A Longitudinal Driving Experience Study with a Novel and Retrofit Intelligent Speed Assistant System

Abstract

Speeding is the primary cause of traffic accidents. To improve road safety, the European Union started implementing a new regulation mandating that all new vehicles coming to the EU market must be equipped with an Intelligent Speed Assistance (ISA) system from 2022 onwards. However, the rule did not include existing vehicles on the roads. Our research aims to fulfill this gap by investigating user experiences and acceptance with a retrofit system developed by V-tron, a Dutch company. Seven participants signed up for our study and our technicians installed the ISA system on their cars. They then used the car for more than one month and reported their experiences weekly. We also recruited a driving school and conducted a focus group with five instructors. Using interview and questionnaire methods to collect their first-person experiences, we saw that all participants acknowledged the vision and potential of ISA systems in reducing speeding and improving traffic safety. While the retrofit system is easy to use, the technology needs to be improved in accuracy and robustness. The overruling mechanism also needs to minimize the latency and consider secondary users unfamiliar with the speed control. Three design concepts were proposed to improve user experiences and eventually promote the adaptation of ISA systems.

1 Introduction

Driver assistance systems have been proven beneficial in several situations, both from a driver's and road environment perspectives. One is the intelligent speed assistants (ISA) system, which has been developed for over two decades. It has constantly been evolving, testing, and improving. A system must communicate with a driver properly to inform without causing unnecessary distraction or interference (Bakker & Niemantsverdriet 2016). For this reason, we see more and more systems being tested and introduced into legal road automobiles (Nidamanuri et al. 2021). Besides the need for safety checking (Dikmen & Burns 2016; Endsley 2017; Ingle & Phute 2016), achieving a high level of user acceptance is essential. This is essential for social support and the actual use of these systems.

When looking at the user response to advanced driver systems like adaptive cruise control (ACC), we see that acceptance and positive attitudes are significantly above 90% (Strand et al. 2016; Xiao & Gao 2010). While this shows potential for systems that aim at improving road safety by assisting in maintaining and regulating speed, there is still quite a significant difference between ACC and ISA systems that provide and control a vehicle’s speed according to the situational speed limits.

While research on roads with dynamic speeds is limited, there is a growing interest in improving driver and other road users' safety through adjustable speeds. One great example is the study performed by Li et al. (2019). Their study used a car simulator to put drivers in a situation where poor visibility hindered their ability to drive safely at a normal road speed. Using a connected vehicle attached to a variable speed limit system, they explored the driver's acceptance and its influencing factors. Within their study, they confirmed that external factors could provide enough support to convince drivers of the importance of dynamic speed limits. They found support for on-board communication and on-road sign displays, with a slight preference for the latter.

Research on these signs generally does involve user acceptance but is limited to the willingness of drivers to adhere to the different speeds. For instance, Janssen et al. (2021) claim that asking for user feedback on dynamic speed sections would not provide actual results, as experiencing and imagining are sometimes significantly different. Furthermore, they address that technology is often only partially used as the inventor/designer intended (Bainbridge 1983). However, their conclusion is promising, indicating that while the willingness to adhere to lowered speed limits is low, it rises significantly when drivers feel the need for the reduced speed is meaningful.

Our study aims to investigate the user experience of the retrofit ISA system (see Fig. 1) developed by V-tron, a Dutch company. The system currently equips two technical components to detect the in-context speed limit, including a front camera for recognizing the road signage and the GPS location. In the near future, V-tron’s ISA system can also retrieve dynamic speed limits from the city’s traffic management system and adjust the driver’s speed accordingly.

We want to understand (1) the effectiveness and usefulness users can perceive, (2) the usability of the system, (3) user experiences and social acceptability, and (4) the possible changes in their driving behaviors. Those insights can help improve the ISA system's designs and bring the products to the market to improve safety and reduce traffic accidents.

We used interviews and questionnaires to understand users' expectations, experiences, and behaviors. We conducted a pre-interview before they started using the system to collect their basic information and understand their driving preferences and experiences, commuting routines, expected expectations related to the system, etc. During the trial period, we sent periodical questionnaires to participants and asked them to report their experiences every week. After the testing phase, we conducted a post-interview to probe their overall experiences and discuss some issues we observed from analyzing their driving data and questionnaire responses.

In the following sections, we will first briefly review theories and measurement methods in related topics. Based on that, we will explain how we developed the interview scripts and questionnaires for collecting participants’ subjective experiences. In Sect. 3, we will present our study setup. The findings and design recommendations are summarized in Sects. 4 and 5, respectively.

Fig. 1.

The V-tron ISA system was used in the study presented in this paper. It consists of a camera (right) that can recognize the speed sign on the road, a round-shape display (middle) that can show the current speed limit, and other safety-related information.

2 Background and Literature Review

2.1 The Definition of Acceptance

Acceptance, acceptability, and social acceptance are all terms used to describe "how potential users will react and act if a certain measure or device is implemented” (Vlassenroot et al. 2010, p. 167). If there is public/social support, the effectiveness and success of an initiative will increase. For the implementation of ISA, it is essential to know whether drivers and other public will accept the system and what factors influence the acceptance of this technology. Although the importance of investigating acceptance and acceptability is well recognized, there needs to be more consistency between studies as to what acceptance is and how it is measured [cited in Regan et al. (2014); Mitsopoulos et al. (2002)]. The definition of acceptance is the basis for the assessment structure and acceptance model, “without a definition, it is not possible to examine the validity and reliability of any assessment methods and/or models” (p. 12). Acceptance has been defined differently but partially overlapping in numerous studies.

Adell (2009) divides acceptance into five categories. The first category is “accept”; the second one considers the need and requirements, which are relevant to the usefulness of the system; the third type views acceptance as the aggregation of attitudes; the fourth category is the intention to use, and the fifth one emphasizes the actual use. They can be seen as an evaluation process from usefulness to actual use. It shows that acceptance is a multifaceted concept and that researchers select the propensity to conform and limit its scope (Regan et al. 2014).

Some researchers have further subdivided attitudinal acceptance and behavioral acceptance, conditional acceptance, and context-based acceptance, as well as distinguished between acceptance and support [see Ch. 2 in Regan et al. (2014) for a detailed review]. Although there are different definitions, what they have in common is that “acceptance and acceptability are … based on the individual’s judgment” (p. 12). Any assistance system only produces the desired effect when used by the user, which means that using the system is essential (Regan et al. 2014).

Since the users recruited in this project use the ISA system realistically, based on Schade and Schlag (2003), we ask users about their forward-looking judgments about the system and their expectations of using it during the baseline phase as acceptability and ask them about their attitudes and behaviors toward using this system during the testing phase as acceptance. The acceptance is closely related to the usage, and the acceptance will depend on how the users’ requirements are integrated into the development of the system. Therefore, in our study, we take the definition of acceptance proposed by Adell (2009), “the degree to which an individual intends to use a system and, when available, to incorporate the system in his/her driving” (p. 31).

Another related concept, social acceptability, considers a broader range of factors affecting acceptance related to security and economics, but also some with cultural, social, and psychological significance (Otway & Von Winterfeldt 1982). Researchers have studied social acceptance and social acceptability to understand the impact of various potential social contexts and their specific factors on human interaction with technology (Uhde & Hassenzahl 2021). Wüstenhagen et al. (2007) propose a triangle of social acceptance of renewable energies, including socio-political, community, and market acceptance. According to Vlassenroot et al. (2010), “social acceptance is a more indirect evaluation of consequences of the system” (p. 168); it involves a broader range of factors beyond the direct operating system. Vlassenroot et al. proposed 14 indicators most likely to influence acceptance and acceptability for the ISA scenario, adding factors such as personal and social aims, social norms, responsiveness awareness, and affordability. Since their subjects were people with no experience using the ISA system, their findings were used by us more in the pre-interview stage of our study.

2.2 Assessment of Acceptance

Since the driver's judgment of the system is based on personal knowledge, understanding, and experience, this may differ from the influence of the system as measured by an external observer (Adell 2009). Several widely applied models exist to measure the impact on acceptance: the technology acceptance model (TAM) proposed by Davis et al. (1989); TAM2 by Venkatesh and Davis (2000), and TAM3 by Venkatesh and Bala (2008), which is the extension of TAM. There are also some other models that extended TAM with other methods, such as the unified theory of acceptance and use of technology (UTAUT, Venkatesh et al. 2003) with the integration of several widely used models and some comprehensive research. Lastly, Osswald et al. (2012) incorporated the influence of contextual information in TAM and proposed the car technology acceptance model (CTAM) based on the abovementioned UTAUT. Two crucial elements were introduced into the CTAM model: anxiety and perceived safety. After conducting a literature review, we selected the following influencing factors from several models that fit our research.

Individual Factors. From Rahman et al.’s (2018) survey, we know that age and gender are the most frequently cited demographic factors in the mobility domain, although some studies have concluded that gender has no significant effect on acceptance (Rahman 2016). Besides, Adell (2009) found the effect of age was more in the perception of usefulness, satisfaction, and keeping the system, with younger drivers rating lower in these categories than older drivers. Some studies also point out that a driver's experience with similar devices can affect the acceptance of new technologies (Höltl & Trommer 2013; Rödel et al. 2014).

Driving habit is a broad phenomenon that covers the choice of driving speed, distance to a preceding vehicle, overtaking other vehicles, and the tendency to commit traffic violations constitute behavioral tendencies of drivers. These habits are often described as ‘driving style.’ Collecting the driver's driving style helps to interpret which habits have more significant effects on their current driving behaviors and to compare them with the use of the ISA system afterward.

Expectations of the System. According to Compeau et al. (1999), expectations are the user's prediction of the system’s purpose and function before actually using a system or product. It is also called outcome expectations, such as increased efficiency and improved quality. This item allows us to understand the user's vision of the new technology and validate it after use.

Perceived ease of use means the degree to which users expect the target system to be easily operated (Davis et al. 1989). It is called Effort Expectancy in the UTAUT model (Venkatesh et al. 2003) and the CTAM model (Osswald et al. 2012). According to Venkatesh et al., the effect of effort expectancy on intention is more pronounced among females and more senior users. In Davis’ (1989) study, this factor was expected to be more prominent in the early stages of using a new system.

Perceived Usefulness presented in TAM (Davis et al. 1989) means the extent to which the user subjectively believes that using the target system will improve their performance. TAM theorized that perceived usefulness strongly influences users' behavior intention (BI), and perceived ease of use also significantly affects them but diminishes over time (Davis et al. 1989). Meanwhile, in their conclusion, perceived usefulness was also influenced by perceived ease of use because the easier the system is to use, the more useful it will be. In the UTAUT model (Venkatesh et al. 2003), it is called performance expectancy; the authors conclude that performance expectancy is a determinant of intention in most cases, and this relation is influenced by gender and age, which is more pronounced among males and younger users (Venkatesh et al. 2003).

Effectiveness is defined as “the extent to which a system performs its intended tasks” (Rahman et al. 2018, p. 136). Many studies recognized that system reliability is an essential factor that affects the effectiveness and act as a barrier to acceptance. Some common examples of poor reliability are false/nuisance alarms and accuracy when using the system.

Attitude is defined as “an individual’s overall affective reaction upon using a system” (Osswald et al. 2012, p. 54). This factor is intended to reflect the user's perception of the use of the system and its impact, going beyond an assessment based on pure functionality. The initial attitude toward the system mainly affects workload, emotional state, and usage (Adell 2009). Venkatesh et al. (2003) argue that attitudes toward the use of technology are not theoretically a direct determinant of behavior intent. But in CTAM, this element is considered because we cannot assess the effect in a car context in advance (Osswald et al. 2012).

Anxiety and Perceived Safety are the two crucial elements in CTAM (Osswald et al. 2012). Anxiety was defined in the car context as “the degree to which a person responds to a situation with apprehension, uneasiness or feelings of arousal” (p. 55). Perceived safety was defined as a judgment of individual driving skills and a sense of safety in relation to other drivers (p. 55).

Social influence is also called the subjective norm in TAM2 (Venkatesh & Davis 2000). It refers to the influence of people who are important to the user's view of whether he/she should perform the behavior or not. Vlassenroot et al. (2010) observed that “peers, e.g., co-workers or specific other road users, will influence the attitudes and behavior of individuals” (p. 176). Venkatesh et al. (2003) also found that “the role of social influence in technology acceptance decisions is complex and subject to a wide range of contingent influences” (p. 452). Based on the TAM2 model (Venkatesh & Davis 2000), people incorporate social influence into their perception of usefulness, i.e., gaining status and influence through the use of systems and thereby improving their job performance. When people gain direct experience over time, they rely less on social information in forming their intentions and instead make judgments based on the potential benefits that come with use.

2.3 Summary

Based on the context and purpose of this study, we integrated the theoretical model described above. As this study focuses on the driver’s actual use of the system, we divided all relevant factors into before exposure to the ISA system and while using the system. Individual factors and expectations of the system are investigated before the system is activated. Perceived ease of use, usefulness, effectiveness, attitude, anxiety, safety, and social influence, which are direct determinants in the driving environment, are investigated when drivers use the system. The figure below shows the specific meaning of these factors in the driving context.

3 Study Setup

In order to investigate (1) the effectiveness and usefulness users can perceive, (2) the usability of the system, (3) user experiences and social acceptability, and (4) the possible changes in their driving behaviors. We interviewed two user groups in this study: A. voluntary people with their vehicles; B. five tutors of a driving school at Helmond. For type A users, we used interviews and questionnaires to understand their experiences. We organized focus group discussions for the B group to know their observations and suggestions for the tested ISA system.

In total, we recruited 12 participants, including seven people with their vehicles and five instructors from the driving school where the ISA system was installed on four cars used for driving lessons.

3.1 Three Phases for Individual Participants

(1) Collecting Participants’ Expectations and Prior Experiences. We recruited our participants through the project website and social media channels of the local communities. After signing up for our study, we arranged a meeting with them to explain our study of collecting their subjective experiences. When they agreed with our plan and signed the consent form, we asked them for some basic information related to this study, including demographic info, gender, and diving experiences. We asked them the average annual kilometers they had driven and their driving style preferences (van Huysduynen et al., 2015), which may help to interpret the events that occurred during the test. We also collected contextual information, like how often the user drove and how familiar they were with the road conditions, which could affect their driving experiences with a new assistant system in place. At the end of the pre-interview, we asked participants about their expectations of the ISA system.

(2) A Weekly Questionnaire Survey for Three Weeks. During the testing period of three to four weeks, we collect their experiences by asking them to fill in a short questionnaire weekly. The interactive questionnaire was implemented LimeSurvey. It covers six dimensions of TAMs, including Perceived ease of use, Perceived Usefulness, Effectiveness, Attitude, Anxiety and Perceived Safety, and Social Influence. There are multiple-choice questions that a participant can easily express their attitudes and experiences. We also used Likert scale to design questions to investigate how much they agreed with the specific criteria. Taking “information accuracy” as an example, we asked, “The information from the system was accurate,” and they can drag a slider from 1 = “ strongly disagree” to 7 = “strongly agree” to indicate their experience. In case of negative feedback, additional open-ended questions would appear to ask participants for more details and explanations.

To protect participants' personal information while filling in a digital questionnaire, we customize a link for every individual participant and send it to their mailbox on the day of their convenience. In this way, they don't need to fill in their personal information every time, and we can proceed with their inputs anonymously.

(3) Post-Interview. After the testing phase, we conduct a post-interview to collect participants’ total experiences, improvement feedback, and possible changes in their driving behaviors. To host the post-interview smoothly and efficiently, we ask participants to fill in a summary questionnaire. It covers the Car Technology Acceptance Model and some specific questions based on the results participants reported in the weekly questionnaire. We synthesize their responses and produce a UX (user experience) curve for each participant to investigate what makes the changes in their experiences and opinions throughout the entire study. We also learned improvement ideas from the participants and discussed possible marketing proposals to understand which subscription models and pricing are affordable for the target customers. The main questions include the following:

• Effectiveness: Do you think driving a car with this system will make you drive differently? Followed-up question: In what ways? Under what conditions?

• Usefulness: How useful would you find the technology? Would it serve a purpose for you?

• Usability: Can you think of any potential problems or concerns that you might have in using the system? For instance: Source of distraction? Potential for over-reliance? Reliability issues? Issues with the look and feel of the warnings?

• Social acceptability: How would you feel if it were compulsory for you to fit this technology into your vehicle?

• Total experiences: From 1 to 10, how satisfied are you with the system? Can you also explain why?

3.2 Focus Group Study with Instructor Participants

In this study, we organized a focus group discussion with five instructors at Rijschool VOC van Oijen in Helmond to learn about their experiences from the second-person view. They are the secondary target users of the ISA system, and we want to understand their observations on their students’ usage of the system. This can help us to understand their expert opinions and examine whether the driving school could be a proper market for the system. We start the discussion with the TAM questions covering five dimensions: Effectiveness, Usefulness, Usability, Social acceptability, and Affordability, and probe their opinions collectively.

4 Findings

Due to technical issues and the delay of needed equipment, our driving experiment started in the second week of December. Due to the Christmas holiday, many participants couldn’t respond to our questionnaires. As a result, we received 18 responses from participants’ questionnaires. In this section, we first present their feedback with diagrams and explanations, followed by a more detailed analysis of the interview data discussed later in this section.

To keep track of the feedback gathered from different participants, we used the following acronyms to mark the source of the specific quote:

• PreIn (pre-interview): the data we collected from the first interview conducted before the ISA system installed in the participant’s vehicle was activated.

• PxQy (the yth weekly questionnaire): the responses we collected with the yth weekly questionnaire participant x filled out.

• PosIn (post-interview): the data we collected from the second interview we did when the participant completed the four-week driving experience experiment.

4.1 Subjective Responses

First, all users perceived the ISA System as easy to use throughout the experiment. Although three participants (5 [28%] responses) were neutral about this option, all other users found the system very easy to use.

Secondly, when we asked participants their perceived usefulness (including, the system is useful and using this system will improve my driving performance), their responses were a bit disputed. Half of them felt positive. They indicated that when the system works well, it helps them to understand the current speed limit and to be more focused when driving. From a more macro perspective, the system can contribute to road safety. And half of the responses were neutral or opposed because of problems with the system that influenced their judgment on perceived usefulness. The two main issues reported by the participants are the information misalignment, and the system does not react quickly. Fortunately, users widely agreed that if the system worked well, it would be beneficial for assisting their driving.

Fig. 2.

The percentage of participants’ perceived values on (A) usefulness, (B) improving driving performance, (C) providing helpful information, and (D) feedback clarity. The amount of total responses is eighteen.

Thirdly, participants expressed some complaints. “Information Accuracy” has the most popular comments: except for one user who rated it neutral, all other users reported inaccurate information (one user thought the information was accurate in the first week and changed to disagree after two weeks of use). Incorrect speed limits may lead to danger, so this point directly affects the user's overall attitude and perceived usefulness of the system. Since the system provides the wrong speed limit, and users need to override it, half of the responses consider the system annoying. We discussed all those issues with the participants in the post-interviews and focus group discussions, and the results were reported in the following sections.

Fig. 3.

The percentage of participants’ perceived values on (A) the accuracy of information provided, (B) capability for avoiding traffic accidents, (C) improvement of safety, (D) believability, (E) assistance, (F) annoyingness, (G) Desirability, (H) pleasantness, and (I) overall satisfaction. The amount of total responses is eighteen.

4.2 Positive User Experiences

In our interviews, most participants agreed that the ISA system could be an effective assistance system for improving road safety. For instance, Participant 2 (P2) mentioned that the situational speed limit could also educate the drivers on why the authority limits the speed to a certain degree in specific areas. This information can persuade drivers to care about other road users’ safety and adjust their driving speed accordingly. The user interface is straightforward to understand without distraction. When used in an old car without an embedded information system, the system can be an easy and affordable upgrading solution for providing situational information and speed control. We synthesize their positive feedback into the following themes.

1. Improving Drivers’ Awareness of Speed Limits

Thanks to the camera sensor and integrated information system, the system reduced drivers’ efforts to check the momentary speed limit and pay attention to road signage. Some participants reported that the system made them more aware of the speed and the safety it brings. For example, Participant 7 likes that “the safety speed is monitored automatically” (P7Q2). By using this system, drivers can be more aware of the speed limit and calmer, “The system monitors speed so I do not commit a violation and can concentrate more on traffic” (P7Q2).

2. Reducing or Changing Speeding Behaviors

Before the testing starts, all participants express positive expectations of the system’s assistance in avoiding speeding and preventing potential accidents or traffic tickets. During the experiment period, the system acted well in restricting their speeding. Furthermore, some participants also found additional benefits in informing them of the correct speed limits and triggering them to reflect on their existing behaviors. For instance, when P7 drove the regular commuting route in the second week of testing, he noticed that the speed limit shown on display was lower than he thought (i.e., 50 km/h). He said, “I now realize that I've been driving too fast in the past” (P7Q2). Most participants acknowledged that the system raised drivers’ mindsets back to the limit. P2 further pinpointed an educational opportunity the system can enable; she said, “ the government would put a little bit more focus on giving the pedestrians and cyclists room, may be aware of environmental issues if you speed and break and speed, what does it mean for the pollution, for CO2. So I think there are so many elements that can be converged to education process towards the drivers” (P2PosIn).

5 Visual Display

Among the ISA system evaluated in this study, there is a round shape display installed on the dashboard within the participants’ vehicles. All participants indicated seeing speed restrictions but no other driving assistance alerts. The placement, size, visibility, and clarity of the display were well received by all the participants. P1 said, “The visual display was placed correctly and easy to see” (P1PosIn). Based on the combined feedback from the participants, the display works correctly and appropriately. The current information suits the way the screen communicates with the participants. However, slight improvements can be made by making the display adapt better to dark conditions (see driving school feedback), and if additional, more urgent information is displayed, sound warnings could be a beneficial option as well. If the future system wants to provide contextual information for educating the driver on the reactionaries of the speed limit set by the authorities (P2PreIn), a larger display or auditory feedback will be needed.

5.1 Issues that Need to Be Improved

Through this real-world evaluation, our participants reported several issues that affected their driving experiences. We synthesized their repones and the problems they encountered into the following five themes:

1. Comparability Issue: Some participants reported that the start-up process was sometimes slow. For example, P2 mentioned, “It could take up to half an hour for the system to turn on after starting a drive. Sometimes it didn’t turn on at all” (P6PosIn). “System frequently does not turn on or drops out for a few minutes while driving” (P1Q2). Another point of interest can be found in the different cars available. P5 indicated how electric vehicles with regenerative capabilities brake significantly when power to the pedal is cut “Electric cars break quite hard when releasing the pedal and regenerating energy” (P5PreIn). Two other participants also mentioned this question (P3Q2, P8Q2). P9 thought, “The over-intervention of the system, not functioning and not thought through for an electric vehicle” (P9Q2). This could also lead to dangerous situations, especially when the system behaves unexpectedly.

2. Misalignment of the Speed Limit Information: In this trial, all participants reported that in some areas, the system’s speed limit setting was different from that shown on road signage or in mobile apps, such as Google Maps or Fitsmeister (a Dutch mobile app for assisting driving). For instance, P5 noticed several inaccurate speed limits when he drove around the city, especially when there were multiple roads close to each other and every road had different speed limits. P5 said,”where I live here, a lot of situations were 30, 50, and 80 (speed limit), the parallel roads just very close… On the right side, that's the different road, that is 30, and you are on the 50 roads. The GPS is not precise enough to recognize 30 or 50; it shows 30 instead of 50…” (P5Posln). P3 reports that “on the highway, it often shows 50 or 70 of the speed limit; fortunately it was on cruise control. Otherwise, it would have reduced the speed suddenly” (P3Q2). Another technical problem is the map accuracy. All participants indicate a need for updated maps, paired with better speed recognition “my advice would be to make sure that the speeds of the system are correct so that the user experience becomes much more pleasant. There will be fewer irritations” (P2Q1). P5 further suggested assessing the speed limit dynamic because “different cars have different dashboards, it’s always different in bandwidth - if the 50 was 50, or if it’s 47 or 53… so which should not be limited to exactly 50 but allows for a 10% variation” (P5Posln).

3. The Latency and Unexpected Acceleration while Overruling the System: Some participants reported that the reaction time was not as quick as expected when they pushed the paddle to temporarily deactivate the system. P2 said, “unexpecting situation when you need to overtake other cars, the system’s reaction is slow” (P2Q1). During the interview, we also heard that “when I want to overtake a slower vehicle and change to the inner lane, my car can’t go beyond the limit quickly. This latency could be dangerous because this delay might increase the chance of being hit by other vehicles” (P2PosIn). On the other hand, P9 reported extreme acceleration caused when overruling the system, “I must kick the system out, which results in an immediate torque of 100% in the vehicle. This, in turn, manifests itself in an extreme acceleration of at least 10km/h extra” (P9Q2).

4. Safety Concerns: In some situations, the restricted speed limit might increase the danger of driving. P9 reported that “Often the car limits itself which leads to severe speed reduction, and on days like today where the road surface is extremely wet, the car has to “pedal” through it leading to slipping tires.” (P9Q2).

5. Trust in the System: Most participants did not fully trust the system due to the above four issues. P2 said, “sometimes the speed [limit] given by the system does not correspond with the reality, which is a major issue when trusting the system” (P2Q1). P1 also mentioned, “Trust in the system will only occur when it is always correct” (P1PosIn).

6. Impact on Other Road Users: A recurring theme within the interviews with all participants was that they felt annoyed about being rushed by other road users. Using the system in an isolated environment (meaning only you have the system, not the drivers around you) indicated that participants frequently experienced speeding cars, and they resorted to tailgating and other dangerous driving behavior. For instance, P4 said, “I have experienced people driving extremely close to me, and at times overtaking me on stretches of road where this wasn’t allowed” (P4PosIn). This situation sometimes impacts other road users. P6 noted that “it is irritating because you still want to go with the traffic if the speed is lower, then it irritates other road users” (P6Q2). When a truck or aggressive driver pushed a user's car in the back, they felt very uncomfortable (P3PosIn). Sometimes, “it could be dangerous” (P2PreIn) if the other cars behind us do not always notice that we are reducing speed quite rapidly” (P8Q2). To improve communication with other drivers, many participants wished to have a signal on their vehicles to tell others their speed was restricted within the safety range. P2 shared a workaround approach, “when I was driving slower than other cars, and I could not immediately pick up the speed, I needed to press the flash to inform others ‘sorry, I cannot speed over 70’” (P2PosIn).

5.2 Findings from the Focus Group Discussion with Five Instructors

The focus group discussion was held on Dec. 8th in the office at Rijschool VOC van Oijen in Helmond. Five instructors participated in this two-hour section, including four full-time driving instructors and one supervisor/team leader who is also a part-time driving school instructor. V-tron installed the ISA-System on four school vehicles in September, and all participants had the experience that students used them during the driving lessons. Throughout the focus-group discussion, the researchers observed and noted strong coherence and support between participants. The main findings and points of interest will be discussed below.

The Benefits and Disadvantages of Having a Mandatory ISA System. We first discuss the European Vision of having the ISA system on all vehicles. One said, “If everyone has a system and the speeds are correct, then yes, I think the biggest irritation is all gone” (FCP5, participant #5 of the focus group session). While their collective knowledge also included driving a truck, which would benefit from a similar system. “The only difference between a truck and a passenger car is (…) the braking deceleration of a truck is not as great as the braking deceleration of a passenger car” (FCP1). According to the two participants who owned a motorcycle license, motorcycles might be the only road-legal vehicle that an ISA-system could not safely fit. One said, “A motorcycle is a balanced vehicle, and suppose I'm in a corner and the gas is reduced in one go, then you have a chance that the vehicle will crash due to an imbalance” (FCP2). When reflecting on their driving and the collective road safety enforced by the ISA system, one participant worried that their driving experiences might be hindered. He said, “from my own point of view, I agree that making ISA systems mandatory would reduce driving pleasure, but from the safety point of view, it would make sense. Those speeds don't count for nothing, do they?” (FCP1).

System Inaccuracies Lead to Decreased Acceptance and Commitment. Similar to the inaccurate problems reported and discussed in the previous section, all instructors shared several observations on the wrong speed limit shown in the display. For instance, FCP3 said, “I wrote down a couple of observations that the system was wrong when there was an overpass or close parallel road.” Two participants continued to share their ideas of how it could be improved. They indicated that intelligent systems could be installed in several cars to collect data, and based on group behavior, correct/appropriate speeds on the road could be gathered. “Collecting data, yes, that does not matter a lot to me” (FCP2). “If we can contribute to road safety in Helmond and a better road network, yes, then I would like to contribute to that” (FCP3).

Other unmet requirements hinder their acceptance and commitment to the study. The first one is the moment of activation of the speed limit. FCP1 said, “Our students also need to look ahead. Yes, you see that there is a 50 sign, so you have to release your gas pedal. […] When you see an increased speed sign, you can already start increasing speed, so you should actually react before the board. Now [the ISA system] only responds two or three seconds after the sign” (FCP1). The other one is the latency while overruling the system. One instructor said, “[A student needs to] be able to overrule the system when it makes mistakes, more directly, so not those two seconds, but immediately” (FCP5). Considering those two unmet needs and the inaccuracy of the speed limit, the instructors only briefly turned on the system for their pupils to experience it for a short time. Then, they turn it off completely during the actual driving lessons. “The students also indicated almost unanimously that they would have preferred systems off” (FCP4). All instructors acknowledged how incorrect speed reductions caused by the system could lead to dangerous situations (FCP3).

They are driving Autonomy and Possible Challenges of Adapting New Assistant System. New driving assistant systems are constantly developed and improved. Most aim to improve driving capabilities and safety. This is something the instructors all confirmed. One example is the navigation systems that display both speed limits as well as current vehicle speed. “It shows … how fast you are allowed to drive; then it says how fast I drive and the faster I drive it changes color, say from green to orange to red. (…) I always find it [quite] useful” (FCP1). However, they further discussed the effect these systems can have and how they regularly observe their students to fully trust and follow a system instead of basing their decision-making on their observations. “[When using the ISA system] they will no longer pay attention to the signs, then it will decrease road safety because there will be mistakes” (FCP3). This is confirmed by FCP2, who said “Because the students are still too much focussed on watching that screen than reading the speed signs” (FCP5). While observed in many students, this effect is not precisely the same in all drivers. The participants explain how self-confidence changed their reliance on the systems and how easy it was for them to stay in control while driving. “Some students have a little more self-confidence than others and one student can handle this better than the others” (FCP2).

6 Discussions and Design Recommendations

In this research, all participants acknowledged the vision and potential of ISA System in reducing speeding and improving traffic safety. One driving instructor said, “I think almost everyone would like to have the ISA system if it works well” (FCP4). They helped us identify several technical and social issues that need improvement and consideration in further developing the ISA-System into a robust product. Based on their feedback and comments, we further specify some concrete recommendations for technology improvement, infrastructure, and citizen education that the city hall and relevant organizations can collaborate on to improve overall traffic safety by utilizing and gradually introducing the ISA System to prospective drivers.

6.1 Technology Improvement

Improving the Accuracy of the Speed Limit. The product development team needs to thoroughly validate the system’s output with the actual speed limit on the road. One approach is to work with the city and follow a well-established test procedure: Link to the document. According to the protocol, manufacturers need to ensure that their system can correctly identify 90% of the speed limits along a 400–500 km test route consisting of a mix of motorways, highways, and urban roads. The instructors from the driving school are pleased to help. Providing them with a method to indicate inaccuracies through, for example, an alert button could help the developers to focus on map areas that could benefit from revision precisely.

Solving the Reaction Time Issues. It includes the slow starting up and the latency while overruling. A typical case worth investigating is the lane-change situations on different roads, such as on low-speed roads or high-speed highways. On the other hand, from the driving school instructors, we learned that there was some delay in showing the information on the display when a speed sign was already visible in the front. This is a crucial requirement for their driving teaching practice.

Re-examining the Speed Control in Different Weather Conditions (such as heavy rain and snow). In our study, we found one case that severe speed reduction might lead to slipping tires in highly wet situations. The technical teams need to test the possible effects of variant weather to ensure the system can function all the time. Whenever there is any uncertainty, the system could automatically deactivate.

Re-evaluating the Existing Design of Overruling by Pushing the Pedal. According to the focus group and the feedback of many interviewees, whether automatic speed limitation by pedal is, the best solution still needs further investigation. Because the system automatically slows down, and the user needs to step on the pedal again to override the speed limit, the user feels that it is not them but the system that controls the car. We propose a new design for installing two steering wheel buttons (see Fig. 4 middle). When a driver wants to overrule the system, he can press both buttons simultaneously and use the pedal to speed seamlessly without latency.

Fig. 4.

A scenario of using the new design buttons for overruling the speed control by pressing buttons on the steering wheel.

Providing Deactivate Mechanisms for Other Users. Providing an easy function for temporarily disabling the ISA system when the vehicle is used by other users unfamiliar with the system, including other family members or students in the driving school. However, this flexibility should encourage the primary user to refrain from overrule the system frequently. To do so, we propose a concept (see Fig. 5) of providing a monthly quota (e.g., ten times) for users to temporarily deactivate the ISA function for others to use the car without the need to get used to it. We expect that this mechanism could help to increase users’ acceptance of the ISA system and eventually to improve road safety.

Fig. 5.

A scenario of using the mobile application and service to temporarily deactivate the ISA function when the other family member or friend uses the car.

Having a Reasonable Margin of Speed Limit (e.g., 10%). Most participants indicated it is fine and normal if their speed is controlled below 110% of the speed limit. This means that the threshold could be considered to align with the users’ mental models, especially when the system is not sure on the correct setting. This will also give some flexibility when the road conditions are empty (e.g., no other visible cars in the front).

Checking and Improving the Comparability with Variant Vehicles. The system is still in development, and unexpected new situations will regularly occur. We recommend exploring the system on different cars, including Electric, Gas, and other less occurring vehicles. Consider these and adapt possible system features to suit these vehicles better.

6.2 Redesigning the Interaction Design to Increase and Promote Acceptance

Drivers with a tendency to speed might often do so consciously and will regularly use an overrule option whenever presented to them. The system could be redesigned in a way that the system can let drivers feel in control and gradually negotiate to take over the control rather than reducing the speed automatically. This strategy could increase users’ awareness of speeding and promote their acceptance. We got one valuable idea from participant P1: “I would continue to feel more in control if the system would first notify me visually, followed by an auditory alert before finally regulating my pedal” (P1PostIn). This, however, could work counterproductive, as a certain level of speeding would become easier and more accessible by users.

Furthermore, many participants believe that accepting and using the system requires users’ willingness to contribute to safety, and local authorities might need to play a role. For people with long driving experience, it might be harder to fully follow the system because they have their own habits, “You need to re-educate yourself… you need to have an open mind for that to drive according to the system” (P2PosIn). This issue is difficult to combat and requires a governmental solution/approach. The city can consider organizing campaigns or educational programs to let citizens understand and experience the ISA system and gradually increase its adoption.

Based on participants’ feedback and suggestions, we propose a design concept (see Fig. 6) of providing visual and auditory feedforward (Chuang et al. 2018; Chuang 2020) to notify the driver of the change in speed limit and explain why the authority set it that way. We hope the additional explanation can reinforce drivers’ understanding and action without speeding.

Fig. 6.

A scenario of providing informative feedforward to notify the upcoming change of speed limit ahead and explain why the authority set the limit for the area. When the driver reaches the exit of the area, the system will speak out to notify the new change of limit.

7 Conclusion

While ISA systems show potential to reduce speeding and improve driving safety, our findings suggest improving the technology and developing the adaptation campaign from broader perspectives of the community and city. Regarding inaccurate or misalignment issues, overruling should be an easily accessible feature that drivers should use without hesitation. This also applies to over-reliance on being able to drive the speed limit while a slower speed might be needed for safety reasons. Our design proposals utilized multimodal feedback and feedforward for communicating essential information to the driver, especially in newer and, thus, less experienced drivers. If the system was to become mandatory, we recommend regulators first introduce the system as obligatory in driving schools before the system becomes mandatory in all privately owned cars.