Behavioral Fingerprinting: Acceleration Sensors for Identifying Changes in Livestock Health
Round 1
Reviewer 1 Report
The manuscript is organized and written well and the reviewed topic is of importance to the relevant research communities.
However, since the submission is a review article, it should be as comprehensive as possible in covering the existing related work. I believe some other recent works on animal behavior classification using wearable accelerometers such as the following ones can also be cited and briefly discussed as required.
D. Pavlovic et al., “Classification of cattle behaviours using neck-mounted accelerometer-equipped collars and convolutional neural networks,” Sensors, vol. 21, 2021.
J. W. Kamminga et al., “Robust sensor-orientation-independent feature selection for animal activity recognition on collar tags,” Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, vol. 2, 2018.
L. Nobrega et al., “Assessing sheep behavior through low-power microcontrollers in smart agriculture scenarios,” Computers and Electronics in Agriculture, vol. 173, 2020.
J. P. Dominguez-Morales et al., “Wildlife monitoring on the edge: A performance evaluation of embedded neural networks on microcontrollers for animal behavior classification,” Sensors, vol. 21, 2021.
R. Arablouei et al., “In-situ classification of cattle behavior using accelerometry data,” Computers and Electronics in Agriculture, vol. 183, 2021.
R. Arablouei et al., “Animal behavior classification via deep learning on embedded systems,” arXiv:2111.12295, 2022.
L. Wang et al., “Animal behavior classification via accelerometry data and recurrent neural networks,” arXiv: 2111.12843, 2022.
J. Cabezas et al., “Analysis of accelerometer and GPS data for cattle behaviour identification and anomalous events detection,” Entropy, vol. 24, 2022.
J. Brennan et al., “Classifying season long livestock grazing behavior with the use of a low-cost GPS and accelerometer,” Computers and Electronics in Agriculture, vol. 181, 2021.
I would also encourage the authors to consider the following recent review article on animal behavior classification using accelerometer data and pay attention to the papers cited therein to make sure they do not miss any important previous work that may deserve to be included in the review.
L. Riaboff et al., “Predicting livestock behaviour using accelerometers: A systematic review of processing techniques for ruminant behaviour prediction from raw accelerometer data,” Computers and Electronics in Agriculture, vol. 192, 2022.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The article entitled ‘Variation in acceleration sensors for identifying changes in animal health’ is intended to provide an overview of the state of the art in the use of accelerometers for monitoring ruminant activity.
This article is not a systematic review, as authors do not present their research and selection methodologies.
The authors focus on the use of accelerometers without specifying the motivation for this choice. Other types of sensors are used on the market or in academic works. Moreover, the term accelerometer is vague, it should be specified whether it is inertial sensors and the number of measurement axes.
Table 1 is very interesting but does not allow a sufficient comparison between the system. In particular, the number of subjects studied and the duration of the measures are lacking. Moreover, the characteristics of the accelerometers are not specified in terms of frequency, resolution and type of post processing of the data. Finally, the use of r or r2 coefficients does not allow evaluating precision and is very controversial from the statistical point of view to make comparisons of the methods.
In the last part of the article, the authors make recommendations on the use of the accelerometer. I find the guidelines very subjective. The object of the article is to do a review, whereas this end of the article is more in a perspective view style. The purpose of the article should be clarified here.
In conclusion, this article presents a relevant subject. However the form presented here does not allow a complete answer to have a state of the art.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Manuscript can be edited for language and better comprehension. What is the difference between physical behaviour and expressive behaviour and how expressive behaviour subtleties can be measured and quantified by accelerometer based data?
Depth is missing in the manuscript as it reads like a mere summary. Critical analysis and insightful information are lacking.
Being a review paper, recent relevant papers are not fully cited.
Discussions surrounding the concerns regarding tall claims from published work on the correlation of health indicating status based on accelerometer data needs demystification.
Challenges associated in the wearables for livestock monitoring needs elaboration ( Doi: 10.1016/j.sbsr.2016.11.004 and Doi: 10.1016/j.bios.2017.07.015 ) are some papers that can be cited in this review.
X, Y and Z axis data from tri axial accelerometer and the physiological correlation information from cattle and sheep needs a detailed discussion.
Table 1 - A bit more critical analysis is required. How many of these published studies have realistic cross validations? Can the researchers were able to identify which side of the body (right or left) the animal is lying based on accelerometer data rather than simply concluding animal is lying. what are the discrepancies associated in the variables? What kind of algorithms or methodologies were employed in the assessment of behaviours and health status based on the obtained data? What is the sample size in each of these studies, number of animals per study and the overall obtained data? Just listing the measurement validity does not make any sense to readers with instrumentation and technology background who work in the animal sector. The provided information in the table and the associated text in the draft is mere superficial.
Both cattle and sheep have been included in the keywords but a large number of relevant papers as per the distribution of species are missing and have not properly cited. What would be the influence on the differences in the size of the body and the associated accelerometer based data? Variations as per species type, life stages of animals (calves vs adult animal etc) and discussions surrounding the data are missing. What is the difference between accuracy and the precision in terms of reproducibility and repeatability of the data from the accelerometer and it’s correlation to the behaviours?
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
I am satisfied with the revisions that address my previous comments.
I recommend that the authors proofread the final manuscript, if accepted, to make sure there is no typo or grammatical error.
Reviewer 2 Report
All points requested were well argued or corrected
Reviewer 3 Report
The authors appeared to have revised the manuscript based on the provided comments and suggestions. The draft paper reads fine and can be accepted in current form.
