the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Nov 2023
Evaluating the use of smart sensors in ground-based monitoring of landslide movement with laboratory experiments
Abstract. Boulders and cobbles embedded on the body of landslides are carried downstream under the action of gravity, and the study of their transport can give important insight on their dynamics and hence the related hazard. The study examines the reliability of smart sensors to track movements of a cobble and discern between intensity and mode of movement in laboratory experiments. A tag equipped with accelerometer, gyroscope, and magnetometer sensors was installed inside a cobble. The experiments consisted of letting the cobble fall on an inclined plane. By tilting the inclined plane at different angles, different modes of movement such as rolling, bouncing, or sliding were generated. Sliding was generated by embedding the cobble within a thin layer of sand. The position of the cobble travelling down the slope was derived from camera videos. Raw sensor data allowed detection of movement and separation of two modes of movement, namely rolling, and sliding. Additionally, reliable values for the position, velocity, and acceleration were determined by feeding a Kalman filter with smart sensor measurements and camera-based positions. Furthermore, by testing LoRaWAN wireless transmission through sand, the study showed that the signal strength tended to decrease for thicker sand layers. These findings confirm the potential to use these sensors to improve early warning systems and further studies are in progress to assess practicalities of their use in field settings.
- Preprint
(2057 KB) - Metadata XML
-
Supplement
(1453 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2023-2596', Anonymous Referee #1, 08 Dec 2023
Recommendation: reconsidered after major revisions.
Comments to the Author
Ms. Preprint egusphere-2023-2596
Title: Evaluating the use of smart sensors in ground-based monitoring of landslide movement with laboratory experiments
Authors: Alessandro Sgarabotto, Irene Manzella, Kyle Roskilly, Miles J. Clark, Georgie L. Bennett, Chunbo Luo, and Aldina M. A. Franco
Overview and general recommendation:
The paper "Evaluating the use of smart sensors in ground-based monitoring of landslide movement with laboratory experiments" investigates the effectiveness of smart sensors for tracking the movement of cobbles in landslide scenarios. It focuses on the ability of these sensors to discern the intensity and mode of movement through laboratory experiments. The study involves embedding sensors in a cobble and observing its movement on different inclines and surfaces, using a combination of accelerometers, gyroscopes, and magnetometers. The data collected is further analysed using a Kalman filter and camera-based positions. Additionally, the paper explores the impact of sand layers on wireless signal transmission. The findings suggest potential applications for improving early warning systems in landslide monitoring.
Thus, the general recommendation is reconsidered after major revisions.Major Comments The table-top experiment in your manuscript presents some interesting ideas, notably the investigation of LoRaWAN signal strength propagation through sand. However, the approach to comparing results with classical equations of motion (EOMs) comes across as unnecessarily grandiose. Additionally, the pendulum analysis comes utterly short, as a single harmonic oscillator is deemed as model comparison, while the data clearly shows a beating effect. A significant oversight is the lack of discussion on scalability, especially pertinent given the context of smart sensors in landslide applications. In conclusion, substantial revisions are needed, particularly in data analysis for the pendulum section and a more grounded portrayal of the comparison to classical EOMs.
Specific comments :
Figure annotations: While there are strong opinions around for this one, feel free to stick with yours: mine would be “x (m) “ instead of “x[m]”. The [] brackets are in physics the unit operator used in dimensional analysis. Hence [x] = m, [mag] = G, and so on.
L 94-97: Leave out the obvious. The committed reader will find out about the structure of the paper.
L115-116: distinguish between a 9DOF IMU and a 3DOF accelerometer. Yes, the web page states the ST LIS2DH tracks motion, but it actually only tracks accelerations. Be precise.
L120 vs L 115: The introduction of the smart sensor as mini-GPS Tracker, and only stating later, that GPS is switched off, is confusing. Sensor experts wonder immediately, what use a GPS tracker is for an indoor experiment.
L115: Please add an image of the stripped down, commercially available Miromico to Figure 2Figure 4: Please add RSSI and SNR explanations for only image-readers. Increase resolution of images, they look pixelated.
Figure 5: Increase dpi.
Figure 6: Increase dpi. Legend with capital letters in Test 1, etc.
Equation 1: I have seen this before, the equation and the ascribing to a specific publication dated from this century. With all due respect for your work, as a respectful note to a fellow scientists, it's crucial to maintain academic integrity and historical accuracy in scientific publications. Attributing well-established scientific principles, such as the equation representing total mechanical energy, to contemporary authors overlooks the foundational work of early physicists like Isaac Newton and Leonhard Euler. Yes, we are talking the big names here. This not only misrepresents the origins of these fundamental concepts but also undermines the rich history of scientific discovery. It's important for all scientific literature to accurately reflect the development of these theories over time and give credit where it's historically due.
Equation 2, 3 and 4: A rigid body travelling down a tilted plane and its associated equations of motion are a common high-school problem. While it is perfectly valid to restate these equations here – again, not really invented or proposed the first time by any individual in the 21st century - as they are applied in a specific contexts, such as the kinematics of a rigid body traveling down a tilted plane with varying curvature, a lengthy derivation is unnecessary.
Pendulum Oscillation: The derivation of the pendulum oscillations equations starting from line 465 in the document is a detailed and lengthy process. It appears to be a standard approach found in textbooks on classical mechanics. While thorough, the extended focus on these foundational equations might be considered excessive for a journal publication, where readers typically expect more concise presentations of well-established theories
Fig. 12 and 13: The significant discrepancies observed in the comparison with the classical equations of motion are a critical issue. The noted beating in the signal is particularly concerning and suggests a fundamental problem that needs to be addressed. It's imperative for the authors to rigorously investigate the origins of these discrepancies. This deeper analysis is crucial for assessing the validity of using the sensor technology in the context of these standard motion equations and for ensuring the accuracy and reliability of the study's findings.
General remark on “modelling”: To clarify, comparing experimental results with classical equations of motion should not be termed "modelling" in a strict sense. This comparison is more accurately described as a validation or verification step, where the experimental data is tested against established theoretical principles. "Modelling" typically refers to the development of new theoretical frameworks or the application of existing theories to simulate complex systems, which is distinct from merely comparing data with standard motion equations. This distinction is important for accurately conveying the nature of the scientific work being done.
Citation: https://doi.org/10.5194/egusphere-2023-2596-RC1 - AC1: 'Reply on RC1', Alessandro Sgarabotto, 21 Feb 2024
-
RC2: 'Comment on egusphere-2023-2596', Anonymous Referee #2, 10 Jan 2024
Referee’s comments on the manuscript “Evaluating the use of smart sensors in ground-based monitoring of landslide movement with laboratory experiments”
The submitted manuscript aims to derive long-term rock movements and landslides through in-situ mounted sensors. Various laboratory experiments were carried out for this purpose.
The long-term goal of in situ-based landslide detection is socially relevant, and its achievement is scientifically desirable. However, how this manuscript is presented is not (yet) relevant for publication in its present form. The applied methods are per se to be embraced (smart sensors and the fusion of their gyroscope data with those of the magnetometer and accelerometer), but the validation presented (e.g. Fig. 11-13) are too far from the compared mental model, namely the center of mass of a rigid body (=single point). This raises both the questions of whether (a) the measured data and its fusion chain are not consistent or (b) whether the applied model is suited for the used complex cobble. Usually, favorable laboratory conditions are chosen. This has not been the case here, as the damping through the cotton pads around the sensors (which should protect them) creates an environment that is not comparable with the rigid body model.
So, in my opinion, the manuscript and its current results are not yet usable for the NHESS readership. On the one hand, the same results should be compared with a new comparison model. Instead of the cobble motions, the different movements of the sensors on its cotton pad are recorded. Therefore, the manuscript better fits in a signal processing journal, especially as the embedding in the natural hazards literature is not without some reservations (see specific comments). On the other hand, the experiments could be repeated with a rigid sensor attachment to the rock. Then, the simple single-point model can be applied, and the outcome will align with the NHESS scope.
Specific comments:
L13:
process type instead of hazardL28:
Sim et al. 2022 are talking rather about risks, not the hazardsL34:
https://doi.org/10.1038/s43247-023-00909-z is also an important source here.L65/557
Citing EGU-General assembly abstracts is not the best practice.L74-76:
The data of pressure sensors were barely used in these studies.L78:
The cited sources here make no sense: They were all published before the smart sensors publications (L74-76) and the mentioned 3D rockfall modeling approach was already validated with the publication by Dorren (2005).L94-97:
This section is unnecessary. If you use meaningful sub-titles (as you do), the reader can easily navigate without this section. However, if you want to add this section, add at least the purpose and content of section 4.L102:
The squared, inclined board was hinged to the rectangular, horizontal board along its shorter side.L108:
Synchronized recordings to what?L155:
How many different training images?
Second, in each training image, …L162:
Which one is the suitable built-in function in OpenCV?L165:
Missing “.” after “occurred”L228-234:
Belongs rather to the methods section.L248-250:
Belongs rather to the discussions section.L280, Fig. 5:
Better comparability if the x-axis comprises in all subplots (a-f) the same interval (e.g., 0.0 s – 2.8 s). The same would also be nice for the y-axis in the corresponding subplots (a and d; b and e; c and f).L306-310:
A procedure description belongs to the methods section.L320:
“We speculate” belongs never in a result section.L324:
Wrong sub figures mentioned (7 c,d instead of 7b,c)L365:
From where do you have the linear velocity? Pronounce it (again) that this is from the video footage analysis.Citation: https://doi.org/10.5194/egusphere-2023-2596-RC2 - AC2: 'Reply on RC2', Alessandro Sgarabotto, 21 Feb 2024
Status: closed
-
RC1: 'Comment on egusphere-2023-2596', Anonymous Referee #1, 08 Dec 2023
Recommendation: reconsidered after major revisions.
Comments to the Author
Ms. Preprint egusphere-2023-2596
Title: Evaluating the use of smart sensors in ground-based monitoring of landslide movement with laboratory experiments
Authors: Alessandro Sgarabotto, Irene Manzella, Kyle Roskilly, Miles J. Clark, Georgie L. Bennett, Chunbo Luo, and Aldina M. A. Franco
Overview and general recommendation:
The paper "Evaluating the use of smart sensors in ground-based monitoring of landslide movement with laboratory experiments" investigates the effectiveness of smart sensors for tracking the movement of cobbles in landslide scenarios. It focuses on the ability of these sensors to discern the intensity and mode of movement through laboratory experiments. The study involves embedding sensors in a cobble and observing its movement on different inclines and surfaces, using a combination of accelerometers, gyroscopes, and magnetometers. The data collected is further analysed using a Kalman filter and camera-based positions. Additionally, the paper explores the impact of sand layers on wireless signal transmission. The findings suggest potential applications for improving early warning systems in landslide monitoring.
Thus, the general recommendation is reconsidered after major revisions.Major Comments The table-top experiment in your manuscript presents some interesting ideas, notably the investigation of LoRaWAN signal strength propagation through sand. However, the approach to comparing results with classical equations of motion (EOMs) comes across as unnecessarily grandiose. Additionally, the pendulum analysis comes utterly short, as a single harmonic oscillator is deemed as model comparison, while the data clearly shows a beating effect. A significant oversight is the lack of discussion on scalability, especially pertinent given the context of smart sensors in landslide applications. In conclusion, substantial revisions are needed, particularly in data analysis for the pendulum section and a more grounded portrayal of the comparison to classical EOMs.
Specific comments :
Figure annotations: While there are strong opinions around for this one, feel free to stick with yours: mine would be “x (m) “ instead of “x[m]”. The [] brackets are in physics the unit operator used in dimensional analysis. Hence [x] = m, [mag] = G, and so on.
L 94-97: Leave out the obvious. The committed reader will find out about the structure of the paper.
L115-116: distinguish between a 9DOF IMU and a 3DOF accelerometer. Yes, the web page states the ST LIS2DH tracks motion, but it actually only tracks accelerations. Be precise.
L120 vs L 115: The introduction of the smart sensor as mini-GPS Tracker, and only stating later, that GPS is switched off, is confusing. Sensor experts wonder immediately, what use a GPS tracker is for an indoor experiment.
L115: Please add an image of the stripped down, commercially available Miromico to Figure 2Figure 4: Please add RSSI and SNR explanations for only image-readers. Increase resolution of images, they look pixelated.
Figure 5: Increase dpi.
Figure 6: Increase dpi. Legend with capital letters in Test 1, etc.
Equation 1: I have seen this before, the equation and the ascribing to a specific publication dated from this century. With all due respect for your work, as a respectful note to a fellow scientists, it's crucial to maintain academic integrity and historical accuracy in scientific publications. Attributing well-established scientific principles, such as the equation representing total mechanical energy, to contemporary authors overlooks the foundational work of early physicists like Isaac Newton and Leonhard Euler. Yes, we are talking the big names here. This not only misrepresents the origins of these fundamental concepts but also undermines the rich history of scientific discovery. It's important for all scientific literature to accurately reflect the development of these theories over time and give credit where it's historically due.
Equation 2, 3 and 4: A rigid body travelling down a tilted plane and its associated equations of motion are a common high-school problem. While it is perfectly valid to restate these equations here – again, not really invented or proposed the first time by any individual in the 21st century - as they are applied in a specific contexts, such as the kinematics of a rigid body traveling down a tilted plane with varying curvature, a lengthy derivation is unnecessary.
Pendulum Oscillation: The derivation of the pendulum oscillations equations starting from line 465 in the document is a detailed and lengthy process. It appears to be a standard approach found in textbooks on classical mechanics. While thorough, the extended focus on these foundational equations might be considered excessive for a journal publication, where readers typically expect more concise presentations of well-established theories
Fig. 12 and 13: The significant discrepancies observed in the comparison with the classical equations of motion are a critical issue. The noted beating in the signal is particularly concerning and suggests a fundamental problem that needs to be addressed. It's imperative for the authors to rigorously investigate the origins of these discrepancies. This deeper analysis is crucial for assessing the validity of using the sensor technology in the context of these standard motion equations and for ensuring the accuracy and reliability of the study's findings.
General remark on “modelling”: To clarify, comparing experimental results with classical equations of motion should not be termed "modelling" in a strict sense. This comparison is more accurately described as a validation or verification step, where the experimental data is tested against established theoretical principles. "Modelling" typically refers to the development of new theoretical frameworks or the application of existing theories to simulate complex systems, which is distinct from merely comparing data with standard motion equations. This distinction is important for accurately conveying the nature of the scientific work being done.
Citation: https://doi.org/10.5194/egusphere-2023-2596-RC1 - AC1: 'Reply on RC1', Alessandro Sgarabotto, 21 Feb 2024
-
RC2: 'Comment on egusphere-2023-2596', Anonymous Referee #2, 10 Jan 2024
Referee’s comments on the manuscript “Evaluating the use of smart sensors in ground-based monitoring of landslide movement with laboratory experiments”
The submitted manuscript aims to derive long-term rock movements and landslides through in-situ mounted sensors. Various laboratory experiments were carried out for this purpose.
The long-term goal of in situ-based landslide detection is socially relevant, and its achievement is scientifically desirable. However, how this manuscript is presented is not (yet) relevant for publication in its present form. The applied methods are per se to be embraced (smart sensors and the fusion of their gyroscope data with those of the magnetometer and accelerometer), but the validation presented (e.g. Fig. 11-13) are too far from the compared mental model, namely the center of mass of a rigid body (=single point). This raises both the questions of whether (a) the measured data and its fusion chain are not consistent or (b) whether the applied model is suited for the used complex cobble. Usually, favorable laboratory conditions are chosen. This has not been the case here, as the damping through the cotton pads around the sensors (which should protect them) creates an environment that is not comparable with the rigid body model.
So, in my opinion, the manuscript and its current results are not yet usable for the NHESS readership. On the one hand, the same results should be compared with a new comparison model. Instead of the cobble motions, the different movements of the sensors on its cotton pad are recorded. Therefore, the manuscript better fits in a signal processing journal, especially as the embedding in the natural hazards literature is not without some reservations (see specific comments). On the other hand, the experiments could be repeated with a rigid sensor attachment to the rock. Then, the simple single-point model can be applied, and the outcome will align with the NHESS scope.
Specific comments:
L13:
process type instead of hazardL28:
Sim et al. 2022 are talking rather about risks, not the hazardsL34:
https://doi.org/10.1038/s43247-023-00909-z is also an important source here.L65/557
Citing EGU-General assembly abstracts is not the best practice.L74-76:
The data of pressure sensors were barely used in these studies.L78:
The cited sources here make no sense: They were all published before the smart sensors publications (L74-76) and the mentioned 3D rockfall modeling approach was already validated with the publication by Dorren (2005).L94-97:
This section is unnecessary. If you use meaningful sub-titles (as you do), the reader can easily navigate without this section. However, if you want to add this section, add at least the purpose and content of section 4.L102:
The squared, inclined board was hinged to the rectangular, horizontal board along its shorter side.L108:
Synchronized recordings to what?L155:
How many different training images?
Second, in each training image, …L162:
Which one is the suitable built-in function in OpenCV?L165:
Missing “.” after “occurred”L228-234:
Belongs rather to the methods section.L248-250:
Belongs rather to the discussions section.L280, Fig. 5:
Better comparability if the x-axis comprises in all subplots (a-f) the same interval (e.g., 0.0 s – 2.8 s). The same would also be nice for the y-axis in the corresponding subplots (a and d; b and e; c and f).L306-310:
A procedure description belongs to the methods section.L320:
“We speculate” belongs never in a result section.L324:
Wrong sub figures mentioned (7 c,d instead of 7b,c)L365:
From where do you have the linear velocity? Pronounce it (again) that this is from the video footage analysis.Citation: https://doi.org/10.5194/egusphere-2023-2596-RC2 - AC2: 'Reply on RC2', Alessandro Sgarabotto, 21 Feb 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 303 | 143 | 21 | 467 | 32 | 11 | 14 |
- HTML: 303
- PDF: 143
- XML: 21
- Total: 467
- Supplement: 32
- BibTeX: 11
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
