Prototype Network for Few-Shot Hazard Assessment of Vehicle Lane Changes in Risk Field

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The main question addressed by the research is to discuss risk of changing lanes in road network.

It is very important problem, especially connected with autonomous vehicles and algorithms which must calculate driving path during driving process.

The problem escalates when we are talking about a few changes in short time. So the original part of the paper is Author's approach and methods to among others determine risk of the situation.

Literature review is enough in my opinion.

Important is specially chapter 4, where Authors proposed prototype network-based risk posture identification model for lane change.

Authors also defined limitation of the paper. In my opinion current version of the paper is ready to publish. However during future it would be great to extend research and include some of the currently excluded scenarios (lines 520-528).

Author Response

Dear Editors and Reviewers,

Thank you for taking the time to review the paper”applsci-3035911: Prototype Network for Few-Shot Vehicle Change Lanes Hazard Assessment under Risk Field”. Those comments are all valuable and very helpful for revising and improving our papers, as well as the critical guiding significance to our research. We have carefully reviewed and adopted the reviewers’ suggestions. The revised portion is marked in red for the new submission. We sincerely hope this manuscript could be published in Applied Sciences. Again, thank you very much for your help, and I am looking forward to hearing from you soon.

 

Response to Reviewers

 

Reviewer#1:

 

Comments:

The main question addressed by the research is to discuss risk of changing lanes in road network.

It is very important problem, especially connected with autonomous vehicles and algorithms which must calculate driving path during driving process.

The problem escalates when we are talking about a few changes in short time. So the original part of the paper is Author's approach and methods to among others determine risk of the situation.

Literature review is enough in my opinion.

Important is specially chapter 4, where Authors proposed prototype network-based risk posture identification model for lane change.

Authors also defined limitation of the paper. In my opinion current version of the paper is ready to publish. However during future it would be great to extend research and include some of the currently excluded scenarios (lines 520-528).

Response:

Thank you for your valuable feedback and constructive comments on our paper. We appreciate your acknowledgment of the main question addressed in our research regarding the risk of changing lanes in a road network, especially in the context of autonomous vehicles and algorithmic driving paths calculations.

We sincerely appreciate your recognition of the significance of the problem, particularly in situations involving multiple lane changes within a short timeframe. The original contribution of our paper lies in the author’s approach and methods, including the determination of risk in such circumstances.

Furthermore, we acknowledge and have outlined the limitations of our study. We are grateful for your assessment that the current version of the paper is suitable for publication. We share your view that future research endeavors could expand on the excluded scenarios (lines 520-528) to enhance the scope and depth of the study.

Once again, we appreciate your time and effort in providing feedback on our manuscript. Thank you for your support and guidance throughout this review process.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The idea of the article is really interesting and valuable in the field of vehicle motion research. Despite this, the manuscript should be improved because the main aspects stated in the abstract, in the title and in the introduction are not clear due to a confusing exposition of the concepts.

Some of the suggestions are listed as follows:

The acronyms must be introduced before using them, e.g., JDL-DBN, even if they are well-knwon for expert in the field, not all the possible readers will be expert.

A scheme detailing the objectives of research could improve the awareness. The objectives are clearly presented but in the bullet point lists they are too verbous and the main focus becomes not intelligible.

L. 107-111 tortuous sentence. Please reshape it making it clear.

Please list in a bullet point list the parameters introduced in Equations 1 and 2.

In Figure 1, which is the unit of measure of the horizontal coordinate?

L.134 should refer to Figure 1 (a) not to Figure 2(a) as in the text, revise it.

L.170-172 revise the use of brackets. Sentences are not clear.

Please provide a comment about the 0.1 s time interval. Do you think it is sufficient for the purpose of research: Would you use shorter time span?

Figure 3. The name of the axes should contain the entire name of the variables.

L. 247-263 are crucial for the manuscript, but they are not as clear as they should be, since they introduce key concepts for the analysis.

Before Table 2 a description of how the values in table are obtained and how they are correlate with the risk factor.

All the equations should detail all the parameters used with a list, to make their explanation more effective.

Figure 6 contains important graphs that need to be highlighted and made bigger.

 

Comments on the Quality of English Language

The Quality of English must be improved, more on the sentences construction then on words. It is often difficult to follow the flow of the sentences because they are tortous and not clear.

 

Author Response

Dear Editors and Reviewers,

Thank you for taking the time to review the paper”applsci-3035911: Prototype Network for Few-Shot Vehicle Change Lanes Hazard Assessment under Risk Field”. Those comments are all valuable and very helpful for revising and improving our papers, as well as the critical guiding significance to our research. We have carefully reviewed and adopted the reviewers’ suggestions. The revised portion is marked in red for the new submission. We sincerely hope this manuscript could be published in Applied Sciences. Again, thank you very much for your help, and I am looking forward to hearing from you soon.

 

Response to Reviewers

 

Reviewer#2:

 

Comments 1:

The acronyms must be introduced before using them, e.g., JDL-DBN, even if they are well-known for expert in the field, not all the possible readers will be expert.

Response 1:

Thanks for your review. We are sorry for our neglect. We have updated the introduction as requested and added yellow highlighting in the document to emphasize it.

Author action:

“The Data Fusion Model maintained by the Joint Directors of Laboratories (JDL) Data Fusion Group is the most widely-used method for categorizing data fusion-related functions. According to Polychronopoulos et al., the JDL was applied to the automotive safety-assisted driving. Based on the technical characteristics of the vehicle-mounted multi-sensors, the JDL model is improved, including structure and composition. Furthermore, the Pro Fusion2 model for situational awareness of driving safety is proposed [1].

With the real-time collision prediction results of the random forest classifier as an example, the probability required by the Dynamic Bayesian Networks (DBN) model was estimated. According to the findings, a well-calibrated collision prediction classifier can provide complementary cues to already developed interactive perception-motion models and enhance real-time risk assessment for autonomous vehicles.”

 

Comments 2:

A scheme detailing the objectives of research could improve the awareness. The objectives are clearly presented but in the bullet point lists they are too verbous and the main focus becomes not intelligible.

Response 2:

Thank you for your valuable feedback. We will revise the bullet point lists to ensure clarity and conciseness in articulating the research objectives. We will focus on simplifying the language and directly stating the intent of each objective.

Author action:

“Utilizing features extracted from both the self-vehicle content and the surrounding vehicle-related content, we aim to refine the representation and discriminability of risk posture discriminating features by sharing the vehicle risk threshold space with the vehicle risk field. This enhancement will be achieved through univariate or multivariate combinations, enhancing the clarity of the risk posture assessment index.}

We propose a risk classification mechanism that integrates the vehicle risk field and prototype network. This mechanism aims to match attributes from various risk fields to identify more crucial attribute features for input data of the prototype network. Extensive experiments conducted with few-shot data demonstrate that our proposed algorithm effectively enhances the accuracy of vehicle risk identification.

We employ a prototype network to predict vehicle risks and introduce a novel vehicle risk assessment method with few-shot capabilities.”

 

Comments 3:

1. 107-111 tortuous sentence. Please reshape it making it clear.

Response 3:

Thanks for your careful suggestion, which can improve the quality of our manuscript. We have reshaped the manuscripts and added yellow highlighting in the document to emphasize it.

Author action:

As lane-changing behavior is continuous, identifying lane-changing status requires selecting driving trajectory data from specific time periods as samples. To locate lane change data, this study initially identified the vehicle’s lane change points by observing changes in the vehicle’s lane ID or the moment the vehicle’s center crossed a lane line. Using a dynamic time window, this study specified the phase of lane change execution when both the vehicle’s lane ID and lateral position changed continuously. Equation \ref{deqn_ex1} demonstrates that vehicles with continuous changes in lateral displacement and the number of continuously changing data points is greater than 100 are positioned as lane change vehicles. Further data extraction is performed in accordance with the three phases of lane change perception phase, lane change execution phase, and lane change adjustment phase.

 

Comments 4:

Please list in a bullet point list the parameters introduced in Equations 1 and 2.

Response 4:

Thanks for your careful suggestion, we have added the explain of these independent variables are chosen and added yellow highlighting in the document to emphasize it.

 

Comments 5:

In Figure 1, which is the unit of measure of the horizontal coordinate?

Response 5:

Thank you for your valuable feedback. In the figure, the unit of the horizontal coordinate is time (seconds).

 

Comments 6:

L.134 should refer to Figure 1 (a) not to Figure 2(a) as in the text, revise it.

Response 6:

Apologies for the oversight, we mistakenly labeled Figure 2 (a) as Figure 1 (a). We will promptly correct this error and update the document with the correct figure number. Thank you for bringing it to our attention, and we will ensure to rectify this mistake in the final version. (See L.137）

 

Comments 7:

L.170-172 revise the use of brackets. Sentences are not clear.

Response 7:

Thank you for your feedback. To improve clarity, I have revised the sentences and removed the brackets. Here is the revised text: 

“The ego vehicle SV is faster than the vehicle PV in the current lane ahead. There is sufficient space ahead of the target lane for a lane change, meeting the lane change requirement of \begin{math}g_{T-2}\end{math}. The distance between the main vehicle LGV and the vehicle behind the target lane \begin{math}g_{T+2}\end{math} exceeds the minimum safe following distance.” (See L.172-175）

 

Comments 8:

Please provide a comment about the 0.1 s time interval. Do you think it is sufficient for the purpose of research: Would you use shorter time span?

Response 8:

The 0.1-second time interval used for outputting the motion information of vehicles in the NGSIM data set is relatively high frequency and provides detailed temporal resolution. Whether this interval is sufficient for research purposes depends on the specific objectives of the study. In certain cases, a 0.1-second interval may be suitable for capturing relevant dynamics in traffic behavior and vehicle interactions.

In our study, the motion information of vehicles obtained at a 0.1-second interval is sufficient to support lane change research, as further data processing is required.

 

Comments 9:

Figure 3. The name of the axes should contain the entire name of the variables.

Response 9:

Thank you for your feedback. We will ensure that the complete names of the variables are included on the axes of the chart to improve clarity and readability.

 

 

Comments 10:

247-263 are crucial for the manuscript, but they are not as clear as they should be, since they introduce key concepts for the analysis.

Response 10:

Thank you for your feedback. We acknowledge the importance of lines 247-263 in our manuscript and recognize the need for clarity in introducing key concepts for the analysis. We will carefully revise this section to ensure that the concepts are conveyed more effectively. Your comments are invaluable in improving the quality of our manuscript. Here is the revised text:

“TTC (Time to Collision) \cite{ref-journal27} can indicate the relative speed and distance between the lane-changing vehicle and the target vehicle, a metric often used in analyzing longitudinal collision safety in vehicles. According to Olsen et al., \cite{ref-journal28}, the TTC tends to infinity when the speed of the leading vehicle exceeds that of the following vehicle in the same lane. Both vehicles continue to drive in their current conditions with minimal risk of collision. At this point, the TTC value can be consistently set at 30 seconds. This study utilizes Olsen’s method to calculate the vehicle’s TTC value during driving. If the leading vehicle in the same lane is moving faster than the following vehicle, the TTC value is set at 30 seconds. When the speed of the vehicle ahead in the same lane is inferior to that of the vehicle behind, the relative distance and relative speed between the two vehicles are used to ascertain the TTC value. Furthermore, an analysis of the TTC value is conducted if the quotient exceeds 30 seconds. Minderhoud et al., \cite{ref-journal29} proposed classifying a Time to Collision (TTC) within 3 seconds as high risk and exceeding 5.5 seconds as low risk. Ma Yanli et al., proposed using the median and the 85th percentile to categorize the risk thresholds, defining a TTC under 2.1 seconds as low risk and over 4 seconds as high risk \cite{ref-journal30}. In this study, a Time to Collision (TTC) below 3 seconds is categorized as a potential risk, and a TTC exceeding 4.4 seconds is classified as high risk. Hence, a reciprocal of less than 0.333 seconds for TTC indicates low risk, while over 0.2273 seconds signifies high risk \cite{ref-journal31}.”(See L.255-272)

 

Comments 11:

Before Table 2 a description of how the values in table are obtained and how they are correlate with the risk factor.

Response 11:

Thank you for your question. As you mentioned, Table 2 already provides a detailed introduction to the threshold division of each variable and its corresponding relationship with risk. Please refer to L. 247-291 of the article for details.

 

Comments 12:

All the equations should detail all the parameters used with a list, to make their explanation more effective.

Response 12:

Thank you for your review. We appreciate your suggestion regarding the equations. We will ensure that all equations in the manuscript are accompanied by a detailed list of parameters used, enhancing the clarity and effectiveness of their explanation.

 

Comments 13:

Figure 6 contains important graphs that need to be highlighted and made bigger.

Response 13:

Thank you for your valuable feedback. We acknowledge the importance of the graphs in Figure 6 and will work on highlighting and increasing their size to improve visibility and emphasis.

 

 

 

 

Author Response File: Author Response.docx