Automating Logistics Operations: Qualitative Insights from Four European Sites

Abstract

Background: Automated vehicles are increasingly entering logistics operations, driven by factors like controllability, standardization, and reduced risk. However, successful automation requires understanding the diverse perspectives of logistics stakeholders. Method: This paper investigates these perspectives through 28 interviews with representatives from five key stakeholder groups within the European Project AWARD’s four pilot sites. Results: Key findings highlight positive expectations for efficiency, safety, and reliability, but also identify critical prerequisites still to be met: further technological advancements, shifts in logistics roles and working conditions, regulatory improvements, and careful narrative building around technology. Conclusions: A deeper analysis of individual stakeholder groups and pilot site representatives reveals nuanced needs and concerns, emphasizing the importance of considering different perspectives and need for further research involving a wider range of stakeholder groups to fully capture the complexities of this emerging field.

1. Introduction

Recent decades have witnessed the progressive automation of a broad range of industries, and the logistics sector is one that is undergoing radical changes. These pose several specific challenges that need to be addressed for the successful realization of the “logistics 4.0” paradigm, a transformative shift toward automation and optimized logistics processes. This transformation entails meeting individual customer needs at fixed costs, increasing the use of digital technologies and cyber-physical systems [1,2], transitioning to software-oriented processes [3], and leveraging big data, robotic machinery, and real-time communication [4].

Because of the intrinsic diversity and ramification of the logistics industry, the extent of automation varies significantly across different contexts. For instance, while automation in warehouses and ports is becoming widespread and more mature by relying heavily on perfected technologies like automated bar-code reading, sorting and storage, conveyor belt systems, indoor Automated Guided Vehicles, and yard management systems, the automation of operations on open roads and with harsh conditions (to enable, e.g., last-mile delivery) faces unique challenges. Autonomous vehicles (AVs) used in these scenarios must navigate highly dynamic environments including varying weather conditions, complex traffic situations, and diverse road infrastructures [5,6].

As these examples show, the core of this transformative process is often perceived as predominantly technical. However, the need to consider how these changes will affect the people involved has recently been acknowledged [2,7]. In this regard, a growing share of research focuses on factors determining acceptance (or lack thereof) of AVs in different use contexts. These factors range from technological features, such as safety, privacy, and environmental friendliness, to individual and social differences like age, affinity for technology, and cultural background, and to supporting conditions such as infrastructures and regulations [8,9,10,11]. Other studies have addressed specific aspects such as interaction between AVs and other road users [12,13], task distribution and the transfer of control [14,15,16], issues with misunderstanding vehicles’ capabilities and functionalities with consequent risks of over-reliance [17,18]. Yet, as these topics have only recently started to attract research interests, many questions remain open. Furthermore, with regard to logistics operations, human presence manifests in a variety of roles including direct operators, indirect process participants, and other employees, employers, and other relevant actors. Each of these has different interests and needs that must be considered and integrated in automation processes. Therefore, more comprehensive investigations of these many perspectives on technology acceptance and adoption in the logistics sector are needed.

By exploring the viewpoints of different logistics stakeholders, representing a variety of interests and experiences, this paper contributes to understanding the factors that influence the extent to which key technologies for automating logistics processes are accepted. While the primary focus of this study concerns AVs, related topics such as design requirements for human–machine interfaces of FMSs, essential for remotely operating and supervising vehicles, are also investigated to promote a deeper comprehension of the transformative processes. Furthermore, this study provides insights into the impact of harsh weather on automating logistics operations, a crucial aspect that is often overlooked. In detail, our study aimed to answer the following questions:

• How do the factors identified in the “Automated Road Transport Logistics Acceptance Model” (ARTLAM) capture stakeholders’ perception of the system in terms of acceptance? (See Section 4.1, Section 4.4 and Section 4.5)
• What functionalities should a fleet management system (FMS) have to support remote operators with task planning, dealing with harsh weather, and issue resolving? (See Section 4.2 and Section 4.3)
• Which differences can be identified among participants belonging to different stakeholder categories and use cases? (See Section 4.6 and Section 4.7)

To achieve this goal, 28 in-depth interviews were conducted and qualitatively analyzed, adopting a ’thematic’ approach. Participants were recruited among the partners of the European Project AWARD (https://award-h2020.eu/, accessed on 30 October 2024). The project’s goal is to enable connected and automated heavy-duty vehicles to operate, autonomously and remotely controlled, in real-life logistics operations and under all weather conditions. Project solutions include a tailored FMS developed precisely for the project and four different logistics vehicles (a port tug, a tow truck, a mid-sized swap body truck, and a forklift). These vehicles feature three operation modalities: manually driven, remotely operated, and autonomous. Although the vehicles are based on different platforms and are meant to carry out different tasks according to the logistics needs of each hub, they are all endowed with similar technical equipment (i.e., sensors, lidar, cameras, etc.). These technological solutions were selected for and tested in four different logistics pilot sites across Europe. Respectively, these were the port of Rotterdam, the Oslo airport, the premises of a hub and a factory (i.e., hub-to-hub) in Gunskirchen, Austria, and a research facility with ad hoc reproductions of open-world loading environments in Seibersdorf, Austria. Collecting observations related to these different logistics contexts and geographical locations was one of the main recruitment criteria, together with a balanced representation of different categories of stakeholders. The logistics stakeholder taxonomy developed by [19], shown in Figure 1, provides the reference framework for the stakeholder categories, as it offers a comprehensive overview of the actors directly and indirectly involved in logistics operations and automation processes, as well as several others with economic or social interests.

Our main findings highlight crucial challenges to overcome, such as further technological improvements to enable higher flexibility and safe use in harsh weather and on open roads. The need for better legislative and regulatory efforts, as well as the evolution of working conditions in logistics (with the support of employers) also emerged as critical nodes on the path to logistics 4.0. At the same time, we also found positive expectations, as most participants believed that technologies for autonomous and remote logistics operations will increase efficiency and safety and will optimize logistics processes, particularly after an initial phase of adjustment and further improvements. Societal acceptance is also expected to increase over time. Possible improvements for the design of the project’s FMS have also been suggested. Additionally, our analysis by use case and stakeholder group shows how different locations, individuals’ expertise, and involvement do indeed yield different results regarding priorities, concerns, and expectations. While our sample of stakeholders was limited in terms of the groups represented, our findings nevertheless underscore the importance of recognizing the heterogeneity of stakeholders in the logistics industry. To gain a more comprehensive understanding, future research should include a broader spectrum of stakeholders, representing a wider range of interests and perspectives, including policymakers, labor unions, and environmental organizations. Only by carefully considering each of these complementary viewpoints can we gain a comprehensive understanding of automation in logistics.

In the remainder of the paper, we first present related work in Section 2; then, Section 3 describes the research questions and provides a detailed overview of the methodology applied. The results are then presented in Section 4 and discussed in Section 5. Section 6 provides closing remarks and presents limitations, suggestions for improving the design of FMSs, and possible avenues for future work.

2. Related Work

In recent decades, a wide variety of models have been proposed to assess technology acceptance. Among these, the more vastly adopted and tested include the “Technology Acceptance Model” (TAM) [20], the “Advanced Unified Theory of Acceptance and Use of Technology” (UTAUT), and its later extension (UTAUT2) [21,22]. Key factors to predict technology acceptance such as perceived usefulness, ease of use, trust, price value, and social influence have been identified. These findings have been further adapted and fine-tuned to study technology acceptance in the specific context of interaction with vehicles. For instance, the “Car Technology Acceptance Model” (CTAM) [23] extends the UTAUT model by introducing safety and anxiety as key factors, while the “Safety Critical Technology Acceptance Model” (SCTAM) refines the understanding of safety expectancy, perceived usefulness, and ease of use in the context of transportation systems [24]. Seuwou and colleagues then developed a technology acceptance model specifically for AV applications (AVTAM) and found performance, safety, and consumers’ trust for AV manufacturers to be the main drives for acceptance and adoption [25]. Approaching the field of logistics, Kapser and Abdelrahman (2020) applied the UTAUT2 model for last-mile delivery and found pricing to be the most significant indicator in this context. They also further extended the model to include risk perception [26]. Lastly, extending the TAM and UTAUT models, Rosic and colleagues (2021) developed an “Automated Road Transport Logistics Acceptance Model” (ARTLAM) [27]. In building on previous work [28], this model refines and emphasizes the following dimensions, crucial for logistics: “job relevancy” (related to employee training and the emergence of new job positions) and “social dimension” (that is, acceptance by different stakeholder groups). Accordingly, it aims to capture the attitudes of stakeholders belonging to different categories, including those that do not typically have direct contacts with AVs, but are nevertheless affected by them.

Several contributions address theoretical aspects of the human dimensions of logistics 4.0. The study of [9] called for more nuanced consideration factors such as operators’ age and interfaces with new digital technologies. The authors also outlined a research and management agenda that goes beyond the analysis of physical requirements to integrate socio-psychological aspects and suggest more frequent adoption of qualitative approaches to determine case-specific needs. Similarly, another study suggested (and exemplified) the use of interviews to support the improvement of worker conditions in warehouses operations and in relation to interaction with new technologies [29]. In their “conceptual framework of logistics 4.0”, the authors identified the “human factor” as one of three key layers [2]. They emphasized how people involved to varying degrees in logistics processes play a crucial role in determining the acceptance or rejection of new technologies, such as human–machine interfaces, the adoption of which implies a shift in employees’ roles (e.g., from operators to administrators, or requiring retraining).

The authors of [7] investigated the impact of new technologies developed in the context of the “industry 4.0” transition for logistics operators. Building on the work of [30], they provided a useful definition of “operator 4.0”, emphasizing updated skills, access to wearable, robotic, and hyper-connected technologies. In relation to the implementation of “Internet of Things” (IoT) technologies in logistics, a study detected a general non-readiness of companies in dealing with the transition [31]. Furthermore, they envisioned two main strategies that logistics companies may pursue. The first implies moving toward plain “automation”, which would relieve operators of most responsibilities, but would also reduce their agency and level of information as most processes would be controlled by technology. The second option requires implementing IoT as a toolkit, hence leaving decision making in the hands of upskilled operators, reducing menial and repetitive tasks, and redefining the sets of skills needed by employees. These considerations are aligned with work suggesting that human presence will persist in automated processes, albeit requiring a re-thinking of human roles in terms of upskilling and retraining [32,33]. Among other crucial challenges, the study of [34] identified the current lack of legislative changes concerning liability and responsibility, as well as the coexistence of human-driven and AVs, as one of the main hindrances preventing wide adoption of new technologies.

Most existing empirical studies have focused on a specific geographical context, individual stakeholder categories, or on the general public. The authors of [35] conducted a large-scale survey among supply chain practicing experts, aiming to identify the main components of the logistics 4.0 model. They selected nine final items belonging to three main categories, respectively, referring to improving organizational capabilities (e.g., retraining of employees), interconnection and material flow (e.g., customer and provider supply chain connections), and processes automation (e.g., automated warehousing).

A recent study investigated the acceptance of automated logistics among citizens in an urban context and found that participants were mostly in favor of the new automation technologies as they found them efficient and reliable but raised concerns in terms of safety, security, and potential losses of jobs [36]. Furthermore, the authors identified age as a factor influencing acceptance in that younger participants expressed higher ratings. In a similar urban context, a stakeholders’ study on city logistics highlighted traffic and congestion as main obstacles hindering the actuation of passengers and freight transport [37]. The authors also found age to be discriminant, as older participants are more concerned with safety aspects, while younger ones are more concerned with environmental impact. Through interviews, requirements for the successful introduction of AVs in port logistics operations were identified [38]. Among these, the authors pinpointed the need for better legislation concerning AVs, updating structural conditions (e.g., by improving sensor networks also on public roads), and organizational processes (e.g., concerning improved data transparency and operations safety). They also reported that, assuming that the requirements would be met, interviewees were positive about AVs, particularly in relation to driver shortage and in restricted areas such as those of port operations. Another study reported similar findings from a comparable seaport context and emphasized the potential advantages in terms of economics and process efficiency for road transport service providers, as well as planning customer service benefits for supply chain stakeholders [39]. The authors of [40] also tackled port logistics operations from a stakeholders’ perspective and identify digital technologies as a tool to increase operational efficiency and support vessel automation. However, they also highlight the current pace of digitalization as one of the main limitations.

Another avenue of empirical research tackles requirements and enabling factors to realize sustainable logistics. Among these, a study identified technology infrastructure and solutions, cybersecurity, commitment to sustainability from company management, a customer-centered approach, and support from governments and institutions [41]. Particularly in relation to the second to last of these factors, a recent study reported how the involvement of external stakeholders (i.e., beyond the companies’ limits, such as final customers) plays a crucial role in determining the adoption of sustainable logistics [42].

Finally, studies exist that address the requirements and viewpoints of specific stakeholder categories. For instance, the authors of [43] highlight a discrepancy in the perspectives of managers of logistics and of forwarding companies on the use of automated trucks in road freight logistics. Specifically, members of the latter group underlined how the full potential of automated trucks will only be realized by completely replacing drivers, whose scarcity is a timely issue. To ensure this transition, the authors maintain that the challenges to overcome are not only technical but also political and legal. Another recent study qualitatively explored the perspective of remote drivers in relation to AVs and teleoperations and found that the operators tend to want to retain one form of control over the vehicles (i.e., either directly controlling the vehicle, or at least monitoring it constantly) and suggested improving sensing technology (i.e., field vision and motion feedback) as an enabler [44].

By examining data from four distinct pilot sites and incorporating perspectives of participants from five categories of stakeholders, this study contributes to the understanding of the complex dynamics surrounding technology acceptance and adoption in logistics. The grouping, sampling, and recruitment of participants was informed by the logistics stakeholder categorization presented in [19].

3. Materials and Methods

3.1. Study Design

3.1.1. Data Collection

The study design was agreed upon with the other project partners, and contact with the potential participants was established through the project network. This study was then organized into two rounds that took place between July–August 2023 and November–March 2024, conducted and recorded virtually using Microsoft Teams (MS Teams). The whole experimental procedure, which consisted of semi-structured interviews and included a survey with 5-point Likert scale answers, took between 50 and 60 min and was conducted in English. The survey comprised 34 items spread out over eight sections, and was accompanied by a total of 18 non-mandatory open questions distributed throughout the procedure.

3.1.2. Participants

A total of 28 participants were recruited via a combination of purposive and quota sampling among the partners involved in the AWARD project according to the following criteria: (i) belonging to one of the stakeholder category defined in [19]; (ii) familiarity with one of the pilot sites; and (accordingly) (iii) some degree of awareness about or experience with the relevant technologies, that is, having had contact with one of the vehicles, the FMS, or the systems for remote operations used in the project. In total, 25 men and 3 women aged between 21 and 70 years of age participated in this study. Their average age was 39.32 years (M = 36.5; SD = 12.31), and they resided in seven countries (Austria: n = 14; The Netherlands: n = 5; Norway: n = 4; France: n = 2; Belgium: n = 1; Denmark: n = 1; and Finland: n = 1). Participants belonged to 5 stakeholder categories derived from [19]. Respectively, these were ‘process participant’ (n = 3), ‘indirect process participant’ (n = 6), ‘technology developer’ (n = 10), ‘employee of authorities’ (n = 1), and ‘other’ (n = 8).

Participants were further categorized in relation to the four project use cases and pilot sites representing four logistics hubs in Europe, to investigate possible differences among them. Respectively, these were the port of Rotterdam (participants n = 8), the Oslo airport (n = 6), the premises of a hub and a factory in Gunskirchen, Austria (n = 8), and a research facility with ad hoc reproductions of open-world loading environments in Seibersdorf, Austria (n = 6). As mentioned in the introduction, the four use cases also imply the use of four different vehicles and three operation modalities. However, the FMS employed was the same in all the pilot sites.

3.1.3. Procedure

Potential participants were informed about the scope of this study and its relation with the AWARD project. If they agreed to participate, they were asked to sign an informed consent form, and a date for the virtual interview was agreed upon. The interviews were recorded, and transcripts were generated automatically by MS Teams and then polished, edited, and prepared for analysis by the researchers responsible to ensure accuracy and fidelity to the participants’ expressions.

At the beginning of the virtual interview, participants were greeted and thanked, informed that their answers were being recorded, and briefed about the structure and purpose of the study. Then, they were asked to picture themselves as a logistics operator in charge of a fleet of three AVs. In this role, their tasks included the scheduling of loading, transporting, and unloading operations as well as resolving issues. They were then shown three videos portraying logistics operations scenarios, similar to those tested in the pilot studies. The videos were played from the point of view of an FMS developed in the AWARD project as a tool to both monitor and (if needed) remotely control the vehicles. Each logistics operation shown in the videos took place in one of the project use case hubs, as seen from the FMS view (see, respectively, Figure 2a, Figure 2b, and Figure 2c). After each video, participants were asked a question concerning functionalities and features of the FMS. Specifically, the questions concerned what aspects of the FMS could be improved or changed regarding, respectively, tasks planning, weather information, and issue resolving.

After this initial part, participants were guided through a questionnaire composed of eight thematic sections, each investigating one facet of acceptance, as defined by the ARTLAM (respectively, usefulness, ease of use, supporting conditions, safety, security, reliability, public acceptance, stakeholders’ acceptance). At the end of each section except for the last one, which was followed by only one general question asking for further thoughts and opinions they did not express until then, participants were asked 2 open questions to answer orally. These questions focused on participants’ positive expectations and concerns in relation to the topic of the section. The aim of these questions was to capture any additional insight concerning the survey sections that were not covered by the Likert-scale questions. For instance, after they rated all the items in the first section labeled “usefulness”, they where asked if they had positive expectation (first question) or concerns (second question) in relation to the usefulness of the technologies for autonomous and remote logistics operations. Furthermore, throughout the procedure, participants were encouraged to ask questions and express their thoughts, doubts, or expectations in general and in relation to specific items of the survey. Demographic questions, including one aimed at determining the stakeholder category to which the participant identified themselves, closed the procedure.

4. Results

A preliminary quantitative analysis of the survey responses did not show any relationships between different demographic groups (age or gender), stakeholder category, geographic location, or logistics contexts. However, we focused on the responses to open questions to capture different perspectives and identify subtle themes that might explain the context-specific factors influencing individual responses.

To analyze the total collected data of 28 interviews, we employed a thematic analysis methodology. Primarily, the deductive coding strategy (i.e., following the structure of the survey and of the ARTLAM) was combined with an inductive approach for new emerging topics. The analysis was conducted using “ATLAS.ti 23” (https://atlasti.com/, accessed on 25 October 2024), a popular software for qualitative analyses. Data saturation was achieved after 21 interviews. A total of 44 codes were assigned to 1109 quotes and then grouped into 5 main categories. These categories, which resulted from the combination of the deductive and inductive approaches, were “Technology features”, “Fleet management system design”, “External enabling factors”, “Human-machine interaction”, and “Public response”. The list of the codes, with occurrences, and examples are presented, with one figure for each category, in the following pages. Several codes (e.g., the time frame, the weather, location dependencies) crosscut through different categories and are therefore reported where relevant (see Figure 3). Below, we report the main findings for each category. Following this, an additional analysis arranged by use case and stakeholder group is also provided.

4.1. Technology Features

This category concerns every technological aspect and quality (or lack thereof) playing a role in determining users’ acceptance, with the exception of topics specifically related to the FMS (see the next subsection). The most important sub-categories that contribute to this theme are illustrated below. Figure 4 shows the pertinent list of codes with frequency and example quotes.

4.1.1. Usefulness and Efficiency

Generally, participants were rather positive in relation to the usefulness and efficiency of technologies for remote and autonomous logistics operations. Positive expectations referred primarily to the potential increase in working hours, better operation and maintenance planning, improved overview and awareness of ongoing operations, optimized use of the fleets, decision making, job conditions, and overall efficiency. Improving the connectivity (e.g., WiFi and 5G networks) and robustness of the hardware mounted on the vehicles was seen as a key enabler, particularly in remote locations and harsh weather conditions.

On the other hand, participants expressed some concerns regarding the true feasibility of 24/7 operations particularly in relation to remotely operated vehicles (and vehicles’ reliability in general), which would still require human presence (as opposed to fully autonomous ones). Similar doubts were raised about technological flexibility under unforeseen circumstances, in relation to harsh weather, and in mixed-traffic situations. The need for constantly available backup solutions, such as the possibility to easily switch to manually operated vehicles in case of failures was also mentioned as a logistics-specific limitation deriving from the need to operate at high pace and capacity. Most of these doubts explicitly referred to a short-term time perspective, meaning that participants expected operation improvement and technological progress to go hand in hand.

4.1.2. Safety

In reflecting on the perception that humans are responsible for most errors, one of the main safety upsides expected by participants concerns the reduced number of accidents and injuries due to limited human presence (for both remote and automated operations). Additionally, the amount of data provided by the large (and likely increasing in the future) number of sensors and cameras was perceived to increase safety (for both remote and automated operations). At the same time, safety is dependent on the technological maturity, but many participant did not think relevant technologies would be deployed before they have been extensively tested and are proven mature (and, hence, safe). To this regard, several participants stated that, having safety set as a priority, AVs might even be “too safe” to be efficient (i.e., stopping for every minor inconvenience). Also, safety was seen dependent on the context of use with several positive comments referring to deployment on private grounds and negative concerns expressed mostly in relation to use in mixed traffic, due to the presence of pedestrians and unpredictable human drivers, and in harsh weather conditions, which might affect vehicles’ sensory perception and communication. Here, and generally with unforeseen situations, human drivers are still seen as more capable and, importantly, flexible. Some participants also stated that mixing automated and manually operated technologies, particularly in the near future, may pose additional safety issues.

4.1.3. Security

Participants’ perspective on security and safety partially overlapped. For instance, many reported that before new technologies enter the market, a lot of money is and will be invested so that systems for remote/automated logistics operations will be secure against external cyber-attacks. This was tied to the hectic nature of logistic operations, which already require high levels of security. At the same time, several participants also expressed the concern that, being set up by humans, these systems will still have vulnerabilities and that, if really motivated (by, e.g., aversion to automation technologies or money), adversary actors may be able to hack such systems. Several participants agreed that such adversary actions may have catastrophic consequences (e.g., forcing long-term operation stops). Concerning illegal smuggling of goods and people, most participants did not see an increase in risk. Rather, connectivity functionalities of the vehicles should provide further security. However, the same technological complexity needed to ensure efficient communication and connection was also reported as the potential cause of data leaks or communication interceptions.

4.1.4. Reliability

Participants’ attitude toward the reliability of the technologies employed is generally aligned with their opinions about safety and security. This topic is also strongly tied to that of usefulness and efficiency. Crucially, participants expressed that a reduction in human presence (and errors) by automating (or remotely handling) tasks will result in improved operations and maintenance planning, less work downtime, and, ultimately, increased overall reliability, efficiency, and cost savings. Most concerns were expressed in relation to harsh weather conditions, which might compromise not only the usability of individual parts and hardware (e.g., cameras and sensors) but possibly also of whole systems and their safety. Also, the connection between stations for remote operations and the vehicles (e.g., if a logistics hub is in a remote location) was seen as a limitation for reliability. Ultimately, reliability and usability are tied to (and limited by) technological adaptability to unforeseen situations, hence implying constant human availability and occasional intervention. Eventually, market adoption will depend on the robustness and reliability of the systems. To this regard, some participants envisioned an initial transition phase, before the systems’ full potential can be achieved.

4.1.5. Technological Maturity

This topic crucially emerged in relation to most other technological aspects and to the timing framework. The key ideas expressed by participants cover a broad spectrum. For instance, several participants stated that the technologies will not be deployed until they are mature and thoroughly tested because, otherwise, they would pose safety and security risks. Others highlighted that if the automation of logistics operations would not be achievable, there would not be such significant capital invested in the development. At the same time, others thought that the technology has not, as of now, reached a level of maturity that ensures safe use. Many such comments referred particularly (but not only) to fully autonomous vehicles, their use on open roads and in harsh weather conditions.

4.2. Fleet Management System Design

This category emerged in the analysis as the result of, especially, the first three questions that were asked after participants watched videos of various operations through the FMS interface. As such, it includes thoughts on displaying information concerning scheduling tasks, the weather, issue resolving, participants’ attitude toward the FMS employed in the AWARD project, and desirable design features to have implemented in user interfaces for remote logistics operations in general. Some participants provided positive feedback on the functionalities and interface of the FMS and, generally, on the positive role of controlling logistics operations through an FMS. These are seen as potentially very useful tools to improve planning, supervision, efficiency, and job conditions. Interestingly, some participants saw remote operations through an FMS as a transition phase leading to fully automated operations. Figure 5 shows the codes belonging to this category, their frequency, and example quotes.

4.2.1. Design Improvements

Upon explicit request, participants suggested a number of improvements and generally desirable features for systems for remote and automated operations to have. One key point concerns the amount of information an operator can handle in relation to the number of vehicles composing the fleet. At the same time, other participants expressed the need for the operators to be well informed about a vehicle’s state (e.g., battery status), past and present movements, future tasks, and other operators’ shifts logs. To avoid overwhelming the operators, some suggested that detailed information may be recalled only upon request (particularly by ’expert’ users). The need for camera views from the vehicles and the site was mentioned several times in relation to task scheduling (e.g., better overview and connection), harsh weather (e.g., to see whether the road is blocked by snow or ice), and issue resolving (e.g., to see if roads are closed or monitor accident severity). Concerning other information related to harsh weather, participants suggested displaying air and the tarmac temperature, road friction coefficient, wind speed, precipitation intensity, and forecasts for better planning. Then, in relation to issue resolving, participants suggested improved use of notifications (e.g., multiple colors/icons, more responsive and clear) and integrated systems for communication with in situ operators, authorities, and emergency services. A topic that was reported by several participants, particularly in relation to the need for human intervention, concerned the division of tasks between vehicles, operators working through the FMS, and other actors. Others mentioned the risk of operators losing feeling for the non-virtual reality of the operations, due to visualizing everything through a screen. The latency between giving commands remotely and vehicles’ implementing them also emerged as a hindrance.

4.2.2. Ease of Use

Ease of use (or lack thereof) also emerged repeatedly, as this topic was featured in the structure of the survey. This topic was primarily interpreted as a quality of the FMS and, only partially, in relation to other technological features. Participants were quite divided in this regard. Some said that, as the interface resembles a phone app, it is fairly easy to use and understand and that no specific IT knowledge should be required. Operators’ training was mentioned several times as a key enabler. Also, future iterations of the system are expected to be progressively easier to use. On the other hand, several participants found the system rather complex and expressed concerns in terms of stakeholder and user acceptance.

4.3. Human–Machine Interaction

This category, which does not strictly reflect the structure of the survey, summarizes the important and multi-faceted topic of human interaction with the technologies involved in the automation of logistics operations. The codes grouped into his category, their frequency, and example quotes are presented in Figure 6.

4.3.1. Human Intervention

Among the main topics that fall under this category, the most important concerns human intervention. To this regard, participants identified optimal communication and interface, as well as robust software and hardware, as crucial to provide the operators with a better situation overview and awareness, plan operations safely, or determine interventions in the presence of harsh weather or issues with the vehicles. Failing to provide these features would translate to increased human on-site presence. In general, the idea of intervention itself is crucially tied to responsibility and task distribution as well as human presence, even if just in the background, for potential takeovers (i.e., as a backup option). In turn, the degree of human intervention is inversely proportional to technological maturity, flexibility (to unforeseen events), and reliability, all of which are expected to improve over time.

4.3.2. Human–Machine Comparison

Another relevant topic often brought up concerns comparisons and compatibility between humans and machines. We already illustrated how the former was often represented in terms of higher human flexibility and adaptability, which is a crucial aspect in harsh weather conditions, but also when other humans are involved (e.g., on open roads). But comparisons also emerged on other levels. A significant share of comments referred to humans as the cause of most small mistakes and accidents, which will diminish by employing AVs. To this regard, many participants reported how automated vehicles have quicker reaction times, do not drink, and get sleepy or sick, which in turn improves safety and efficiency. At the same time (as we will show more in detail later on), this aspect contributes to the perception that people might lose their jobs.

4.3.3. Human–Machine Compatibility

Regarding compatibility, the main concern referred to safety in relation to the shared use of roads and city spaces, as well as coexistence in the form of mixed fleets. This was often seen as a result of human unpredictability and careless behaviors (both when driving and as pedestrians). Some participants even mentioned risks associated with people testing the vehicles’ safety out of curiosity or high expectations. Also here, time played a role as it not only allows for technological improvements, but also for people to become used to AVs and learn how to interface with them. Another issue that was reported concerns how operators may become bored and even detach from real-life events, with potentially negative consequences. However, participants also brought up the idea that, once the systems are up and running, logistics operations may benefit by combining strengths from both sides (i.e., humans and automation), particularly at the intersection between remote and fully autonomous operations. Importantly (as the next section will show), ensuring the retraining and upskilling of employees was seen by most as a crucial enabler. In turn, failing to achieve this may result in accidents, losses of jobs, and lack of acceptance.

4.4. External Enabling Factors

In general, supporting conditions in the form of employees training (strongly linked to the topic of public response and acceptance), connectivity and communication, physical infrastructure, and regulations are seen as key requisites and challenges for the deployment of automated technologies, as Figure 7 exemplifies.

Several participants highlighted that regulations and legislation are, particularly in Europe, not keeping the pace of technological development and expressed a sense of urgency. This topic often reflected the perception of the broader European issue of properly regulating autonomous technologies resulting from high-safety standard legal requirements (https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:32019R2144 and https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32022R1426, accessed on 30 October 2024). Other participants were uncertain, either because they thought support will be different from location to location, or because they felt that the strong drive of technological development will force legislation processes to accelerate. Adequate physical infrastructures were also reported as crucial, to optimize and facilitate operators’ job (e.g., with distributed cameras) and for integrating vehicles into open roads (e.g., intelligent traffic lights). As these were seen as costly investments, a critical issue here concerned the cost distribution among the interested parties. Some participants suggested that private investments will support improvements on companies’ sites, but that improvements on public roads will be more difficult. Economic interest was reported as a key enabler in relation to safety, usefulness, and reliability. High costs in the short term are seen as a temporary hindrance but, at the same time, long-term returns due to increased efficiency represent one of the main drives. In this regard, some participants also expect a snowball effect with growing interest and investments, on top of significant initial capitals invested.

4.5. Public Response and Acceptance

Here, we report on topics that most directly relate to the general public’s and stakeholders’ acceptance of the relevant technologies. As exemplified by Figure 8, important facets refer to the job market evolution, narratives created around the technologies, and their environmental impact. Concerning acceptance in general, participants expressed mixed feelings, most of which concerned AVs. Many envisioned a transition phase that, as the technologies prove safe and efficient, will eventually lead to well-calibrated acceptance. Some expected either resistance or too high expectations as the result of individual attitudes, while others thought acceptance varies from location to location, with more cosmopolitan and technological contexts at the forefront of adoption. Only very few participants did not see any issues with acceptance.

The idea (fear for some) that many jobs will be rendered obsolete by the automation and optimization of logistics processes plays a central role, as reported by many participants. However, it was also noted how the automation process might actually be beneficial, due, for instance, to the current scarcity of human drivers. Additionally, participants referred to the evolution of jobs in logistics in terms of increasing monotony and losing feeling for what happens on site, but also of increased safety (e.g., less contact with the vehicles and exposure to the weather), comfort (e.g., working from home), and efficiency (e.g., one worker controlling multiple operations). Also, while the fear of job loss is still dominant, several participants expected new (and overall better) jobs to be created. The responsibility to provide more and better resources for employees’ training and to increase their awareness and understanding of the new technologies and their impact was mostly expected to be on employers. This aspect is also tied to the broader concept of narratives created around technology and pushed by companies and, importantly, the media. In this regard, various participants criticized how the hype and sensationalism around AVs in the news selectively magnify the perception of danger. Interestingly, albeit most comments were positive, not many participants considered environment-related aspects crucial for acceptance, primarily because automating logistics is not perceived as having any particular environmental impact (as opposed, for instance, to electrification).

4.6. Analysis by Use Cases

Here, we briefly report the most relevant differences that emerged between the four project use cases. These are the Oslo airport (n = 6), a realistic reproduction of warehouse environments in Seibersdorf, Austria (n = 6), the port of Rotterdam (n = 8), and a logistics hub and a factory in Gunskirchen, Austria (n = 8). Most considerations on what functionalities of the FMS to improve were very similar across the use cases and are therefore not mentioned below (as they are already reported in Section 4.2.1), unless unique to one of the contexts.

4.6.1. Oslo Airport

Generally, airport participants referred to AVs more than to other technologies. One recurrent topic among the airport interviewees was the impact of harsh weather on the vehicles’ capability to operate efficiently. In turn, this calls for safe, manual backup solutions, as airports experience huge volumes of traffic and operations cannot be delayed. Consequently, while airport stakeholders saw efficiency and safety as the main potential benefits of automated logistics operations, they also expressed doubts concerning the current levels of technological readiness (i.e., vehicles stopping for minor inconveniences, issues deriving from proximity to planes) with regard to the airport’s high-pace operations. Hence, human supervision and, occasionally, intervention was seen as a crucial requirement, at least for the near future.

4.6.2. Warehouse Forklift

Also among the warehouse participants, AVs were the most discussed technology. One issue that emerged a few times concerns the higher resonance accidents involving AVs have in the media (as opposed to proper functioning). However, participants also believed that people will become used to AVs, which will gradually mitigate the initial negative perception. This shall be achieved also by clear communication around AVs. Logistics operations were expected to overall improve thanks to automation, but primarily and, at least initially, on private grounds (as opposed to open roads). This was also reflected in the opinion that private entities will spearhead the adoption of automated technologies. Finally, another topic that emerged repeatedly concerns the idea that controlling the vehicles remotely might lead to laziness and operators’ detachment from real-life events.

4.6.3. Rotterdam Port

The most prominent topic that emerged from the interviews with the port stakeholders was efficiency, with mostly positive expectations but also some concerns regarding the flexibility and adaptability of current technologies, particularly regarding environments like the docks, which were originally designed for manual operations. Participants thought that, initially, people will be skeptical toward AVs. Consequently, companies should act promptly to mitigate the fear of job losses and provide suitable infrastructures, while institutions should improve regulations to keep up with the pace of technological development. Few participants expressed the idea that, albeit increasing safety, remotely supervising and controlling vehicles might be a boring job.

4.6.4. Hub-to-Hub

Also, in this case, efficiency was the most relevant topic, particularly in relation to expected improvements in working conditions and workflow optimization. However, these benefits were expected to materialize mostly on private grounds, as the technology was not seen as mature enough to operate on public roads. Likewise, vehicles were not believed to currently be able to operate consistently in harsh, rapidly changing weather. Manual backup (either as remote control or physical intervention) was seen as a necessity for the near future, which, in turn, will require employee retraining. Overseeing the vehicles remotely was seen as a more desirable job than driving and as the ideal “bridge” leading to full automation. Due to the comparatively remote location of this site, participants noted how mobile networks may pose connection and lagging problems, which, in turn, might make it even more difficult to operate the vehicles on open roads.

4.7. Analysis by Stakeholder Groups

The next paragraphs report the key differences among the five stakeholder categories that participants belonged to. However, of the five categories, only four were accounted for individually in the coding, as only one participant belonged to the “Employee of authorities/regulatory agency/certification agency” category. This material was therefore coded with the category “Other” (n = 10), which included, for instance, members of research organizations, of companies’ management or marketing teams. The other categories used in the coding were “Direct process participant” (n = 3), “Indirect process participant” (n = 5), and “Technology developer/equipment manufacturer” (n = 10).

4.7.1. Direct Process Participant

This category comprises personnel operating in logistics yards, fleet managers, and operators working in proximity of the vehicles. The main topic that emerged concerns contextual conditions, as the compatibility of existing infrastructures (e.g., uneven dock pavements and airport pathways) and technologies (e.g., container ships, manual vehicles, and planes) with AVs was questioned. This is also reflected in some participants’ rather negative perception of mixing manual and automated operations. Due to the high pace and volume of logistics operations, the availability of manual backup solutions was seen as a requisite. The time frame is also crucial, as legislation, technologies, logistics processes, and infrastructures were not seen as mature enough yet to support the transition to fully automated operations. However, these were expected to change in the (mid-term) future.

4.7.2. Indirect Process Participants

This category comprises operation monitoring personnel, employees of logistics service providers, and other companies involved in logistics operations. Several participants expressed optimism concerning operation efficiency, safety, and accuracy thanks to reduced human presence and errors. The option to remotely control multiple vehicles was seen as one of the main benefits (with employee training as a requisite), particularly in the transition toward full automation. However, doubts were raised about the monotony of working as a remote operator. AVs’ limited adaptability, particularly in harsh weather and mixed traffic, was perceived as a limitation. Public response was another key theme in this category, with fear of losing jobs being the main concern (followed by fear of change). At the same time, participants also expected acceptance to grow over time, as technology improves and people become used to AVs and experience the benefits. Communication and openness about the new technologies were seen as enablers.

4.7.3. Technology Developer/Original Equipment Manufacturer

Participants in this group are primarily engineers, developers, and employees of manufacturers and OEMs. They belong to the broader category of “general stakeholders”, as described in [19]. AVs and their efficiency were the most relevant topics among these participants. Some expected increases in safety and efficiency, but not in mixed traffic or harsh weather, at least for the immediate future. Participants also noted how the core technologies (e.g., the sensors, physical and digital infrastructures) still need to improve to ensure AVs’ reliability and adaptability. Until then, constant human oversight (and interventions) was considered by several participants as a requisite. In this regard, remote operations were seen as problematic by several participants for different reasons (e.g., too much or poor information from sensors, monotony, cyber-security). Time was also a crucial factor for acceptance (i.e., the initial fear of losing jobs or suspicion toward AI and automation) and interest from investors. Finally, participants from this group also gave several suggestions in terms of FMS design improvements. These included pop-up notifications with differentiated icons, real-time camera views, better task planning information for all vehicles (e.g., a task priority list, vehicles’ status), and weather details (e.g., road temperature, wind speed).

4.7.4. Other

This category mostly comprises members of research institutions, infrastructure, and transport organization and marketing operators. Participants mostly perceived efficiency (the main topic in this group) and optimization as the main benefits derived from reduced human presence and the possibility to manage multiple vehicles. AVs were generally perceived as safer than manual vehicles, but also less adaptable and flexible. Harsh weather and mixed traffic were seen as open challenges calling for manual backup solutions. Participants also provided several potential design improvements for the FMS. These mostly overlap with those mentioned in the previous category, but more layperson-friendly nuances (e.g., showing route options like car navigation systems) were also suggested. Time was also an important factor in relation to technological readiness, adequate regulations, infrastructures, the evolution of jobs, and acceptance. Communication and openness about the technologies were mentioned by some as a way to increase the public’s awareness and acceptance. Remotely controlling the vehicles was seen as an evolution of current jobs and a bridge toward full automation, but not necessarily a safer (because of operators’ potential detachment from events) or more efficient (due to frequent interventions) alternative.

5. Discussion

This paper extends previous work on logistics stakeholders’ acceptance of new technologies by providing an analysis across five different stakeholder categories and four European logistics hubs [38,40,43,44]. In particular, our work contributes by relating these variables, highlighting connections and differences, to changing working conditions, technological maturity and flexibility, and operations in harsh weather and on open roads. Additionally, it provides insights on user-oriented requirements for the design of human–machine interfaces for automated and remotely controlled logistics operations.

In general, participants expressed the idea that automating logistics operations will result in safer, more reliable, and more efficient processes, confirming previous findings [36,38,39]. However, our results also show two crosscutting and intertwined issues that currently hinder the automation process. The first concerns technological maturity and flexibility, as many participants mentioned that the current technological level is not yet sufficient to safely enable fully automated processes. Accordingly, the current feasibility of 24/7 operations and AVs’ overall efficiency and usefulness were questioned. This issue often influenced participants’ attitude toward employing AVs in mixed traffic. This limitation calls for near-constant human supervision (and possible intervention). In this regard, remote operations were repeatedly mentioned as a “transitional” solution until fully automated operations can be achieved. Diverging from previous work on the time factor in determining the acceptance of AVs [45], many participants seemed confident that, over time, enough resources will be invested into the technological development, hence progressively mitigating the current downsides and realizing the envisioned increases in efficiency.

Interestingly, while technological immaturity is also believed to have negative effects on security, this topic also partially raised the opposite problem, as participants mentioned that the more processes automated, the higher the stakes, and the more tempted adversary actors might become. In line with existing work [41], this shows that the security of systems for remote and automated logistics operations shall grow alongside complexity and interests.

The other crosscutting issue concerns automated and remote operations in harsh weather. Although this does not necessarily represent a total impediment, corroborating previous findings [46], many participants expressed doubts in relation to operators overseeing or remotely controlling AVs in such conditions (e.g., heavy rain, fog, snow). For instance, blocked sensors and cameras, as well as critical road conditions, are seen as potential dangers to safety, reliability, and efficiency. As this at least partially represents an effect of technological readiness and flexibility, in the frequently adopted narrative of comparing humans’ and machines’ capabilities, participants consistently pinpointed human intervention as the best solution currently available. However, outside of operations in harsh weather, this comparative narrative emerges also in the opposite direction, as machines are often seen as more reliable than humans (being more accurate and lacking traits such as impatience). This finding is important for two reasons. First, strictly speaking, the topics of human–machine comparisons and compatibility were not part of ARTLAM’s structure, which may represent an additional dimension to technology acceptance (i.e., concerning compatibility and competition) that is typically overlooked. Second, they reflect crucial aspects of individuals’ perception and experience of automation as a social process. Therefore, they should be more carefully considered in future studies.

Both of the abovementioned issues (i.e., technological maturity and the impact of harsh weather) were often related to the time framework of evaluation. This idea that, eventually, technologies will be mature enough to be safe, secure, and reliable to operate (also in harsh weather) recurrently emerges in participants’ view, although most of them struggled to provide clear estimations. Interestingly, in this unspecified transition phase, remote operations were seen by several participants as a viable alternative, and yet others raised doubts concerning how such jobs may become boring.

Corroborating previous findings related to digital tools, IoT, and remote operations [2,7,40], participants were often positive about the potential of FMSs to improve efficiency and working conditions. Specifically, this may be achieved by finding a balance in terms of amount of information provided (not too much or too little), ensuring clear tasks distribution, user friendliness, and efficient communication (e.g., from the vehicle to the operator and with road infrastructure). Furthermore, the process of employee upskilling was identified as a requisite for successfully shifting operation control from sites to offices. These improvements were consistently tied to broader considerations on external support from companies as well as institutions regarding financial and regulatory aspects, in addition to the availability of resources (e.g., to train employees). In line with previous work [31,34,35], several participant indicated how such support does not necessarily move at the same pace of technological development. This consideration underlines how Europe suffers persisting structural imbalances in support around the development and regulation of innovative technologies.

Consistent with previous findings [36], participants expected other stakeholders and the public to positively react to the technologies for automating logistics operations, particularly after an initial adjustment phase. However, they also identified the construction of a fair narrative around the technologies as a requisite. In particular, this topic concerned people’s perception and fear of losing jobs to automation. In line with previous work [2,32,33] and reflecting the ARTLAM’s ’job relevancy’ component, several participants pointed out how the transformation of logistics operators’ roles (from hubs to the office) will enable better overall working conditions and provide a solution to the current scarcity of drivers [43]. Communication efforts should therefore focus on such positive prospects engendered by automation.

The idea of “narratives around technologies” is a common topic whenever new technologies are discussed. In this regard, participants suggested that companies should put more resources into demonstrating to the public that the technologies being employed are safe, efficient, and might benefit society at large. As suggested by other studies [45], the role of the media emerged as crucial, as several participants noted how AVs are disproportionately portrayed in the news as dangerous. Interestingly, not many participants commented on the environmental impact of these technologies. This primarily reflects participants’ perception of automation per se not having as positive of an environmental impact as, for instance, electrification.

While several of the key themes reported above emerge consistently from the site- and stakeholder group-based analyses, there are also some noteworthy differences. Regarding the pilot sites, the requirements, peculiarities, and challenges of each environment show up in participants’ concerns and expectations. For instance, participants from the port and airport, hubs characterized by very busy and tight schedules, raised more doubts about technological readiness, flexibility, and efficiency. Likewise, participants from the airport and hub-to-hub sites, which experience very harsh weather conditions, were generally more concerned about the vehicles’ (and annexed technological infrastructure) capability to function reliably in bad weather.

A similar trend also emerges from the analysis of stakeholder groups. For instance, participants with more direct experience with the vehicle and hub conditions expressed concerns about the operational compatibility of existing infrastructures and technologies with automation. Comparably, participants with a technical background more frequently recognized the current limitations of core technologies such as sensors and interfaces. While our analysis was constrained by a limited range of stakeholder categories and representatives of four European pilot sites, the emergence of distinct issues, interpretations, and expectations is extremely relevant and timely, and it underscores the potential for even greater variation across a wider and more representative sample.

6. Conclusions and Limitations

This study investigated stakeholder perspectives on autonomous and remote logistics operations within the European AWARD project. It contributes to research on technology acceptance and human–machine interfaces in automated logistics, providing insights into factors influencing the acceptance of automation technologies among diverse stakeholder groups and pilot sites.

In general, participants expressed positive opinions about automating logistics operations as a way to increase efficiency, usefulness, safety, and reliability. However, they also identified several requirements and challenges. One factor that emerged repeatedly concerns the current level of technological maturity and the time frame of the transition towards automation.

The dominant feeling was that logistics operations can already be automated, but are mostly limited to enclosed hubs and stable conditions. Operating safely and consistently in harsh weather conditions and on public roads is still seen as partially out of reach. Importantly, however, several participants pointed out how remotely controlling the vehicles and supervising operations may be a suitable compromise during this transition phase.

These considerations call for more committed support from employers regarding employee retraining. Additionally, to promote the emergence of a new image of the logistics operator, a shift in public discourse is needed, focusing on fairer narratives around technology that prioritize new employment perspectives over the fear of being replaced by automation. To complement these efforts, European legislative bodies should act promptly to ensure that regulatory endeavors keep pace with technological developments.

One limitation of this study is the small sample size, particularly for certain stakeholder categories like “direct process participant” (n = 3). Additionally, some categories, such as “cities”, “traffic and road operators”, and “national/local authorities” were not represented at all. This underlines the need for further research with larger and more diverse samples to capture a wider range of views and experiences.

A broader consideration must be made regarding the stakeholder taxonomy adopted in this study. The topics of changing working conditions and the fear of being replaced by automation are central in ARTLAM and have been discussed consistently by participants. However, in this regard, one key perspective that is missing from the study sample, and more importantly, from the categorization, is that of union workers and representatives. This shortcoming should be addressed by revising the current taxonomy. Furthermore, this study used ARTLAM as a guideline to assess technology acceptance in logistics. Future work should extend this and evaluate the model’s validity.

Another limitation regarding the participants concerns the gender imbalance of the sample. The gender imbalance was primarily due to the limited availability of female stakeholders among the AWARD project partners. Besides contextual considerations limited to our study, this aspect might highlight a persisting gender imbalance in logistics and supply chain representatives [47]. Further work may focus on what requirements must be provided, for instance, by institutions, governments, and private companies to close this gap.