Effect of Proactive Interaction on Trust in Autonomous Vehicles

Round 1

Reviewer 1 Report

In my opinion, the paper is well described and thorough in its approach to explaining methods and the context around the work. Some comments to address:

P4L163 the descriptions of the PI should provide more detail as to exactly what is said to the driver. 

P6L241 I would like to see a more detailed technical description of the VR headset used in terms of effective resolution, refresh rate and field of view. Please add comment around whether the authors feel the technical capabilities of the headset provide suitable immersion for the results to be applicable to real-world driving. 

Figure 5- the familiarisation run should be added to this diagram

P7L260 Please indicate what questionnaire was used for the self-assessment of driving proficiency

Figure 6- the x-axis labels should align to the low, med and high PI labels that have been used throughout the paper. Otherwise this inconsistency can be confusing

Conclusion- please draw out the original contributions more clearly in the conclusion

Key point- in the abstract you say that the paper "provides guidance for interaction designs to increase trust, thereby enhancing the acceptance...".But this is not evidently clear from the concluding paragraphs. Please add a section whereby the key design recommendations are highlighted, ideally as bullet points, for a reader to get the value out of the paper. This is the key issue with the paper in its current form and why I recommend a major revision. Please ensure that this is evident throughout the paper, with clear guidance provided from the results wherever possible

Author Response

Dear Reviewer:

Thank you very much for your time and consideration. We do appreciate your valuable comments and suggests, which helped improve our work a lot.

Based on your comments, we uploaded a revised manuscript, and a letter of our responses to the comments.

We hope the revised manuscript is acceptable for publication in Sustainability.

Sincerely,

Authors

Responses to comments are as follows:

1. P4L163 the descriptions of the PI should provide more detail as to exactly what is said to the driver.

Thank you for pointing out the need for clarification. In response, we have added more specific details to the manuscript regarding the interactions between the system and the driver:

(P4L160) “(i) High PI: The system provides the driver of situation of the surroundings or traffic circumstances based on information gathered from sensors. It then suggests decisions according to those sensory inputs and requests confirmation from the driver. After receiving feedback from the driver, it makes a received respond and executes or does not execute the previous decision based on the driver's feedback.

(ii) Medium PI: The system provides the driver of situation of the surroundings or traffic circumstances based on information gathered from sensors. The system then pro-vides the driver with a decision according to those sensory inputs and then directly executes the decision.

(iii) Low PI: The system executes directly without interaction with the driver.”

2. P6L241 I would like to see a more detailed technical description of the VR headset used in terms of effective resolution, refresh rate and field of view. Please add comment around whether the authors feel the technical capabilities of the headset provide suitable immersion for the results to be applicable to real-world driving.

Thank you for your reminder. To demonstrate that the VR headset can effectively simulate the real-world driving experience, here is a supplemental technical description of the PICO neo 3:

(P5L251)“…The PICO Neo 3 headset (4K resolution, 120Hz refresh rate, 98° FOV) features integrated spatial audio and high-quality stereo speakers, capable of playing white noise from the car engine during the experiment to maximally simulate the real driving environment.”

 3. Figure 5- the familiarization run should be added to this diagram.

Thank you for your suggestion. The familiarization has been added in Figure 5 as recommended.

4. P7L260 Please indicate what questionnaire was used for the self-assessment of driving proficiency

Thank you for your query regarding the self-assessment of driving proficiency. This assessment aimed to investigate indicators related to the driver's self-efficacy rather than the objective level of driving skills (which was reflected through questions about driving experience and total kilometers driven). Self-efficacy may be a personal factor influencing trust, as previous studies have indicated that individuals with higher self-efficacy are more likely to trust and use AVs because they believe in their driving ability to control the technology and effectively handle potential challenges. Therefore, we conducted a simple survey on the participants' "self-assessment of driving proficiency", which consisted of a single-item measure with 5 options ranging from "I think my driving skills are very poor" to "I think my driving skills are very good". We believe this method suffices for the scope of our study instead of a specific questionnaire, and we appreciate your understanding.

The reason you might have raised this concern could be due to a lack of elaboration in our manuscript. To address this, firstly we have supplemented our literature review with additional research on self-efficacy:

(P3L122)“… Additionally, research shows that individuals with high self-efficacy are more likely to trust and use a system because they believe they can master the technology and effectively handle possible challenges [2].”

Furthermore, we have expanded on the personal information questionnaire section of our study, including the purpose of collecting personal information and the form of the questionnaire items:

(P7L269)“… They were investigated by a single-item survey for basic demographic details…and self-assessment of driving proficiency to fully understand the features of the participants.” 

5. Figure 6- the x-axis labels should align to the low, med and high PI labels that have been used throughout the paper. Otherwise this inconsistency can be confusing

Thank you for your suggestion. We have supplemented the “Low PI”, “Medium PI” and “High PI” below the x-axis labels in Figure 6-8.

6. Conclusion- please draw out the original contributions more clearly in the conclusion.

Key point- in the abstract you say that the paper "provides guidance for interaction designs to increase trust, thereby enhancing the acceptance...".But this is not evidently clear from the concluding paragraphs. Please add a section whereby the key design recommendations are highlighted, ideally as bullet points, for a reader to get the value out of the paper. This is the key issue with the paper in its current form and why I recommend a major revision. Please ensure that this is evident throughout the paper, with clear guidance provided from the results wherever possible

Thank you very much for your suggestion. We have adjusted the structure of the conclusion section to clarify the logic, making it easier for readers to quickly get the value of the paper. The first paragraph now highlights the original contributions:

(P13L478)“ Our study designed different PI levels for AVs, subdivided voice PI behavior, and explored the differences in trust toward different PI levels by driving supervisors as well as the impact of personal factors on trust in AVs. The results provide practical guidance for the design of AVs that will increase drivers’ trust. Research on gender and personality traits helps understand how individual characteristics modulate the relationship between interaction behaviors and trust, offering a new perspective for personalized design and optimization of interaction strategies in AVs.”

The following paragraphs emphasize the interaction design recommendations, clearly listing out in bullet point form how the research results guide practical interaction design:

(P14L487)“(i) In the future, trust in the system can be enhanced by increasing the PI level. During voice interaction with the driver, it will be more helpful for increasing trust when AVs proactively report current situation changes to the driver, delegate the authority for driving behavior decisions to the driver under suitable conditions, and inquire whether to execute the provided suggestions.

(ii) When L4 AVs detect that the driver is handling other tasks, they may need to be cautious about informing the driver of driving decisions, especially male drivers, as this may have a negative impact on trust. Instead, AVs can choose to inquire, or simply execute the decision directly without any interaction.

(iii) It is also helpful to gather information about the driver’s personality beforehand to select interaction schemes for different extroverted groups. For drivers with extroverted but not extremely extroverted personalities, a higher level of PI can be utilized more frequently to enhance their trust in AVs.”

Reviewer 2 Report

Based on virtual reality (VR) simulated driving, this paper studies the influence of different levels of active voice interaction on the trust of L4 autonomous vehicle drivers. The experiment is carried out from three aspects: system, personal and situational. The repeated-measures ANOVA of SPSS software is used to study the difference of different factors on trust. The research results obtained have certain significance for improving the acceptance of autonomous vehicle.

1.     Why did the experimental VR environment choose a ramp merging scene to control risk perception variables? Can other scenarios be used?

2.     This experiment was conducted on the basis of voice proactive interaction. To complement the previous discussion, it is recommended to elaborate on the characteristics and advantages of voice proactive interaction.

3.     Please explain the symbolic meanings of inter group effects F3,51, F3,24, F3,25 in the Task and Non task sections of the analysis results 4.1.

In section 4.3, there is no explanation provided on the degree of extroversion represented by levels 3, 4, and 5 of Level of Extraversion. Please provide additional information on the degree of extroversion represented by levels 3, 4, and 5.

minor revision

Author Response

Dear Reviewer:

Thank you very much for your time and consideration. We do appreciate your valuable comments and suggests, which helped improve our work a lot.

Based on your comments, we uploaded a revised manuscript, and a letter of our responses to the comments.

We hope the revised manuscript is acceptable for publication in Sustainability.

Sincerely,

Authors

Responses to comments are as follows:

1. Why did the experimental VR environment choose a ramp merging scene to control risk perception variables? Can other scenarios be used?

Thank you for your question regarding the choice of the ramp merging scene to control risk perception variables in our experimental VR environment. We did consider other scenarios, but the ramp merging scenario was chosen for its effective control of risk variables for several reasons:

1. Ramp merging is a common driving scenario that all participants are likely to have experienced, thereby controlling for driving experience. This familiarity helps ensure that drivers do not perceive increased risk due to unfamiliarity with the situation.
2. In this scenario, particularly at high-PI compared to other levels, a safe distance is maintained between Car1 and Car2, regardless of whether the driver accepts the suggestion to accelerate. This is because Car2 is decelerating at this point. Consequently, the difference in risk between different interaction levels within the scenario is minimal, allowing us to isolate the effect of PI on trust without confounding factors related to varying risk levels. In other possible scenarios, if participants choose to accept or reject the suggestions made by the AVs, different driving behaviors could result in the exposure to varying levels of risk. Consequently, any observed differences in trust might be attributed to differences in risk perception rather than differences in the level of PI.

By using the ramp merging scenario, we aimed to create a controlled and realistic setting in which to study the impact of PI on driver trust, with minimal variation in risk perception across different levels of interaction. To make it more clearly, we supplemented the first reason :

(P5L199)“…Ramp merging is a familiar driving scenario for participants, reducing the likelihood of increased risk perception due to unfamiliarity.”

 2. This experiment was conducted on the basis of voice proactive interaction. To complement the previous discussion, it is recommended to elaborate on the characteristics and advantages of voice proactive interaction.

Thank you for your suggestion. We have taken your comments into consideration and have added advantages of voice proactive interaction:

(P2L59)“… Among various PI methods, voice proactive interaction is particularly noteworthy for its natural and intuitive communication.”

3. Please explain the symbolic meanings of inter group effects F3,51, F3,24, F3,25 in the Task and Non task sections of the analysis results 4.1.

Thank you for your question. In the results of the analysis presented in section 4.1, the symbolic terms F_3,51, F_3,24, and F_3,25 refer to the F-statistics and their respective degrees of freedom from the ANOVA test for the Task and Non-task sections. The 'F' stands for the F-statistic, a ratio that compares the variance between different interaction groups to the variance within each group. This ratio is used to determine whether the differences in means across the groups are greater than would be expected by chance.

The first number after 'F' represents the degrees of freedom between groups, which is the number of groups minus one. Here, '3' indicates we have four levels of the independent variable (since 3+1=4), which could be different levels of PI plus pretest in this case.

The second number represents the degrees of freedom within groups, which is calculated by subtracting the number of groups from the total number of observations. The '51', '24', and '25' are the degrees of freedom within for the Task and Non-task groups, respectively. These values are used in conjunction with the F-statistic to determine the p-value, which indicates the probability that the observed differences are due to random chance.

In our study, an F-statistic followed by degrees of freedom, such as F_3,51, indicates a specific test within the ANOVA. The resulting p-values associated with these F-statistics (reported in the same table) tell us if the observed group differences in mean scores for trust are statistically significant across the levels of PI.

The question raised regarding the symbolic meanings of the F-statistics and their corresponding degrees of freedom may due to a format mistake in our initial manuscript. We have since corrected this by placing the degrees of freedom as subscripts to the F-value( from F3,51 to F_3,51), adhering to the standard convention for reporting statistical results. The changes can be seen in the revised manuscript. We appreciate your attention to detail, and we believe that this clarification will make the results more transparent and easier to interpret. We have added a brief explanation of the statistical measures used in the study within the manuscript to ensure that our findings are accessible to all:

(P8L319)“…The ANOVA results indicated significant between-group effects, as evidenced by an F-statistic of F_3,51=6.703. This value is calculated from the ratio of the variance between the groups to the variance within the groups…The ANOVA results indicates a strong statistical significance between-group effects, as evidenced by an A p-value of less than 0.001(p<0.001). Specifically, significant main effects were observed for both the task group (F_3,25=15.549,p<0.001) and the non-task group (F_3,24=12.507,p<0.001), suggesting that the performance differences within these groups were consistent and not random. The effect sizes, measured by partial η², ranged from medium to large, which implies a substantial impact of the independent variable on the dependent variable.”

 4. In section 4.3, there is no explanation provided on the degree of extroversion represented by levels 3, 4, and 5 of Level of Extraversion. Please provide additional information on the degree of extroversion represented by levels 3, 4, and 5.

Thank you for your inquiry regarding the degree of extraversion for levels 3, 4, and 5 in Section 4.3 of our manuscript. We would like to point out that this information is indeed included in the manuscript in Section 3.6, where it is mentioned that ：

(P7L262)"All participants were required to complete a Big Five personality questionnaire [41], which classifies each dimension into five stages, numbered 1–5 from lowest to highest."

To clarify, we added the explanation of 3,4,5 in section 4.3:

(P11L413)“…Levels 3, 4, and 5 on the extraversion scale of the Big Five personality questionnaire correspond to moderate, high, and very high degrees of extraversion, respectively. “

We apologize for any oversight and will ensure this is clearly highlighted in the revised manuscript for the ease of the reader.

Author Response File: Author Response.docx

Reviewer 3 Report

The research implemented various levels of PI for autonomous vehicles, a notable approach given the significance of psychological behaviors in traffic dynamics.

Introduction is well written, and paper is well structured. However, I have few concerns which are given as

1. Text in Figure 1 needs to be clearly visible. Particularly, headings indicating Car 1 and 2.

2. What does the number 1, 2 ,3, and 4 represents in Figure 5. It should be explained clearly in text.

3. The statistics in section 4.1 are not clear. What does (F3,51=6.703, p<0.001), and (F3,25 = 15.549, p<0.001) mean. All these terms should be explained in detail for superficial readers.

4. Likewise, the terms in table 1 such as F, P, t and P, should be explained. Similar applies to the rest of sections.

5. What does P<=0.05, P<=0.01 and p<=0.001 indicates? It needs to be mentioned in paper.

6. What is meant by t test? This should be explained clearly.

7. Heading of Table 1 is mentioned below the table, whereas, for other tables it is above the table. Please follow journal formatting guidelines.

8. Authors have mentioned that this study enhances the sustainability of AVs. Can you explain it how?

Comments for author File: Comments.pdf

English language needs improvement.

Author Response

Dear Reviewer:

Thank you very much for your time and consideration. We do appreciate your valuable comments and suggests, which helped improve our work a lot.

Based on your comments, we uploaded a revised manuscript, and a letter of our responses to the comments.

We hope the revised manuscript is acceptable for publication in Sustainability.

Sincerely,

Authors

Responses to comments are as follows:

1. Text in Figure 1 needs to be clearly visible. Particularly, headings indicating Car 1 and 2.

Thank you for your suggestion regarding the visibility of the text in Figure 1. We have added a background color to the content labels for Car 1 and Car 2, which significantly improves their prominence and clarity against the figure's background. We believe these changes address your concerns and ensure that the text are now clearly visible for all readers. Please find the updated figure attached for your review.

2. What does the number 1, 2 ,3, and 4 represents in Figure 5. It should be explained clearly in text.

Thank you for your valuable feedback on the labeling of the trust questionnaires in our experimental flowchart. We have revised the labels to "Q1," "Q2," "Q3," and "Q4." These correspond to the first, second, third, and fourth administrations of the questionnaire, respectively. This streamlined approach ensures clarity without compromising the flowchart’s readability.

3. The statistics in section 4.1 are not clear. What does (F3,51=6.703, p<0.001), and (F3,25 = 15.549, p<0.001) mean. All these terms should be explained in detail for superficial readers.

Thank you for your suggestion. Thank you for your feedback regarding the clarity of the statistical data presented in Section 4.1. The notation (F_3,51=6.703,p<0.001 )refers to the results of an ANOVA test, where “F” represents the F-statistic, the two numbers in the parenthesis are the degrees of freedom for the between-groups comparison and the within-groups error respectively, and “p” signifies the p-value. In this case, (F_3,51=6.703) suggests that the variance between the groups is significantly larger than the variance within the groups. The p-value (p<0.001) indicates that the probability of observing such a large F-statistic under the null hypothesis is less than 0.1%, which is statistically significant.

Similarly, (F_3,25=15.549,p<0.001) denotes another ANOVA test's result with the respective degrees of freedom and an even smaller probability of the F-statistic being due to chance, also indicating a statistically significant difference.

For the convenience of readers not familiar with these terms, we have added a brief explanation of the statistical measures used in the study within the manuscript to ensure that our findings are accessible to all:

(P8L319)“…The ANOVA results indicated significant between-group effects, as evidenced by an F-statistic of F_3,51=6.703. This value is calculated from the ratio of the variance between the groups to the variance within the groups…The ANOVA results indicates a strong statistical significance between-group effects, as evidenced by an A p-value of less than 0.001(p<0.001). Specifically, significant main effects were observed for both the task group (F_3,25=15.549,p<0.001) and the non-task group (F_3,24=12.507,p<0.001), suggesting that the performance differences within these groups were consistent and not random. The effect sizes, measured by partial η², ranged from medium to large, which implies a substantial impact of the independent variable on the dependent variable.”

4. Likewise, the terms in table 1 such as F, P, t and P, should be explained. Similar applies to the rest of sections.

Thank you for your suggestion. We have now included brief explanations of these terms in the relevant sections of the manuscript to aid in the interpretation of our statistical results. We explained the terms of “F” by adding:

(P8L320)“… as evidenced by an F-statistic of F_3,51=6.703. This value is calculated from the ratio of the variance between the groups to the variance within the groups.”

In addition, the explanation of “P” has been supplemented:

(P8L322)“… In the context of our study, the p-value indicates the likelihood that the observed differences could have occurred by chance.”

As for the explanation of “t” we added:

(P9L332)“…The t-value indicates the significance of the difference between two groups. A larger absolute t-value signifies a more significant difference, with its sign indicating whether the task group's mean is higher or lower than the non-task group's.”

5. What does P<=0.05, P<=0.01 and P<=0.001 indicates? It needs to be mentioned in paper.

Thank you for your question and suggestion. We have now added a section that clarifies the meaning of different p-values:

(P8L323)“… A p-value of 0.05 or less (p ≤ 0.05) is considered evidence of a statistically significant difference. A p-value of 0.01 (p ≤ 0.01) or 0.001 (p≤ 0.001) represents increasingly stronger evidence against the null hypothesis of no effect or no difference.”

6. What is meant by t test? This should be explained clearly.

Thank you for your question and suggestion. In section 4.1, we conducted independent-sample t-tests to assess whether significant differences exist in trust scores between male and female participants under task and non-task scenarios. The independent-sample t-tests are statistical tests that compare the means of two independent groups to determine if the observed difference between the groups is statistically significant. In the context of our study, this was used to understand how trust towards AVs is influenced by gender under different conditions. We have supplemented the aim of utilizing t-test:

(P10L388)“… To further analyze the data, we conducted independent-sample t-tests for the task and no-task groups to evaluate whether there is a significant difference in trust between the two independent gender groups under specific conditions.”

7. Heading of Table 1 is mentioned below the table, whereas, for other tables it is above the table. Please follow journal formatting guidelines.

Thank you for your correction. We have adhered to the journal's formatting guidelines and have adjusted the heading to appear above the table, consistent with the formatting of other tables in the manuscript. We appreciate your attention to detail and guidance in maintaining the publication's standards.

8. Authors have mentioned that this study enhances the sustainability of AVs. Can you explain it how?

Thank you for your inquiry regarding the sustainable impact of our research on AVs. As highlighted in our manuscript, both in the introduction and the conclusion, we have discussed the significant role of trust in enhancing the sustainability of AVs.

In the introduction, we mentioned that “Trust may reduce negative emotions toward AI [3] and play a key role in increasing AI acceptance, making the importance of human trust in AI even more significant [4]. AI is a key technology for achieving functionality and performance of autonomous vehicles (AVs), and can improve energy efficiency by optimizing travel speed and reducing fuel consumption by reducing unnecessary acceleration and braking [5]. Therefore, AVs represent a significant step toward sustainable development [6].”(P1L27) We referenced key literature that underlines the importance of trust in reducing negative attitudes toward artificial intelligence (AI) and increasing the acceptance of AI, which is essential for the functionality and efficiency of AVs. We elaborated on how AI optimization in AVs can improving energy efficiency by optimizing travel speeds and minimizing unnecessary acceleration and braking, which contributes to sustainable development goals.

Furthermore, we mentioned that “The results of our study could serve as a reference for PI between drivers and AVs, which will increase drivers’ trust, acceptance of technology, and sustainability.”(P14L500) in the conclusion. It reiterates that the outcomes of our research provide a reference for proactive interactions (PI) between drivers and AVs. It is critical for increasing drivers' trust and acceptance of the AVs, which in turn, contributes to the sustainability of AVs by fostering wider adoption and optimal utilization.

We believe that these sections emphasize our research's contribution to AVs sustainability through enhanced trust and technology acceptance. We hope this clarification addresses your query, and we have taken care to ensure that the discussion of these points abides by the relevant literature and findings presented within our study.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thanks for addressing my comments