Challenges and Solutions for Autonomous Ground Robot Scene Understanding and Navigation in Unstructured Outdoor Environments: A Review

Round 1

Reviewer 1 Report

The authors have conducted a comprehensive review of the state-of-the-art in each step (perception, mapping, and planning) of autonomous mobile ground robots in their paper. This work could assist researchers in this domain to better understand the challenges, weaknesses, and strengths of existing approaches, and thus improve the efficiency of robot navigation in outdoor and unstructured environments.

A few syntactical modifications are necessary to do as follows:

Line 5, 237, 700, 825, 935, 1021, 1091: In academic literature, it is generally not common to use the apostrophe "s", as in "robots's external environment understanding" and "Microsoft's Kinect camera", and "method's benefit", etc.

Line 50: There is a grammatical error such as "for weeding".

Line 213: The numbering format as 2.1.1 is normally used for manuscripts in books. It is better to use bullet points or numbers for subsections of each sub-section in the article as follows:

• Vision-based sensor types

1. Vision-based sensor types

Line 242: For the phrase "time-of-flight (ToF)", the first letter of each word must be capitalized, like "Time-of-Flight (ToF)". The same notation should also be used for CSPN in line 437.

Line 535: Replace the phrase "image feature detection" with "image feature extraction".

Line 572: The sentence "In [84] Two CNN-based networks process point cloud and RGB front-view images in parallel" is grammatically incorrect.

Line 578: The number of iterations for which the network was trained seems to be mistyped. It should be corrected.

Line 732: The phrase "scene graphs" appears twice in the sentence.

Line 741: The word "of" is missing in the sentence.

Remove the information from the arXiv paper from the manuscript. We cannot trust the content of an arXiv paper because no journal/proceeding has accepted it for publishing yet.

Comments for author File: Comments.pdf

Author Response

We thank the reviewer for their considered and thorough review of our manuscript, and the valuable suggestions provided for improvement. We have made the following changes in response to these recommendations, and believe that the manuscript is significantly improved as a result.

• The sentences with apostrophe “s” in lines 5, 263, 728, 852, 963, 1049, 1119 have been modified.
• The grammar of line 55 has been improved.
• The numbering format of the subsections in each section was changed (e.g., line 239).
• The acronyms of the phrase “time-of-flight” in line 263 and the notation for CSPN in line 463 have been corrected.
• The phrase "image feature detection" at line 562 has been replaced with "image feature extraction".
• The grammar on line 599 has been revised.
• The number of iterations of the network training step in line 605 has been added.
• Line 760 has been amended to remove the additional phrase “scene graphs”.
• The missing word “of” has been added in line 769.
• Information from the arXiv paper has been removed from the manuscript and alternate peer reviewed citations added.

Reviewer 2 Report

This paper provides a comprehensive review and critical analysis of these methods in the context of their applications to the problems of robot perception and scene understanding in unstructured environments, and the related problems of localisation, environment mapping and path planning. The innovation points of this article are clearly described. However, there are still some minor problems that need to be modified, as follows:

 

1. In "Introduction" section Related Works, I feel the current coverage of the state of the art is not satisfactory as the related work section does not cover many contributions that likely provide the building blocks of the proposed approach. For example,

(1) Flexible gait transition for six wheel-legged robot with unstructured terrains

(2) Towards Hybrid Gait Obstacle Avoidance for a Six Wheel-Legged Robot with Payload Transportation

It is suggested to cite the above articles and analyze the differences in Section Related Works.

 

2. There are some mistakes including grammar, words and English expression in this paper. Please check the overall paper carefully. 

 

After the above problems are solved, the manuscript can be considered to be accepted.

This paper provides a comprehensive review and critical analysis of these methods in the context of their applications to the problems of robot perception and scene understanding in unstructured environments, and the related problems of localisation, environment mapping and path planning. The innovation points of this article are clearly described. However, there are still some minor problems that need to be modified, as follows:

 

1. In "Introduction" section Related Works, I feel the current coverage of the state of the art is not satisfactory as the related work section does not cover many contributions that likely provide the building blocks of the proposed approach. For example,

(1) Flexible gait transition for six wheel-legged robot with unstructured terrains

(2) Towards Hybrid Gait Obstacle Avoidance for a Six Wheel-Legged Robot with Payload Transportation

It is suggested to cite the above articles and analyze the differences in Section Related Works.

 

2. There are some mistakes including grammar, words and English expression in this paper. Please check the overall paper carefully. 

 

After the above problems are solved, the manuscript can be considered to be accepted.

Author Response

We thank the reviewer for their considered and thorough review of our manuscript, and the valuable suggestions provided for improvement. We have made the following changes in response to these recommendations, and believe that the manuscript is significantly improved as a result.

• The introduction section has been amended to include discussion of some additional related works, including references to the suggested publications.
• The paper has been thoroughly checked for grammar and English expression and a number of revisions have been made.

Reviewer 3 Report

This paper provides a comprehensive review and critical analysis of autonomous mobile robot navigation, identified the research gaps and challenges, and suggested promising future research directions for improved autonomous navigation in outdoor unstructured terrains. 

I believe that this work is valuable and could be published. However, authors need to address the following issues:

1. Abstract is too long, I recommend to shorten the Abstract section.

2. In this work, authors have investigated several research works. However, authors missed a few significant research works in the field, including:

A. Vision-based holistic scene understanding towards proactive human–robot collaboration

B. Outdoor scene understanding of mobile robot via multi-sensor information fusion

C.  A Semantic Classification Approach for Indoor Robot Navigation

D. Scene understanding in a large dynamic environment through a laser-based sensing

I recommend investigating the above research works and include them in your study if needed. 

3. Authors need to add a short paragraph in the Introduction Section to present the importance of their work. 

4. Authors are required to analyse the reason behind focusing on outdoor navigation systems. I believe that indoor navigation systems are challenging and significant tasks as well. 

I believe that the paper is well written. 

Author Response

We thank the reviewer for their considered and thorough review of our manuscript, and the valuable suggestions provided for improvement. We have made the following changes in response to these recommendations, and believe that the manuscript is significantly improved as a result.

• Some content was removed from the abstract to reduce its length.
• In the Introduction section, the importance of publications B, C, and D were discussed.
• The introduction section has been amended to include discussion of some additional related works, including references to three of the four suggested publications.
• A new section has been added to the Introduction highlighting the importance of the work.
• A section justifying the reasons for focussing specifically on outdoor robot scene understanding and navigation has been added to the Introduction section.

Round 2

Reviewer 2 Report

All my comments have been answered in the revised paper and the revised paper can be considered for publication.