A Big Data System Architecture to Support the Monitoring of Paved Roads

Round 1

Reviewer 1 Report

It is not clear which are the main sensors that can be installed and/or used to monitor a road infrastructure and which could be their purpose, in relation to a specific infrastructure element (pavement damages, structures inspection for bridges, viaducts, and tunnels, restraint systems, and so on). It is known, in fact, that various kind of smart infrastructure monitoring systems are available nowadays: they are generally based on innovative sensors, systems and integrated technologies, and allow to gather and collect all the data useful to monitor and analyze the real state of each infrastructure component. In the specific case, just sensors to monitor the pavement surface have been analyzed, why? Probably, the Title of the paper has to be modified in respect to the main studies of the work.

1.      In the Abstract no examples neither calculation have been reported to demonstrate the effectiveness of the study.

2.      The Introduction section is too short, and it is not clear which are the aim and objectives of the research.

3.      Which is the importance of the Section “Monitoring Road Systems”? Is it a Literature review related just on road monitoring technology? In that case, please merge the first 2 sections together.

4.      Section 3 and 4 are too long and not so useful to the purpose of the study.

5.      The case study is not well described, and no numbers appear into the paper.

6.      The Result section is absent; have you got any numbers?

7.      The Conclusion section must be improved.

In general, Figures and Tables have low quality and are most of the time copied from previous work or software’s output.

Moderate editing of English language is required.

Author Response

It is not clear which are the main sensors that can be installed and/or used to monitor a road infrastructure and which could be their purpose, in relation to a specific infrastructure element (pavement damages, structures inspection for bridges, viaducts, and tunnels, restraint systems, and so on). It is known, in fact, that various kind of smart infrastructure monitoring systems are available nowadays: they are generally based on innovative sensors, systems and integrated technologies, and allow to gather and collect all the data useful to monitor and analyze the real state of each infrastructure component. In the specific case, just sensors to monitor the pavement surface have been analyzed, why? Probably, the Title of the paper has to be modified in respect to the main studies of the work

Answer: The title of the article has been changed to include the word 'pavement'. However, the aim of this article is not to present the physical structure installed, namely the type of sensors and the installation process. The aim of this article is to propose a big data architecture for collecting, transforming, storing, analysing and visualising sensor data, which will allow the monitoring of a section of road, and possibly even be used to monitor another section of road, or structures such as bridges, viaducts, etc. The monitoring of a road section is not limited to the condition of the pavement, but rather to the various layers that make up the road, i.e. the aim is to monitor the deformations of the structure that makes it up.

1. In the Abstract no examples neither calculation have been reported to demonstrate the effectiveness of the study.

Answer: This article, when proposing an architecture, does not justify the presentation of calculations. The only mathematical evidence that allows to demonstrate the efficiency is the reduction of the amount of data collected by the sensors, namely in one day, 24 hours, 15Gbytes are collected and through data transformations it is possible to compress them to 300Mbytes, it is a reduction of 98%, the authors do not consider it relevant to include these calculations in the abstract, but they are explained in the text of the article.

2. The Introduction section is too short, and it is not clear which are the aim and objectives of the research.

Answer: The introduction refers to the main objective of the research, which is to propose a Big Data System, i.e. an architecture to support the monitoring of road pavements..

3. Which is the importance of the Section “Monitoring Road Systems”? Is it a Literature review related just on road monitoring technology? In that case, please merge the first 2 sections together.

Answer: Sections 1 and 2 have been merged

4. Section 3 and 4 are too long and not so useful to the purpose of the study.

Answer: The authors consider section 4 to be the most important of the article, as it presents and describes the Big Data System (a big data architecture). This architecture will support the case presented in the following section.

5. The case study is not well described, and no numbers appear into the paper.

Answer: We have improved the case description, we have better explained the real-time and batch data processing component. There is a better link to the architecture presented in the previous section. We have improved the pictures and removed the prediction part, which is not yet complete but is expected to be developed.

6. The Result section is absent; have you got any numbers?

Answer: We present a results section that describes the transmission of data from the sensors to their visualisation, both in real time and in batch, and presents their transformation, namely the reduction of the data volume.

7. The Conclusion section must be improved.

Answer: We have reviewed the conclusions, namely future work in terms of predicting road health.

In general, Figures and Tables have low quality and are most of the time copied from previous work or software’s output.

Answer: We have changed the figure: number 3, where we have highlighted the components used in the case; number 5, where we have separated the real-time and batch data flows; and improved the quality of all of them. The text of the figure captions has been improved.

Moderate editing of English language is required.

Answer: The English of the paper has been completely revised.

Reviewer 2 Report

This study aims to propose and implement a Big Data System to monitor a section of a road in Portugal as a case study. In this work, it is proposed to demonstrate the whole data lifecycle, from collection to analysis, considering the different existing data processing speeds.

Please see further comments for consideration below:

In the Abstract, the reviewer suggests adding numbers that can support the results obtained from the study.

The reviewer suggests merging section 2 Monitoring Road System with the Introduction or specifying that it is a literature review. In addition, road monitoring considers several aspects of the infrastructure, not only associated with the deterioration of road pavements but also regarding traffic and speed operating conditions. Regarding the latter, I suggest adding to the long list from lines 70 to 104, a paragraph regarding sensors useful for monitoring implemented speeds and traffic flows. The reviewer suggests consulting the following recent studies:

https://doi.org/10.3390/sci4020018

https://doi.org/10.1109/TITS.2020.2988801

In Section 5, "Big Data Analytics: a Case Study," the optical sensor adopted is introduced at line 287; therefore, the reviewer suggests including a table listing and demonstrating the related and specific techniques.

Figure 7 is an extract of a table from software, so the reviewer suggests either replacing it with a new table or citing the software from which it was extracted.

Although the outline of the article is clear, there is at the same time the absence of definite and precise results. First, the Results section is absent, part of which we read in Section 5 however is not so obvious and easy to read. In fact, Subsection 5.3 seems to allude to a direct application of the Big Data management scheme for road health prediction, but no prediction model is reported.

Indeed, Section 6 Conclusions are weak and concise. Therefore, the reviewer believes the article lacks a closure worthy of the work performed by the authors. The reviewer suggests that a section of the results must be included, demonstrating the validation of the proposed Big Data management system, and then included in the Conclusions as well.

The quality of writing in this manuscript is sufficient, so moderate editing of the adopted English is recommended.

Author Response

This study aims to propose and implement a Big Data System to monitor a section of a road in Portugal as a case study. In this work, it is proposed to demonstrate the whole data lifecycle, from collection to analysis, considering the different existing data processing speeds.

Answer: That is indeed the aim of this paper.

In the Abstract, the reviewer suggests adding numbers that can support the results obtained from the study.

Answer: This paper, when proposing an architecture, does not justify the presentation of calculations. The only mathematical evidence that allows to demonstrate the efficiency is the reduction of the amount of data collected by the sensors, namely in one day, 24 hours, 15Gbytes are collected and through data transformations it is possible to compress them to 300Mbytes, it is a reduction of 98%, the authors do not consider it relevant to include these calculations in the abstract, but they are explained in the text of the paper.

The reviewer suggests merging section 2 Monitoring Road System with the Introduction or specifying that it is a literature review.

Answer: Sections 1 Introduction and 2 Monitoring Road Systems have been merged.

In addition, road monitoring considers several aspects of the infrastructure, not only associated with the deterioration of road pavements but also regarding traffic and speed operating conditions. Regarding the latter, I suggest adding to the long list from lines 70 to 104, a paragraph regarding sensors useful for monitoring implemented speeds and traffic flows. The reviewer suggests consulting the following recent studies:

https://doi.org/10.3390/sci4020018

https://doi.org/10.1109/TITS.2020.2988801

Answer: The aim of this study is not to monitor speed, but rather strain and temperature (the sensors actually measure strain and temperature). The speed is obtained by calculating the number of vehicle passes through the sensors, i.e. it is not measured directly. In our opinion, and there is a lack of studies in this area, speed does not affect the deterioration of the road structure. It is a variable that has not yet been considered relevant to deterioration. However, it is an indicator of the type of vehicle passage. Strain and temperature were therefore the main variables used. The studies you refer to are designed to monitor vehicle speed, not road health.

In Section 5, "Big Data Analytics: a Case Study," the optical sensor adopted is introduced at line 287; therefore, the reviewer suggests including a table listing and demonstrating the related and specific techniques

Answer: Throughout the text, it has been mentioned that the sensors are fibre optic. However, it should be noted that the aim of this paper is not to present the technological infrastructure for pavement monitoring, i.e. the type of sensors, their installation, sensor technologies, etc., but rather the technological structure for collecting, transforming, storing, analysing and visualising data, both in real-time and in batch mode.

Figure 7 is an extract of a table from software, so the reviewer suggests either replacing it with a new table or citing the software from which it was extracted.

Answer: The captions of figures 6 and 7 have been changed. Figure 5 has been changed to explain that there is real-time data processing and batch data processing, and the text presenting Figure 5 has been changed to make it clear that there are two flows. And the case explanation text in sections 5.1 (now 4.2) and 5.2 (now 4.3) has been revised.

Although the outline of the article is clear, there is at the same time the absence of definite and precise results. First, the Results section is absent, part of which we read in Section 5 however is not so obvious and easy to read. In fact, Subsection 5.3 seems to allude to a direct application of the Big Data management scheme for road health prediction, but no prediction model is reported.

Answer: Section 5.3 has been removed as at the time of writing no work had been done on road health prediction. A results section has been added to demonstrate the capability of the proposed architecture.

Indeed, Section 6 Conclusions are weak and concise. Therefore, the reviewer believes the article lacks a closure worthy of the work performed by the authors. The reviewer suggests that a section of the results must be included, demonstrating the validation of the proposed Big Data management system, and then included in the Conclusions as well.

Answer: As mentioned in the previous response, a results section was included to demonstrate the capability of the proposed architecture. The conclusions were mainly reviewed for future work.

The quality of writing in this manuscript is sufficient, so moderate editing of the adopted English is recommended.

Answer: The English of the paper has been completely revised.

Reviewer 3 Report

The primary innovation of the study is the proposal and implementation of a Big Data system for road monitoring. The authors suggest a more generic architecture that can adapt to the needs of road monitoring systems, whether they are cloud-based or not. This system employs Machine Learning algorithms to predict road health, enabling proactive road maintenance and potentially leading to significant cost savings.

There are several questions that could enhance the paper:

The specific tools or software used to address the problem are not explicitly mentioned in the document. The authors discuss the use of a Big Data architecture that facilitates the ingestion, processing, and availability of data for subsequent analysis and visualization. They also mention the use of databases or distributed storage systems like Hadoop's Distributed File System (HDFS), Google's Big Table, or Amazon's S3 in a Big Data context. However, they do not specify which of these tools were used in their implementation.

In terms of methodology, the authors propose a technological artifact based on a design science research methodology. They have integrated specific layers and components into a data workflow from data collection to data analysis and visualization. However, the authors could provide more details about the specific steps taken in their methodology, the rationale behind their choices, and any challenges they encountered during the implementation. For example, how did they select the section of the road in Portugal for their case study? How did they ensure the accuracy and reliability of the data collected from the sensors? How did they choose which Machine Learning algorithms to use for their predictions?

The results of the study could have been better explored. The authors have implemented a Big Data system to monitor a section of a road in Portugal, demonstrating the entire data lifecycle, from collection to analysis. However, the results of this implementation are not thoroughly discussed. For instance, the authors mention that the selected technologies worked together without integration or interoperability problems, but they do not provide a detailed analysis of the performance of these technologies or how they contributed to the overall results. Furthermore, while the authors mention the use of data visualization and some AI, they do not delve into the specific insights gained from these analyses or how they could be used for decision-making processes.

In conclusion, while the study presents a novel approach to road monitoring using Big Data and Machine Learning, it could benefit from a more detailed discussion of the results and the specific tools used in the implementation. Additionally, the authors could provide more transparency about their methodology to allow for reproducibility and further research in this area.

Author Response

The primary innovation of the study is the proposal and implementation of a Big Data system for road monitoring. The authors suggest a more generic architecture that can adapt to the needs of road monitoring systems, whether they are cloud-based or not. This system employs Machine Learning algorithms to predict road health, enabling proactive road maintenance and potentially leading to significant cost savings.

Answer: This is indeed the aim of this paper. Unfortunately, it was not possible to implement the road health prediction component in the case description.

The specific tools or software used to address the problem are not explicitly mentioned in the document. The authors discuss the use of a Big Data architecture that facilitates the ingestion, processing, and availability of data for subsequent analysis and visualization. They also mention the use of databases or distributed storage systems like Hadoop's Distributed File System (HDFS), Google's Big Table, or Amazon's S3 in a Big Data context. However, they do not specify which of these tools were used in their implementation.

Answer: The architecture presented is intended to be generic, allowing the use of multiple technologies for a given component (box) of the architecture. The aforementioned technologies have been used as examples. However, in the case presented, we refer to the technologies used.

In terms of methodology, the authors propose a technological artifact based on a design science research methodology. They have integrated specific layers and components into a data workflow from data collection to data analysis and visualization. However, the authors could provide more details about the specific steps taken in their methodology, the rationale behind their choices, and any challenges they encountered during the implementation. For example, how did they select the section of the road in Portugal for their case study? How did they ensure the accuracy and reliability of the data collected from the sensors? How did they choose which Machine Learning algorithms to use for their predictions?

Answer: The DSR methodology was chosen. Of the 6 steps of the methodology, the problem, the definition of the objectives, the state of the art, the proposal of a big data architecture and its implementation in a case to demonstrate its potential have been described in detail in this paper, as well as the search for dissemination of the results, which ends with the writing of this paper.

The results of the study could have been better explored. The authors have implemented a Big Data system to monitor a section of a road in Portugal, demonstrating the entire data lifecycle, from collection to analysis. However, the results of this implementation are not thoroughly discussed. For instance, the authors mention that the selected technologies worked together without integration or interoperability problems, but they do not provide a detailed analysis of the performance of these technologies or how they contributed to the overall results. Furthermore, while the authors mention the use of data visualization and some AI, they do not delve into the specific insights gained from these analyses or how they could be used for decision-making processes.

Answer: The case study section has been revised, we have removed section 5.3 as it does not present any results as not enough data has been collected yet to be able to predict the state of the road. We have added the results section which explains how the case fits with the proposed architecture.

In conclusion, while the study presents a novel approach to road monitoring using Big Data and Machine Learning, it could benefit from a more detailed discussion of the results and the specific tools used in the implementation. Additionally, the authors could provide more transparency about their methodology to allow for reproducibility and further research in this area.

Answer: We have addressed all the suggestions in a resolution. Namely, the rewording of the case presentation section, the addition of a results section, and a revision of the conclusions. The entire paper has been revised in terms of the English language.

Round 2

Reviewer 1 Report

I want to thank the authors for the valuable efforts in improving their work.

Please modify the following parts:

1.       Clarify what the current “Big Data System” consists of is the most important thing to do in your abstract. Which is the architecture of the system? The aim of the study and its novelty still remain unclear.

2.       The keywords are too many and are not related to what you have declared in the abstract. Data lake? Optical sensors? Information visualization? What do they mean?

3.       In combining the first two sections of the paper you have to pay attention to the references: a change in the numbering occurs.

4.       Please expand the state of the art and add new references, such as:

·         Liao, W., Zhuang, Y., Zeng, C., Deng, W., Huang, J., & Ma, H. (2020). Fiber optic sensors enabled monitoring of thermal curling of concrete pavement slab: Temperature, strain and inclination. Measurement, 165, 108203.

·         Bruno, S., Del Serrone, G., Di Mascio, P., Loprencipe, G., Ricci, E., & Moretti, L. (2021). Technical proposal for monitoring thermal and mechanical stresses of a runway pavement. Sensors, 21(20), 6797.

·         Xue, W., Wang, L., Wang, D., & Druta, C. (2014). Pavement health monitoring system based on an embedded sensing network. Journal of Materials in Civil Engineering, 26(10), 04014072.

·         Di Graziano, A., Marchetta, V., & Cafiso, S. (2020). Structural health monitoring of asphalt pavements using smart sensor networks: A comprehensive review. Journal of Traffic and Transportation Engineering (English Edition), 7(5), 639-651.

·         Barriera, M., Pouget, S., Lebental, B., & Van Rompu, J. (2020). In situ pavement monitoring: A review. Infrastructures, 5(2), 18.

·         Xue, W., Wang, D., & Wang, L. (2012). A review and perspective about pavement monitoring. Int. J. Pavement Res. Technol, 5(5), 295-302.

Minor editing of English language required

Author Response

1. Clarify what the current “Big Data System” consists of is the most important thing to do in your abstract. Which is the architecture of the system? The aim of the study and its novelty still remain unclear.

We changed the title and abstract to clarify the Big Data System is an architecture to support paved road monitoring.

2. The keywords are too many and are not related to what you have declared in the abstract. Data lake? Optical sensors? Information visualization? What do they mean?

We changed the keywords. We try to be more accurate with the paper content

3. In combining the first two sections of the paper you have to pay attention to the references: a change in the numbering occurs.

The error is corrected.

4. Please expand the state of the art and add new references

We accept and include new two references suggested:

26. Liao, W., Zhuang, Y., Zeng, C., Deng., W., Huang, J. & Ma, H. (2020). Fiber optic sensors enabled monitoring of thermal curling of concrete pavement slab: Temperature, strain and inclination, Measurement, Volume 165, https://doi.org/10.1016/j.measurement.2020.108203.

35. Barriera M, Pouget S, Lebental B, & Van Rompu J. (2020). In Situ Pavement Monitoring: A Review. Infrastructures. 5(2):18. https://doi.org/10.3390/infrastructures5020018

We did some improvements in English language 

Reviewer 2 Report

The suggestion to include numerical results in the abstract was intended to help the reader immediately identify the benefits of the proposed study. Their failure to include them and the clarification that the only mathematical efficiency is in the paper's text underscore the proposed study's weakness.

The suggestion to include a paragraph on speed monitoring sensors had been proposed to complement and argue the former Section 2 "Monitoring Road Systems."  The study mainly focuses on strain and temperature monitoring sensors, however, this was not very clear in the first version of the article.

The request to add a table listing and demonstrating the related and specific techniques of the optical sensor I do not think was very onerous to implement. The suggestion was intended to demonstrate the validity of the input data collected.

The Results section has been added and the Conclusions section expanded, but the problem still remains that the real impact of the proposed methodology cannot be quantified.

The reviewer suggests to consider the paper after a major revision, because the paper does not show that it is still ready to be published in the current version.

The quality of writing in this manuscript is sufficient, so minor editing of the adopted English is recommended.

Author Response

1. The suggestion to include numerical results in the abstract was intended to help the reader immediately identify the benefits of the proposed study. Their failure to include them and the clarification that the only mathematical efficiency is in the paper's text underscore the proposed study's weakness.

We rewrite all the text of the abstract to help the reader identify the benefits of the study.

2. The suggestion to include a paragraph on speed monitoring sensors had been proposed to complement and argue the former Section 2 "Monitoring Road Systems."  The study mainly focuses on strain and temperature monitoring sensors, however, this was not very clear in the first version of the article.

In fact, the use case uses strain and temperature optical sensors embedded in a pavement. However, the aim of this paper is to propose an architecture (a big data architecture) to guide who wants to implement a data stream (in this case: real-time and batch), from data sources to be used by persons or machines.

So we changed some sentences to be more clear the aim of this study.

We made some improvements in the English.

Reviewer 3 Report

Following the meticulous revisions, the work was deemed satisfactory: 'Accept in present form.

Author Response

We changed the paper to answer the other reviewers. 

Thank you.

Round 3

Reviewer 1 Report

Thank you.

Fine.

Reviewer 2 Report

It is good now.