the sense that people might not be so keen to take public transport, but they do not have any choice, particularly when the cost is concerned. But there is a latent demand if they had even the slightest chance to take a luxury bus or a luxury car that can carry five or six people. Maybe that is not expressed yet. So, there is another potential demographic group. The London Tube was discussed by analogy, where very wealthy individuals in high-status roles are using that public transport facility regularly because the surface network is congested.

• • Ease of Use

Table 3. The frequency of items mentioned by participants in each category (%).

Category	Subcategory		Participant
		1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24	25	26	Total No.	(%)
Service quality	Reliability	●	●	●	●		●		●		●	●	●				●				●	●		●	●			14	54
	Accessibility		●						●	●								●	●			●	●	●		●	●	10	38
	Speed	●			●	●					●									●		●			●	●		8	31
	Comfort				●	●	●				●	●	●										●				●	8	31
	Status	●					●		●					●														4	15
Ease of Use	Booking/travel		●								●								●				●			●		5	19
Perceived risks	Safety	●	●	●	●	●		●	●	●	●	●	●		●		●	●	●	●	●	●	●	●	●	●	●	23	85
	Privacy											●				●		●										3	11.5

Many respondents discussed ease of use, predominantly around the elderly. One interviewee, though, expressed a contrary view to the aforementioned ones (E2).

• Safety

The uncertainty regarding the perceived risks amongst experts is evident in their reported opinions. All of them discussed safety and privacy. Across all stakeholder groups, there was agreement that it is crucial to ensure that issues are attended to for safety reasons and reliability of the technology that will impact the quality of the service, which would be the trade-offs between traffic flow improvements and route identification versus the actual smoothness of the ride.

• Privacy

Privacy concerns regarding shared space as well as personal data were challenging. Several interviewees noted that some people preferred very specialised, individualised services. In the more individualised service, just a person or his/her family/immediate friends would be able to book and then they could use the service exclusively. Albeit that would be a challenge to government providers because they base their whole transport system and mass transit on the general public as a whole. Very few interviewees were worried about the privacy of the data. Figure 7 provides a visual representation of how the participants rated the the service user expectations in each category.

Figure 7. The frequency of main categories and subcategories.

3.3.4.2. Road user expectations

• Safety

Part of the safety issue translates into concerns that might occur from the sensitivity of other road users. During the interviews, the participants indicated that even though vehicles are safe, if they move in ways that are not familiar to humans, it may cause errors by human drivers of other vehicles.

• Equity

People have a very strong sense of equity and fairness and so if these vehicles are seen to be exploiting road rules or doing something that is not considered to be fair, that could cause resentment.

For instance, one interviewee mentioned

Young people probably tend to make take more risks and they know that the vehicle is meant to yield to a human which those interventions or interactions can be if you are in a less controlled environment when you have got younger people, cyclists, and drivers. The automatic vehicle will always currently operationally yield which will have a detrimental impact on the customer experience in the vehicle that stops every few meters, so I think you’d need to have a designated operational design domain that limits deployment opportunities. I think when you try to allocate it a 24-hour bus lane or something like that, make sure that there’s not a perception that you are taking away from the vehicle space. (State transport authority).

People will have high expectations from the system and if providers do not address these matters properly, even if they are not users, it will likely drive negative community reaction. As acknowledged by alive Perera, Ho, and Hensher (2020), The lack of marketing was a significant factor in the lacklustre patronage of the Shellharbour, NSW trials. Previously, there was hesitation among customers to embrace ridesharing, but with the rise of sharing economies like Uber Pool, this attitude is rapidly changing. To ensure the success of the service, it is vital to implement strong marketing strategies and communicate with customers to manage their expectations.

3.3.4. Deployment/Adoption mitigation strategies

All interviewees mentioned the significance of the partnership between transport authorities, government, industry and academia in one form or another to help increase societal acceptance of autonomous shuttles in preparation for ADRT. As can be seen in Figure 8, strategies provided by the experts can be classified into the following categories:

• Mindset change

Figure 8. Proposed mitigation strategies for deployment/adoption of ADRTs.

Several respondents highlighted the importance of getting people to shift away from an ownership mindset to a shared-use mindset. This is not dependent on technology; rather, on changing human behaviour, which is a much more difficult proposition. As indicated by one comment

if people understand that moving away from privately owned vehicles to shared, autonomous, and electric motor transportation to reduce accidents and emissions is important and urgent, then that drives adoption. Otherwise, it would become just an electronic gadget if they do not change their behaviour. A lot of people are saying we need special lanes for autonomous vehicles, but my argument is we need to just have fewer cars on the road, so the government needs to discourage car ownership and make all the cars extremely expensive just like they do in Singapore. If they make it expensive to own a car and public transport is good, then people will use public transport that needs to be shared, and I think that’s going to be the biggest success factor. (Academia)

• Education and exposure

The majority of respondents acknowledged that the differences between the age groups regarding the uptake of ADRT reduce the levels of awareness and knowledge, which are essential in gaining individuals’ trust. There should be awareness and educative campaigns as more vehicles enter the fleet. The community should be given a tangible experience via book/play with the system and obtain feedback from them. Regulators need to be prepared to understand how this technology is going to be regulated and introduced.

For example, one participant claimed that

in the first instance, getting people to book and be able to play with the system is probably the critical aspect. So, you build that awareness in your pilot trial, which is what the government is currently doing. NSW did some good work on it. you’ve got to have a behavioural change aspect in your policy, and the engineering aspect of building these new things drives people’s behaviours to change. So, like you’ve adopted the mobile phone over the decade, you are planting seeds for future opportunities with your work in these trials. that’s the idea of leading and shaping expectations and communities. Appetite for these things will drive investment decisions and therefore if you’ve got general warmth towards the product, people will be much more responsive to taking that AV journey. (State transport authority)

• Government subsidies

Transit systems throughout the world are mostly not financially viable proposals, so are subsidised by government as a community service. Many experts suggested that government could offer a discount to incentivize the use of traditional mass transit not to make existing investment redundant and encourage multimodal public transit trips rather than using a shuttle for the entire trip. They also mentioned that subsidy is needed at least in the initial phases to strike the right service fee for people. On the contrary, some experts expressed that any formal incentives/subsidies need to be justified just in case of some form of market failure.

One interviewee described that

in a fairly liberal market democracy, the default assumption is the private sector is more efficient and the public sector needs only step in when there is a market failure. So, any formal incentives or subsidies need to be justified just in case of some form of market failure. Hence if ADRT services are commercially viable, then it’s probably best left to the private sector to provide these services. Alternatively, the government might need to step in more selectively where they provide subsidies on particular routes where there is not as much demand to ensure a minimum level of service to all people as a government does more about equity or allocate incentives for suppliers that offer ADRT services in future contracts/tenders. Thus, the government right now should not be thinking about incentives and subsidies so much as just helping the process of development and deployment through integrating ADRT service within the current offering of public transport services and facilitating that shift away from human-driven buses to autonomous systems. (Academia)

• Consumer incentives

According to interviewees consumer incentives concern managing the expectations of the customer. Across all stakeholders, there was agreement that using incentives would not need to be too wide because the economics or usability of the service should sell itself and the government and providers should not need to add extra incentives. Notwithstanding, some forms of pricing and non-pricing incentivization from a commercial perspective could be considered in the future. Some examples of non-pricing incentives were as follows:

• – Offering attractive & faster service, not necessarily disrupting the cost.

• – Incentives can be just provided service for those who did not have any.

• – Use of ADRT covers commute/parking costs of regular patients with no carers in the health precinct.

• – Provide upfront customer service/passive surveillance to make it an attractive offer to overcome concerns.

• • Liability and insurance

It was acknowledged during the interviews that we cannot overcome public acceptance challenges without building trust in the technology and properly integrating the system within the whole transport network. Diffusion into society will be much easier as people understand the benefits and potential risks of the technology through word of mouth. The strategies provided in this regard are as follows:

• – Conducting trials to test/make sure technology works safely to build public confidence.

• – Running autonomous delivery initially rather than carrying people will allow researchers to obtain a lot more data on the economics, reliability, and acceptance of other road users.

• – Build up roadside equipment & visual signals/signs around the vehicle to ensure it is responding to danger.

• – Overcome low insurance with the cooperation of Australian state governments to share insights to set up regulations.

• – Prevent entry of several providers/operators in the market with different vehicles & offering different products.

4. Discussion and implications

This study sought to develop an overall picture of opportunities and challenges of ADRT deployment and user adoption by employing qualitative techniques through the lens of actors involved in decision-making who were engaged with Australia’s industry environment, comprising state transport authorities, consultants, peak bodies, service providers, and academia. The expert interviews generated significant new knowledge, opening up the possibility for further research in this area. The study has revealed the optimistic prospects of ADRT regarding market development and sustainability. These findings hold great importance for expanding research in the context of autonomous shuttle buses where user acceptance will perform a pivotal part in the diffusion of these innovative technologies.

The comprehensive detailing of the implications of findings enables them to identify priority areas within the broad agenda which they can then investigate with the general public. Therefore, providers would be able to take measures to substantially increase the potential users’ behavioural intention and ultimately actual usage. Interviewing the transport stakeholders and asking questions systematically regarding the adoption of ADRT services also engenders knowledge and heightens awareness in the sectors that have been interviewed. The major overarching challenges discussed by the interviewees emerged in four core themes comprising: technical, financial, regulatory, and behavioural challenges. These are all critical areas that are linked together because they depend on each other and need to be focused on before the full-scale launch of ADRTs.

The main policy recommendations mainly contributing to ensuring the success of the ADRT scheme, especially autonomous shuttle buses, were extracted from the literature and the above findings as follows:

• – Intervene throughout the innovation process by stimulating the use of shared transport modes among consumers or manufacturers and discourage private ownership. These transport system outcomes typically align with broader governmental goals, such as carbon reduction strategies highlighted by Sindi and Woodman (2021), which could provide tax increases on fuel, cars, licenses and registration; Or disincentivising private single-occupant car driving to dedicate road space to space-efficient modes or restricting 1–4 occupant vehicles to access/park in e.g. airports or university campuses.

• – Anticipate unfolding risk scenarios in regulatory approaches while providing a robust framework for innovators to operate, rather than lagging (Mordue, Yeung, and Wu 2020). Local governments for instance hold ‘key regulatory powers’ (e.g. managing the right-of-way, articulating policies for use of land) and can plan for AVs (Freemark, Hudson, and Zhao 2019) if they have the goals and resources. Through deliberate and stringent regulation, governments can prompt technological responses, which influence the direction of innovation development (Beerepoot and Beerepoot 2007). As Taylor, Rubin, and Hounshell (2005) describe, the existence and anticipation of regulation, as well as the degree of stringency and certainty, are important drivers of innovation. The regulatory frameworks should not be too restrictive at this point to create a favourable environment for tech, changing the algorithm of road-based, slow-speed, controlled, none on-demand trials, and setting safety regulations, performance standards & economic regulations around market entry/behaviours, and facilitating multi-operator, multi-tech trials in different locations with different use cases.

• – Have a test and validation centre for benchmarking all kinds of autonomous technology, even through partnering with private companies as they may have private roads to deploy autonomous shuttles. Adjust the legal voids to enable the testing and operation of autonomous technology on public roads (Ferreira et al. 2020). The pace at which researchers and manufacturers can advance technology and suitability for practical integration of the ADRT service within current public transport environments, relies on the efficiency of policymakers to keep up (Skeete 2018). Further complicating these regulatory challenges are the often-misaligned interests across different levels of government, private sector, and public stakeholders at local, national and international levels (Stone, Legacy, and Curtis 2018; Freemark, Hudson, and Zhao 2019). These pressures sit against a backdrop of multiple countries competing to become attractive markets for AV industry development by offering innovator-friendly regulatory environments (Lee and Hess 2020 Schepis et al. 2023).

• – Change the public transport subsidisation model to include ADRT where it fits best in the community. According to Currie and Fournier (2020), most ADRT failures are caused by excessive costs. It is essential to consider local factors when introducing a new ADRT service, as previous ADRT experiences in other locations may not be directly applicable (Papanikolaou et al. 2017). According to Davison et al. (2014), many unsuccessful projects result from insufficiently realistic cost estimates and a lack of understanding of the target market. Therefore, practitioners and policymakers need to pay more attention to the funding and commercial prospects of ADRT; e.g. providing subsidies on particular low-demand routes to ensure equity or allocating incentives for suppliers that offer ADRT services in future contracts or tenders, or offering a discount on ADRT fares for people who need point-to-point services on weekends or during work weeks to quit private vehicles to reach an entertainment venue, work, or home; or offering loyalty points to buy coffee along the way in the case of capturing ADRT.

• – Build trust in technology and have a clear, definitive message about this technology or service offering through social media. Many consumer studies have highlighted low trust and safety perceptions relating to AVs, which Cunningham et al. (2019) suggest can be addressed by the government through controlled demonstrations or simulations. Moreover, it can be beneficial to identify early adopter consumer feedback loops that can drive uptake in the positive feedback loop or to identify breakdowns in feedback loops to incentivize reluctant technology adopters. Government is typically the sole actor available to sponsor critical basic research which achieves technological breakthroughs that later permeate the private sector (Salmenkaita and Salo 2002). These investments complement corporate R&D, which tends to focus on applied research and commercial development, often in a closed manner that protects intellectual property. Governments can use policy instruments, such as taxation or grant incentives, to subsidise investments and reduce privacy risks while adding conditions that necessitate certain social benefits outside of capitalist motivations (Taylor, Rubin, and Hounshell 2005; Pinkse, Bohnsack, and Kolk 2014).

• – Have a consolidated proactive policy view to support technology development as ADRT typically involves a greater focus on marketing and the development of strategic partnerships, compared to conventional bus services (Enoch et al. 2006; Perera, Ho, and Hensher 2020). A strong market signal contributes to reputation and expectation effects that increase managerial willingness to invest in areas predicted to receive continuing government support. A coordinated network-wide provision of services such as investments in human capital or technical infrastructure is also necessary to accelerate the pace of innovation (Moon and Bretschneider 1997). As suggested by Enoch et al. (2006), achieving a harmonious balance between providing customers with flexible services and managing technology costs is crucial for an operator. The most effective approach to achieving this balance is by adopting an incremental strategy. This approach allows the operator to gradually implement new services and technology, assess the impact, and make necessary adjustments. By taking this step-by-step approach, the operator can minimise upfront investment and ensure a smoother implementation process. The development and diffusion of AV rely on complimentary transport infrastructure developments such as communications and physical road networks, which are currently provided by the government (Li et al. 2019). They are also facilitating electrification and other business models (e.g. shared use) (Sperling, van der Meer, and Pike 2018). More broadly, companies benefit from government activities that are directed towards a more efficient and effective transfer of knowledge within emerging technological domains, as well as the promotion of common visions which can guide collaborative efforts (Salmenkaita and Salo 2002). Additionally, it can be beneficial to provide ADRT services to meet demand for in new housing developments that do not have many car parking spaces.

• – The trials of automated vehicles have provided substantial valuable insights encompassing technology, safety, project management, road users, occupants, reporting, and infrastructure. However, it is worth noting that these trials are not solely focused on the technology itself. Therefore, governments must capitalise on these lessons to shape future trials. Industry, government, and universities can work together to seize the opportunities that have been uncovered, leading to a brighter future for the emerging trends that will shape the policy agenda (Liu, Nikitas, and Parkinson 2020). Yet very few local governments have started preparations for aVs and many city officials are sceptical about the benefits of aVs (Fraedrich et al. 2019; Freemark, Hudson, and Zhao 2019). Given the variations between the environments in which governments operate and the high degree of anticipated social and market disruption, governance approaches and the application of policy instruments will vary greatly.

Even though the qualitative method allowed for an in-depth discussion with a multitude of stakeholders from different sectors, the findings should be regarded as provisional. Hence, further work is required to generate broader data from which more nuanced recommendations could well be prepared. Moreover, future studies might go beyond the current work’s primary focus on Australian stakeholders to include representatives from a broader variety of international organisations and sectors.

The snowball sampling technique used for the current study may have overlooked some sectors that could be considered for incorporation in future studies, particularly the perspectives of emerging new ride-sharing organisations, which are expected to deepen the study. The repercussions for developing countries are also of great interest since ADRT has the potential to develop and expand the current transport services and provide advantages to larger population groups.

5. Conclusions and research directions

The emergence of ADRT creates a responsibility for transport authorities to give proper attention to public passenger transportation policy and to enhance the public transport efficiency. Based on National Transport Commission (2020),

there are several reasons why governments and trialling organisations become involved in trials. Governments must be clear about the objectives they are trying to achieve through trials in their jurisdiction and evaluate them in light of those objectives. There is an opportunity to place Australia in a better position to be ready for the commercial deployment of automated vehicles through sharing learnings across jurisdictions.

Accordingly, we conducted an explorative qualitative study through interviews with transport experts in the field to gain a more profound insight into the challenges and opportunities of ADRT deployment and user adoption, especially autonomous shuttles in the SEQ metropolitan region. Thematic analysis was used to derive the main themes and provide comprehensive insights regarding the ADRT service concept, potential customer groups in terms of the supply and the demand side, and the potential use cases. Challenges regarding the development and adoption of ADRT challenges in terms of supply (technical, regulatory, and financial barriers) and demand (the behavioural challenges) were explained as depicted in Figure 9.

Figure 9. ADRT deployment/adoption challenge framework.

These are all critical areas that are linked together because they depend on each other and need to be focused on before the full-scale launch of ADRTs. Such novel findings, as well as further validation through wider studies, have the potential to enrich the knowledge of transportation specialists and add to the body of research on ADRT user adoption behaviour. In addition, the four main themes of challenges obtained in this study represent key areas of focus for future research regarding how the introduction of ADRT services can be best aligned with the needs of SEQ residents, especially transport disadvantaged populations. These findings open an avenue for developing a scale for measuring latent constructs underlying SEQ residents’ transport mode choice behaviour and using them alongside present behavioural or psychological theories. It is important to emphasise that solving these issues will require a cooperative effort from sociologists, psychologists, engineers, planners, and transport and urban policymakers.

Disclosure statement

No potential conflict of interest was reported by the author(s).