Gaussian Processes for Signal Processing and Representation in Control Engineering
Round 1
Reviewer 1 Report
This review introduced that Gaussian processes have enormous potential for signal processing and representation in control engineering, which is of practical importance. Based on the analysis, various applications of Gaussian processes in the real process were proposed. However, there are still some problems that need to be modified to improve the paper.
- Enrich your references with recent related articles. Replace some older references with the related and latest ones.
- What is d(x_i, x_j) in Equation (8), there is no relevant explanation in the paper.
- In Section IV, Anomaly Detection in Industrial Processes, a detailed development process of Gaussian processes applications and comparisons are missing, and it is suggested to supplement the recent related work.
- There are related papers in Gaussian Processes area, such as “Loopy belief propagation: convergence and effects of message errors” and “Accuracy analysis for distributed weighted least-squares estimation in finite steps and loopy networks”. They could help to improve the paper.
- The limitations of Gaussian processes are not presented.
- introduction is not attractive and should need to revise again.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The authors reviewed Gaussian Processes (GP) for signal processing in control engineering. First, they defined a mathematical basis of the Gaussian Process and then gave a retrospective of computational methods. Furthermore, they proceeded with GP's applications, such as anomaly detection in industrial processes, predicting a battery condition parameters, electricity usage prediction, decision support for the bike-sharing system, tracking problem applications (mainly GPS based systems), cost optimization, diesel engines post-injection prediction of an oil change, model predictive control, and image processing application. The authors concluded that given today's availability of computational power, the practical usefulness of Gaussian processes increases significantly.
This is an interesting research review report. In my opinion, The manuscript is well organized, has a good structure, is easy to follow, and it would be an addition to the body of knowledge.
Before publication, I have the following comments and suggestions for authors:
- Please, consider deleting the "see" word before providing references (lines 27, 35, 42, 50,65, 73, 78, 113, 140, 153, 280, 302, 329, 330, 365, 383, 400, 442, 498, and 525). It is not necessary.
- Please, define the term IoT (Internet of Things) first time by introducing it in the manuscript.
- Please, consider providing block schematic representations for the crop-based precalculation process (L573 - L582). It would add to the article's readability.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The presented paper is interesting and shows the GP in an explicit way in terms of its application in control engineering applications. However, I would like to make the following comments.
Introduction: Regardless of the fact that the review focuses on control engineering processes, it is necessary to include some more reference to the application of GP to the proposed problems by the authors. In the final part, detail in a more concise way the structure of the article.
Having in account that this is a review article, it would be necessary to include a section where works carried out with other methods used and the justification of the importance of GP from a comparative point of view are discussed.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
