A Conflict Measures-Based Extreme Value Theory Approach to Predicting Truck Collisions and Identifying High-Risk Scenes on Two-Lane Rural Highways
Round 1
Reviewer 1 Report
Interesting use of drones to study traffic conflicts, but the analyses and results are extraordinarily complex. It would help if the authors add some plain-language summary of the approach and of the conclusions - without using any abbreviations. A reader who hasn't done extremely similar studies has to constantly look up what is PET, POT, TLA1, TLA2, TLO, BEVT, UEVT, D_BR, D_BT, etc. The limitations section references a future analysis of collision severity, but how can this be computed from just conflicts? While conflict rate may predict collision rate, estimating severity seems to be a stretch.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This study uses TTC and PET as dependent variables to predict collision occurrence. It seems to be modified to improve the manuscript for consideration for publication in this journal. Please see the below I commented.
- Please insert relevant research about rural highway? How different crashes are occurred in rural highway compared to urban highway? That's the starting point of this study to reveal the merit of it.
- You are not analyzing accident but crash, make clarify the manuscript give confusion to audience.
- I am not sure the data extracted from drone video gets reliability. Because high resolution of the data should be the first priority to conduct this study. Please deal with how accuracy does the data have?
- I believe truck crash occurs during night and cause severe accidents. Could you describe why your study did not include the crashes during that time?
- using the extreme threshold of negative TTC or PED were not accepted. Usually in real data, negative TTC showed in lower speed stream condition on roadway. If you still want use the threshold, rationale and academic evidence is needed.
- How to select the threshold in CRITIC method?
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
This can be a useful paper overall, with interesting data collection and analysis methods applied toa relevant road safety risk factor- truck crashes on rural two lane roads. The application of EVT has a core logic- serious crashes are extreme and compared with the amount of driving rare events. This is a good contribution to the literature. It deserves publication, but needs significant revision before being accepted. Most importantly it seems to be a paper written for statistical modelling people not for road safety people and thus it is of limited value to the main audience for Sustainability and the special issue. Fixing this is the most critical change needed. Examples of the need for revision are listed below.
1. There are important gaps in explanation. For example: " In addition, the study of collision prediction based on traffic conflicts, despite being an effective tool for active traffic safety prevention and control, the subjectivity in choosing the threshold for identifying traffic conflicts and the lack of ability to consider different levels of conflict severity may cause bias in the prediction results, while the Extreme Value Theory (EVT) based on traffic conflicts can overcome these problems well." This requires a detailed account of HOW it can overcome this, not just a blunt claim that it can.
2. The modeling is based on trajectories for collisions. However, on rural roads many crashes are the vehicle leaving the road, not hitting another vehicle. The model does not appear to accommodate this vital crash type. However, while this is an issue its not reason for dismissing the paper. If it does please explain how. If not, then please add this as an important limitation in the discussion.
3. The abstract gives conclusions of interest to a modelling person on types of modelling but NONE of interest to a road safety person- which is basic. please add a few examples of the most important road features which predict truck crash risk. This is critical to readers getting value from the paper.
4. Most critical to the conclusions drawn- the models are validated against the total number of relevant crashes on the study highway. This is NOT a very useful validation for road safety. There is no use in predicting the total number of crashes without being able to predict the road features and thus locations at which they occur, so that these can be improved. This validation could be interesting for a modelling conference, but its of little use in road safety. Critical change- please validate these models for their ability to predict WHERE crashes occur along the road, or which best model predicts the road features involved in actual crashes? Then, some of the road features which predict truck crashes should be discussed in the discussion and noted in the abstract and conclusions- as actual road features not obscure modeling parameters (as now in the conclusion, which talks in terms of DBR, etc. being predictors- no value to road safety in this type of writing of a conclusion).
1. The manuscript is immediately difficult to follow- there are multiple abbreviations in the abstract not explained (e.g., PET, TTC, EVT...). The abstract is simply incomprehensible and other parts of the paper are difficult to follow. Abbreviations need sorting throughout the paper.
2. The paper makes many bold claims- which I expect are just wrong, and provides no references for them. For example: "Further, because of the narrow roads and high proportion of trucks on rural highways, trucks are prone to speeding, overloading, fatigue driving, and other high-risk driving behaviors, which result in frequent accidents." Why would narrow roads make trucks prone to speeding? Please provide a reference.
3. English is generally good with a few exceptions. E.g., "truck crash data is difficult to be collected."
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Paper sustainability-1856415 “A Conflict Measures-Based Extreme Value Theory Approach to Predicting Truck Collisions and Identifying High-Risk Scenes on Two-Lane Rural Highways”
Comments
This study proposes a conflict measures-based extreme value theory approach to predicting truck collisions and identifying high-risk scenes on two-lane rural highways. I think the paper fits well the scope of the journal and addresses an important subject. However, a number of revisions are required before the paper can be considered for publication. There are some weak points that have to be strengthened. Below please find more specific comments:
*The abstract should be expanded a bit. In particular, I suggest for the authors to add a sentence or two clearly highlighting contributions of this work to the state-of-the-art and outcomes from the experiments.
*The introduction section could benefit from additional statistical information to highlight the importance of the subject at hand. Please also use supporting references for the statistical information provided.
*Literature review: Please check for the most recent and relevant studies that have been published over the last years. It is essential that the literature review is up to date.
*Please provide more details regarding the input data used throughout this study. Some supporting references would be helpful to justify the data selection.
*The manuscript contains quite a lot of figures and tables. Please double check and try to provide a more detailed description of these figures and tables where appropriate to make sure that the future readers will have a reasonable understanding of what these figures and tables represent.
*Future research: This study primarily used a custom approach for predicting truck collision probabilities. In the future research, the authors should explore advanced optimization algorithms for this decision problem. Therefore, the authors should create a general discussion regarding the importance of advanced optimization algorithms (e.g., heuristics, metaheuristics) for challenging decision problems. There are many different domains where advanced optimization algorithms have been applied as solution approaches, such as online learning, scheduling, multi-objective optimization, transportation, medicine, data classification, and others (not just the decision problem addressed in this study). The authors should create a discussion that highlights the effectiveness of advanced optimization algorithms in the aforementioned domains and their potential applications for the decision problem addressed in this study. This discussion should be supported by the relevant references, including but not limited to the following:
An online-learning-based evolutionary many-objective algorithm. Information Sciences 2020, 509, pp.1-21.
An integrated optimization method for tactical-level planning in liner shipping with heterogeneous ship fleet and environmental considerations. Advanced Engineering Informatics 2021, 48, p.101299.
An augmented self-adaptive parameter control in evolutionary computation: a case study for the berth scheduling problem. Advanced Engineering Informatics 2019, 42, p.100972.
A comprehensive evaluation of weak and strong mutation mechanisms in evolutionary algorithms for truck scheduling at cross-docking terminals. IEEE Access 2018, 6, pp.65635-65650.
Ambulance routing in disaster response considering variable patient condition: NSGA-II and MOPSO algorithms. Journal of Industrial & Management Optimization 2022, 18(2), p.1035.
Such a discussion will help improving the last section of the manuscript significantly. Plus, this future research direction will allow better assessing the computational performance of the approach presented in this article.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
The authors have made serious revisions to directly address each of kye key concerns. The paper is clearer and the abstract is more relevant to road safety readers.
I have one remaining concern: the new text has several areas of poor English and is very hard to follow. For example the end to the Abstract [my notes are in square parentheses]:
the horizontal curve radius has the most critical impact on truck collision; when a truck is driving on two-lane rural highways with a horizontal curve radius of 227m or less, the frequency and probability of collision will be higher, but it can install deceleration devices and central separation belts to slow down the collision [1. poor English- suggest change say xxx can be installed to reduce risk....2. Not sure what is meant by belts- does this mean guardrail barrier?] Following that [Second?] is the driving behavior risk, when it is less than 43 points [these points mean nothing to the reader at this stage- suggest leave the points out here and just refer to the high risk behaviour], indicating that the driving behavior has a high risk, which can be installed [do you mean addressed?] through the early warning device to remind the 30 driver to control the speed and route of the truck to avoid [ there is no certainty to such solutions- why not suggest speed cameras as well?] dangerous driving. This study extends the evaluation method of truck collisions on two-lane rural highways from univariate to bivariate and provides a basis for the design of two-lane rural highways and the development of real-time dynamic warning systems [and enforcement] for trucks, which will help prevent and control truck collisions and alleviate safety problems on two-lane rural highways.
Other added sections also need refinement, and my suggested addition of speed cameras will need to be added in the body of the text. There is excellent evidence that speed cameras work - there is an older but sound Cochrane library review and a very recent paper in Sustainability: Job, R.F.S. (2022). Evaluations of Speed Camera Interventions Can Deliver a Wide Range of Outcomes: Causes and Policy Implications. Sustainability, 14(3), 1765. https://doi.org/10.3390/su14031765.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
The authors took seriously my previous comments and made the required revisions in the manuscript. The quality and presentation of the manuscript have been improved. Therefore, I recommend acceptance.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
