Reinforcement Learning-Driven Framework for High-Precision Target Tracking in Radio Astronomy

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper reports research that uses reinforcement learning (RL) to improve the tracking accuracy of radio telescopes. This is an unprecedented method and will provide useful information to radio engineers in the field, making it worthy of publication.

However, there are some areas in the paper that need to be revised. These minor corrections should be made before publication. The necessary corrections are listed below.

Figure 3 and equations (1) and (2) are inconsistent. The definitions of X and Y shown in the figure or equations (1) and (2) are likely incorrect, and the authors are requested to investigate this.

"2.3 The Tracking Control System" measures the deviation of the pointing coordinate relative to the object. This method can be used for objects that emit strong radio waves, but measuring the deviation will be difficult for celestial radio sources that emit weak radio waves. A brief explanation of this is recommended.

Please enter the frequency in Table 2.

There are some overlapping statements in the text. For example, lines 329 - 338 of "3. Experimental Setup and Design" contain the same statement twice. Please check your entire paper and revise it to make it easier for readers to read.

Line 307 cites Figure 10, followed by lines 353 and 359 citing Figure 14 and Figure 15. Please number the figures in sequential order.

Line 363 contains a caption for Figure 15, but the figure itself appears to be missing. Please add the figure.

Author Response

We appreciate the reviewer’s observation.

The authors have revised the manuscript accordingly, as detailed below.

Comment : Figure 3 and equations (1) and (2) are inconsistent. The definitions of X and Y shown in the figure or equations (1) and (2) are likely incorrect, and the authors are requested to investigate this.

Response: The authors have carefully reviewed the issue and found that there was an error in the definition of X and Y in Figure 3. The figure has been corrected accordingly to ensure consistency with equations (1) and (2).

Comment : "2.3 The Tracking Control System" measures the deviation of the pointing coordinate relative to the object. This method can be used for objects that emit strong radio waves, but measuring the deviation will be difficult for celestial radio sources that emit weak radio waves. A brief explanation of this is recommended.

Response: The authors agree that measuring the deviation of the pointing coordinate can be challenging for celestial sources with weak radio emissions. The signal level obtained indeed depends on the entire receiving chain, including the overall gain and receiver sensitivity. In the proposed concept, signal acquisition must be achieved before initiating the tracking process. Recognizing the limitation in received signal power, the authors introduced a “parser–staging module” that includes a “dB-scale to linear converter,” which transforms the SNR value from logarithmic to linear.

Because the absolute maximum signal strength cannot be predetermined, the system accumulates all measured SNR values during both the scanning and aiming phases in an “SNR History Array.” The highest recorded value within this array is then used as a reference for normalization. Subsequently, the measured SNR is normalized and clipped through an “SNR Normalization Function,” generating an “observation vector” whose elements range between 0 and 1. This normalized representation is supplied to the RL agent to estimate the next aiming point, as illustrated in Figure 11.

With this approach, both strong and weak radio sources are represented proportionally in the same relative scale, where a normalized value of 1.0 corresponds to the strongest received signal and 0.0 corresponds to the noise level.

 

Comment : Please enter the frequency in Table 2.

Response : The authors have added the operating frequency of the renovated 12-meter radio telescope in Table 2.

 

Comment : There are some overlapping statements in the text. For example, lines 329 - 338 of "3. Experimental Setup and Design" contain the same statement twice. Please check your entire paper and revise it to make it easier for readers to read.

Response : The authors have carefully reviewed the manuscript and removed redundant or overlapping statements in Section 3 [lines 336–349] and throughout the paper to improve readability and clarity.

 

Comment : [Line 307 cites Figure 10, followed by lines 353 and 359 citing Figure 14 and Figure 15. Please number the figures in sequential order.] and [Line 363 contains a caption for Figure 15, but the figure itself appears to be missing. Please add the figure.]

Response : The authors have revised the figure arrangement and corrected the numbering sequence accordingly. The missing figure has also been added to ensure that all figures appear in proper order and alignment with their corresponding captions. [Figure 11, 12]

Reviewer 2 Report

Comments and Suggestions for Authors

The paper describes a novel and interesting result of applying machine learning techniques to pointing a large dish antenna, and is likely to be of interest to a small but potentially growing, readership. The basic approach is sound and experimental method well designed.

However I have an important point of clarification I would like to see addressed before publication. In addition there are problems with the Figures, once of which appears to be missing, and some of which are of of sequence, I also have a number of minor corrections and points of clarification that I feel would improve the quality of the paper. 

I attach my comments in full as a Word document.

I attach a Word document giving more detail.

 

Comments for author File: Comments.pdf

Author Response

We appreciate the reviewer’s observation.

The authors have revised the manuscript accordingly, as detailed below.

Comment : The above sentence including “temporal encumbrance” and the immediately following sentence (“a more efficient strategy”) be rewritten to clarify their meaning. If the intended meaning is that the procedure described in the paper involves a periodic loss of sensitivity (SNR) owing to repeated scanning, but that more efficient strategies could be sought in future using a posteriori (historical) information about the target position, this should be made clear, as this really goes to the heart of the paper.

Response : The authors have expanded the discussion to clarify the concept of observation loss during the scanning intervals, which is similar to that encountered in other scanning techniques such as CONSCAN. In the revised manuscript, it is further explained that future enhancements could employ historical tracking data related to the past aiming position and it 's SNR to allow the AI to continuously infer and predict future trajectories. Moreover, a hybrid approach combining real-time scanning and historical tracking data is proposed, enabling the AI to autonomously decide when to perform additional scans under high prediction uncertainty or to rely on retrospective information when the prediction accuracy is sufficient. [line 466-479]

Comment : I recommend either a) giving one or more examples of real systems with “auxiliary tracking antennas” or b) deleting the references to these.

Response : The authors acknowledge that the term “auxiliary tracking antennas” may have been unclear. It has been revised to “auxiliary tracking system” in the manuscript. For example, in the Deep Space Communications & Navigation Systems (DESCANSO) network, several stations such as DSS-11, DSS-41, DSS-42, and DSS-51 employ monopulse tracking systems [W. A. Imbriale, Large Antennas of the Deep Space Network, JPL Publication 02-6, Pasadena, CA: Jet Propulsion Laboratory, California Institute of Technology, Feb. 2002]. The authors have also added clarification that such systems can serve as alternatives to auxiliary tracking setups, which typically require higher design and maintenance costs, thereby highlighting the potential cost-reduction advantage of the proposed approach. [line 134]

Comment : The caption for Figure 19 should note the variation of the Signal (SNR), e.g. “The variation of the AI tracking curve shows significant variations which are discussed in the text”.

Response : The authors acknowledge that the original caption and description may have been unclear. The sentence
“whereas the RL-based method maintained a higher average and successfully tracked peak variations over the full 30-minute interval.” has been revised to “whereas the RL-based method maintained higher average and peak levels, successfully sustaining continuous tracking throughout the full 30-minute interval.” [line 446-448]

Comment : Figures & Captions

Response : The authors have revised the figures and captions, corrected typographical errors, and rearranged them to ensure consistency and clarity throughout the manuscript.

Comment : Minor corrections and recommendations. The following is a list of minor corrections and recommendations.

Response : 

• Section 1

The sentence has been corrected [line 25, 57, 105], and the suggested reference has been added[line 65, 86, 123].

• Section 2

The sentence has been corrected [line 166, 168, 186, 219 and table2], And a definition and explanation have been added for the equation [line 289].

• Section 3

The sentence has been corrected and revised for conciseness and clarity[line 339-343, 356].The unit “degree (°)” has been added to the table 4.

• Section 4

Time zone information has been added to clarify the local time [line 408, 426].

• reference 

The reference error has been corrected [line 569,576, 583].

Reviewer 3 Report

Comments and Suggestions for Authors

This is well-structured and demonstrates sufficient improvements over existing methods. Nevertheless, it is suggested that the authors enhance both the abstract and the conclusion by including more quantitative descriptions that highlight the key contributions of the work.

 

Figure 11 appears to be overlapping with another figure. It is recommended that the authors revise the layout to improve clarity.

The resolution of the legends and axis labels in Figures 12 to 19 appears to be insufficient, making them difficult to read. It is recommended that the authors improve the image quality for better readability.

Author Response

We appreciate the reviewer’s observation.

The authors have revised the manuscript accordingly, as detailed below.

Comment : Figure 11 appears to be overlapping with another figure. It is recommended that the authors revise the layout to improve clarity.

Response : All figure has been rearranged to avoid overlapping and improve clarity.

 

Comment : The resolution of the legends and axis labels in Figures 12 to 19 appears to be insufficient, making them difficult to read. It is recommended that the authors improve the image quality for better readability.

Response : The resolution and readability of the legends and axis labels in Figures 12 to 19 have been improved to ensure better clarity.

 

Reviewer 4 Report

Comments and Suggestions for Authors

Main obstacle for  positive review of the manuscript is low "RADIO ASTRONOMICAL"  applications of the this antenna.  You did not present the aim of the research.  12-m antenna at ~50 cm wavelength  with  the beam ~2 deg fits only for solar observations, not for M33. You will never see M33. Errors ~0.3deg for beam-width of 2deg is too much for accurate pointing and tracking. You did not refer on classical authors in the radio telescopes:  J. Kraus,  J. Baars  and others.

Comments on the Quality of English Language

English vocabulary  did not always follow the traditional radio astronomical terminology.

 

 

Author Response

Comment : Main obstacle for  positive review of the manuscript is low "RADIO ASTRONOMICAL"  applications of the this antenna.  You did not present the aim of the research.  12-m antenna at ~50 cm wavelength  with  the beam ~2 deg fits only for solar observations, not for M33. You will never see M33. Errors ~0.3deg for beam-width of 2deg is too much for accurate pointing and tracking. You did not refer on classical authors in the radio telescopes:  J. Kraus,  J. Baars  and others.

Response : The authors would like to thank the reviewer for the valuable comments and would like to clarify that the primary objective of this research is not to conduct direct astronomical observations, but rather to develop and evaluate a large-antenna control system for space object tracking using reinforcement learning (RL). The goal is to adapt and apply this approach to a renovated telecommunication antenna for future radio astronomy research.

The mention of M33 serves only as an example in the simulation used to evaluate the system’s stability and tracking accuracy, and does not involve real signal observation. Additional clarification has been included in the revised manuscript, stating that the simulation represents the tracking of the hydrogen line (21 cm) [line 397, 402] to reflect realistic operational conditions of the refurbished 12-meter satellite dish. This study focuses on assessing the system’s capability to maintain continuous long-duration tracking under real environmental conditions — a crucial step toward deployment with high-sensitivity radio telescopes in future applications.

Furthermore, the authors have revised the Conclusion section to emphasize that tracking accuracy remains a key factor requiring further improvement, particularly through the development of more precise and efficient AI models, to enable future practical use in astronomical target tracking [line 514].

Finally, the Introduction section has been revised to explicitly state this research objective, and additional references to classical works — J. D. Kraus, Radio Astronomy, 2nd ed., 1986, and J. W. M. Baars, The Paraboloidal Reflector Antenna in Radio Astronomy and Communication, 2007 — have been incorporated to strengthen the theoretical foundation of radio telescope principles [line 118-120].

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

 Please  submit yout manuscript  to Sensors or any more relevant journals