Route Optimization for UGVs: A Systematic Analysis of Applications, Algorithms and Challenges

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Distinguished authors:

Your work is very interesting and addresses a leading topic. Of special interest and usefulness are the figures and tables you show as they are concrete, concise and help the reader to clearly identify the techniques used, the applications and illustrative references. From a bibliographic classification point of view, I consider it to be a good work. However, I consider that it could be improved in some aspects such as a formal description of the algorithms or techniques used in those papers, a quantitative comparison of the results obtained in those papers, information related to the operating systems used in the cited works, execution times, etc. that is, related to efficiency (for example, let's imagine that we were comparing or describing simulation algorithms of certain random variables, the ideal would be to have a comparative table of the efficiency of the algorithms. Something like this I think should appear in your paper). A comparative quantitative evaluation of the most commonly used algorithms, using standard metrics or benchmarks from the field, should be included. Also, a deeper description of certain approaches, especially those based on deep learning. In addition, it would be of interest to include a deeper description of the contents of the papers you cite, the main results obtained, so that the reader could see what is done and what line of research is open. 

Respectful regards.

 

 

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.

Comments 1: Your work is very interesting and addresses a leading topic. Of special interest and usefulness are the figures and tables you show as they are concrete, concise and help the reader to clearly identify the techniques used, the applications and illustrative references. From a bibliographic classification point of view, I consider it to be a good work. However, I consider that it could be improved in some aspects such as a formal description of the algorithms or techniques used in those papers, a quantitative comparison of the results obtained in those papers, information related to the operating systems used in the cited works, execution times, etc. that is, related to efficiency (for example, let's imagine that we were comparing or describing simulation algorithms of certain random variables, the ideal would be to have a comparative table of the efficiency of the algorithms. Something like this I think should appear in your paper). A comparative quantitative evaluation of the most commonly used algorithms, using standard metrics or benchmarks from the field, should be included. Also, a deeper description of certain approaches, especially those based on deep learning. In addition, it would be of interest to include a deeper description of the contents of the papers you cite, the main results obtained, so that the reader could see what is done and what line of research is open.

Response 1: Thank you for pointing this out. We agree with this comment. Therefore, we have modified Table 2. In the column “Control Techniques” we have described more explicitly the algorithms used for the described application and the contribution of each author to the specific field can be better seen. We have also added the column “Description of Algorithms” in which we concisely state what each author used the algorithm for in their respective study. The “Hardware/Software” column was modified to better identify the hardware and software used by each author. Likewise, the column “Execution Time” was added in which the time used in the different types of algorithms was placed in most of the cases where it was possible to extract the information. Finally, we added the column “Evaluation criteria” in which we included the metrics used by the authors to measure the efficiency of the algorithms used.

In addition, we have added the following at the bottom of Table 2: The data reveals that Cuckoo Search reduces time by 89% vs. PSO in route optimization, while HWGO achieves convergence times of 10.7s vs. 13.2s of WOA in FOPID tuning. For dynamic environments, ISAHS shows superiority with 0.82s vs. 1.15s of classical HS. The evaluation includes benchmarks such as TSPLIB (pr2392 instance solved in 1,216.14s with ACO+SA) and 400x400 cell maps processed in 1.2s using DEM-AIA. Advanced techniques such as Deep Neural Networks (DNN) are being employed for planning in unstructured 3D environments, where bio-inspired IDA reduces computational time by 38% versus ACO. For multi-agent systems, hybrid RL-DDPG architectures are described that improve the success rate to 92% in mobile obstacle avoidance.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents a systematic literature review (SLR) on route optimization techniques for unmanned ground vehicles (UGVs), with a focus on applications in precision agriculture, logistics, and surveillance. The authors adopt the PRISMA methodology to analyze recent studies, highlight trends, and identify technological challenges in the domain.

The systematic approach and structure of the paper are mostly appropriate, and the manuscript includes a broad set of references. However, the paper presents several major issues that require attention before it can be considered for publication.

• The review is largely descriptive. A stronger critical perspective is needed, highlighting the strengths, limitations, and comparative performance of the approaches discussed.
• Regarding the methodology, while the use of PRISMA is mentioned, the paper does not provide sufficient details on the review process (e.g., search dates, exact inclusion/exclusion criteria, quality assessment strategy). A more rigorous description of the review protocol is necessary. Table 1 and Figure 1 seem to be focused only on the agricultural sector. However no details are included in the field of other areas such as logistics or surveillance. 
• Additionally, the paper occasionally shifts its focus to general embedded systems or mobile robotics, moving away from the specific topic of UGVs. A clearer definition of scope is recommended.

Author Response

Comments 2: The manuscript presents a systematic literature review (SLR) on route optimization techniques for unmanned ground vehicles (UGVs), with a focus on applications in precision agriculture, logistics, and surveillance. The authors adopt the PRISMA methodology to analyze recent studies, highlight trends, and identify technological challenges in the domain.

The systematic approach and structure of the paper are mostly appropriate, and the manuscript includes a broad set of references. However, the paper presents several major issues that require attention before it can be considered for publication.

The review is largely descriptive. A stronger critical perspective is needed, highlighting the strengths, limitations, and comparative performance of the approaches discussed.

Regarding the methodology, while the use of PRISMA is mentioned, the paper does not provide sufficient details on the review process (e.g., search dates, exact inclusion/exclusion criteria, quality assessment strategy). A more rigorous description of the review protocol is necessary. Table 1 and Figure 1 seem to be focused only on the agricultural sector. However no details are included in the field of other areas such as logistics or surveillance.

Additionally, the paper occasionally shifts its focus to general embedded systems or mobile robotics, moving away from the specific topic of UGVs. A clearer definition of scope is recommended.

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.

Response 2: Agree. We have, accordingly, done/revised/changed/modified “table 2” to emphasize this point. A stronger critical perspective is needed, highlighting the strengths, limitations, and comparative performance of the approaches discussed. The following paragraph was also added at the end of Table 2: The data reveals that Cuckoo Search reduces time by 89% vs. PSO in route optimization, while HWGO achieves convergence times of 10.7s vs. 13.2s of WOA in FOPID tuning. For dynamic environments, ISAHS shows superiority with 0.82s vs. 1.15s of classical HS. The evaluation includes benchmarks such as TSPLIB (pr2392 instance solved in 1,216.14s with ACO+SA) and 400x400 cell maps processed in 1.2s using DEM-AIA. Advanced techniques such as Deep Neural Networks (DNN) are being employed for planning in unstructured 3D environments, where bio-inspired IDA reduces computational time by 38% versus ACO. For multi-agent systems, hybrid RL-DDPG architectures are described that improve the success rate to 92% in mobile obstacle avoidance.

To address the observation “Regarding the methodology, while the use of PRISMA is mentioned, the paper does not provide sufficient details on the review process (e.g., search dates, exact inclusion/exclusion criteria, quality assessment strategy).” We added in Table 1 in the inclusion criteria section: Research that includes efficiency metrics, Research that has at least 10 citations, Research from indexed journals. Likewise, in the exclusion criteria we add: Aerial (UAV) or maritime (USV) systems not hybridized with UGV’s, Applications that do not use UGV's. Finally, we add the row “search dates” with the following information. The information search was carried out from October 2024 to March 2025.

At the end of Figure 1, we have added the following paragraph: “The analysis of the thematic distribution in the reviewed studies reveals a predominance of the agricultural sector with 32 investigations (57%), followed by logistics with 18 papers (32%) and surveillance with six studies (11%). This agricultural predominance is explained by the urgent need to optimize routes in sowing/harvesting operations where algorithms such as WOA and GA reduce ground coverage time by up to 40%. In logistics, PSO and ACO approaches stand out for improving throughput in automated warehouses, although they present scalability limitations for networks of more than 1,000 nodes. The few works in surveillance (11%) employ hybrid RRT*-DRL techniques that achieve re-planning rates of 92% in dynamic environments, but require further validation in real scenarios. This distribution reflects current sectoral priorities where precision agriculture leads technology adoption, while surveillance emerges as a critical area with outstanding methodological challenges.”

Regarding the comment, “Additionally, the paper occasionally shifts its focus to general embedded systems or mobile robotics, moving away from the specific topic of UGVs. A clearer definition of scope is recommended”. We appreciate the comment on the need for a clearer definition of scope. The main objective of our work is to analyze route optimization specifically for unmanned ground vehicles (UGVs). We acknowledge that, in the introduction, general references to embedded systems and mobile robotics were included to contextualize the importance of edge computing and hardware platforms relevant to UGVs. To avoid potential confusion, we will explicitly reinforce in the introduction and methodology section that the central focus of the study is on UGVs, and adjust the wording of the opening paragraphs to make it clear that the technologies and platforms mentioned are addressed only in relation to their application in UGVs. In this way, we ensure that the scope of the manuscript remains clearly delineated and aligned with the main topic. At the end of the second paragraph of the Introduction, we have added the following: While there are multiple applications for these platforms, in this paper we will focus exclusively on their use and relevance for unmanned ground vehicles (UGVs). Similarly, at the end of the first paragraph of the “Review Protocol” subsection, it was added: Additionally, it should be made clear that the inclusion and exclusion criteria were only applied to research that specifically examined UGV's. To keep the scope of the analysis limited to autonomous UGV's, any reference to embedded systems or mobile robotics in this assessment only relates to their use in UGV's.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Distinguished authors:

Thank you very much for your reply, I think my doubts and observations have been clarified.

Best regards.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have provided a thorough and satisfactory response to the reviewers' comments. In light of the revisions made, I find the manuscript to be suitable for publication in its current form.