A Two-Stage Co-Evolution Multi-Objective Evolutionary Algorithm for UAV Trajectory Planning

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Review of „A two-stage co-evolution multi-objective evolutionary algo-2 rithm for UAV trajectory planning”

 

Summary: The presented work focuses on trajectory planning for single assets under multiple (two), potentially conflicting, mission objectives (flight distance and distance to obstacles) and five constraints (safe distance between trajectory and obstacle, maximum altitude, speed, flight slope, turn angle). Solving the bounded MOO problem splits in an exploration and an exploitation phase of an evolutionary algorithm. The exploration phase relies on two populations (one with constraints, one without). Exploitation phase identifies the pareto front based on exploration information. Three test scenarios were tested. The proposed algorithm was able to successfully avoid obstacles while also respecting the objective of low flight distance. However, calculation time was higher compared to alternative algorithms.

 

Chapter 1: Introduction:

General remark: The Introduction is well written and well structured. The most common procedure – reducing the MOO to a SOO – is only mentioned in a very brief section. However, there is more than only a weighted sum to reduce the MOO to a SOO (e.g., see doi:10.1017/aer.2023.68). Extending the discussion a little bit would cover a much bigger field of academic literature.

This work also relies on an evolutionary algorithm. The introduction should hold a brief description on the working principles of evolutionary algorithms (or meta-heuristics as general). Chapter 2.2 gives more details of the different populations, etc. This is hard to follow without a rough understanding what populations are, why they are needed, etc. Having a brief description in the introduction would also help readers to understand chapter 2.2.

Line 62: Dobbins curve should be Dubins path

 

Chapter 2: Related research background

Chapter 2.1:

The chapter is well written and well structured. The mathematical fundamentals are explained in an adequate extend. Several references (16-23) are given.

Eq. 1: not all element so of the function are explained. What are F, f, p?

Chapter 2.2:

Without a rough understanding of the working principles of evolutionary algorithms this chapter will be very hard to understand. I already suggested to add some descriptions on the working principles of evolutionary algorithms in the introduction.

Fig 2: The Figure itself and the figure description are on different pages.

Chapter 3 General framework of MOUTP based on TSCEA

Chapter 3.1

The chapter is well structured and clear.

Chapter 3.2

The chapter is well structured and clear.

Chapter 4

Table 3 header and body are on different pages.

Chapter 5

The conclusion is well written and structured. It is different from the abstract and contains additional, more detailed information.

Chapter 6 Patents

Patents does not seem to be the correct chapter heading.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The present work addresses a two-stage coevolution multiobjective evolutionary algorithm for UAV trajectory planning. A comparison with the classical DE algorithm, as well as three constrained algorithms. The proposed algorithm performed better than the existing algorithms. As a minor comment, the authors could consider adding a list of acronyms before the references. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

After reading the manuscript here are the recommendations:

Abstract:

1. Specify the optimization objectives and constraints clearly (e.g., minimizing travel time and energy consumption; considering obstacle avoidance and UAV performance limits).   
2. Simplify complex sentences to improve readability by breaking them into shorter ones. The following lines can be improved ": "Then, in order to effectively deal with the UAV constraints on the object space limitation, the TSCEA was divided into an exploration phase and an exploitation phase: the exploration phase adopted a two-population strategy, where the main population was not limited by the UAV constraints and the auxiliary population treated the constraints as an additional objective, with the aim of improving the algorithm's exploration ability under the constraints.
3. Ensure a logical flow from introducing the problem to describing the methodology, and then presenting the results and conclusions.
4. Highlight the significance of the findings by explicitly stating the advantages of the proposed method over existing approaches.
5. Use consistent terminology throughout the abstract, avoiding redundancy and ensuring clarity.
6. Eliminate unnecessary phrases that do not add value, such as "At the same time" and "Therefore."

Introduction: The introduction of the abstract can be improved in several aspects. Firstly, it could benefit from a clearer and more concise presentation of the problem statement, specifically highlighting the inadequacy of single-objective optimization in addressing the complexities of UAV trajectory planning. Secondly, the introduction should more directly introduce the concept of multi-objective optimization for UAV trajectory planning (MOUTP) and its significance in providing more flexible and comprehensive solutions. Thirdly, the discussion on existing research examples could be simplified to maintain the flow and readability of the paragraph. Additionally, the presentation of gaps and challenges should be more concise and focused, avoiding repetition and ensuring clarity. Finally, the introduction should clearly and early on present the proposed solution and its contributions to addressing the identified challenges in UAV trajectory planning. By addressing these aspects, the introduction can become clearer, more concise, and better structured, providing a comprehensive overview of the study's purpose, methodology, and key contributions.

Experiments: To improve the section on validating the TSCEA algorithm, it should be clearer and more concise, with simplified sentences and removal of redundancy. The structure should be well-organized, clearly delineating test instances, comparison algorithms, parameters, and evaluation metrics. Figures and tables need to be well-integrated with clear explanations. Consistent terminology and detailed descriptions are necessary to avoid ambiguity. The section should also focus on highlighting the most significant results and insights gained from the experiments. These improvements will make the text more accessible and effectively communicate the validation process and outcomes.

Conclusions: The conclusion of the study could be improved in several aspects. Firstly, it would benefit from a clearer and more concise summary of the key findings and contributions of the research. This could include a succinct restatement of the objectives, methodology, and main results, focusing on the most significant outcomes. Secondly, the presentation of the proposed algorithm and its stages could be streamlined for better clarity and readability. Providing a more structured overview of the TSCEA and its implementation process could enhance the understanding of the algorithm's functionality and effectiveness. Additionally, the discussion on the verification of the proposed algorithm could be more concise, focusing on the comparison with existing methods and highlighting the superiority of the TSCEA in terms of performance and solution quality. Finally, the conclusion could end with a clear statement of the study's implications and potential future research directions, emphasizing how the findings contribute to advancing the field of UAV trajectory planning. By addressing these aspects, the conclusion can become more impactful and effectively communicate the significance of the research findings.

 

Comments on the Quality of English Language

The quality of English can be improved for better clarity and readability. The text contains some complex sentences and technical jargon that may be difficult for readers to follow. Simplifying the language, breaking down long sentences, and improving the overall flow can enhance comprehension. Additionally, minor grammatical corrections and consistent terminology use would further improve the quality. Here are some specific suggestions:

1. Simplify and clarify complex sentences.
2. Break down long paragraphs into shorter, more focused ones.
3. Ensure consistent use of technical terms and abbreviations.
4. Correct minor grammatical errors and improve sentence structure for readability.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The paper discusses the trajectory planning for unmanned aerial vehicles (UAVs) in three-dimensional space. It introduces a multi-objective evolutionary algorithm called TSCEA (Two-Stage Co-Evolution for UAV Trajectory Planning) to solve the MOUTP (Multi-Objective Unmanned Aerial Vehicle Trajectory Planning) problem.

The TSCEA algorithm aims to increase the searchability of the population across the search space and find feasible domains. It uses a two-stage constraint-handling technique with two populations to promote diversity and convergence of solutions. A runtime diagram of the TSCEA algorithm is provided for better understanding.

The exploration stage of the TSCEA algorithm ignores UAV constraints to find as many feasible domains as possible. The assistant population in the TSCEA algorithm uses constraints as additional objective functions to find promising solutions. 

Here my general comments on the paper: 

The document provides equations and algorithms without explaining their significance or how they contribute to solving the problem at hand.

A table containing technical terms and abbreviations with proper explanation would make it less challenging for non-experts to understand the content.

Section 3: The manuscript briefly mentions the proposed algorithm (TSCEA) but fails to provide a detailed explanation of its principles or how it addresses the multi-objective UAV trajectory planning problem.

Section 3 and results: The authors do not discuss the limitations or potential drawbacks of the proposed algorithm or methodology, which could impact its practical implementation. Does the complexity of the proposed algorithm play a role in the trajectory design?

Section 4:  in the results, the authors dont discuss the motivation for the chosen benchmark algorithms.

The terrains used for evaluating the results should be classified between simple or hard for the UAV trajectory.

The document contains inconsistent formatting (especially the format of the algorithms), making it difficult to follow the structure and flow of the content.

Comments on the Quality of English Language

acceptable, some typos

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

All my comments have been addressed