Curve-Localizability-SVM Active Localization Research for Mobile Robots in Outdoor Environments

Round 1

Reviewer 1 Report

The paper describes a novel method for Curve-Localizability based on support vector machines to assist active localization of robots in outdoors environment. The method is relied on the modeling of 3D point cloud data extracted from 3D lidar to create the motion model of the robot. Overall the paper is technically sound and well written, yet some minor improvements are required to improve the quality of the manuscript.

 

The abstract describes the method superficially, it is too generic. It should be improved to reflect the exact novel content introduced in the paper

The introduction could be benefited from a summary bullet points from the major contributions introduced in this work.

The related work should be enchanted with more SVM path planning methods based on 3D LIDAR data. At the current form, only on paper is referred with SVM local path planning and this is for indoors environments. The authors could also have a look at and refer to the works presented in [1] -[3] and have a qualitative comparison with them in the related work.

[1] Charalampous, K., Kostavelis, I., & Gasteratos, A. (2015). Thorough robot navigation based on SVM local planning. Robotics and Autonomous Systems, 70, 166-180.

[2] Tennety, Srinivas, et al. "Support vector machines based mobile robot path planning in an unknown environment." Dynamic Systems and Control Conference. Vol. 48920. 2009.

[3] Qingyang, C., Zhenping, S., Daxue, L., Yuqiang, F., & Xiaohui, L. (2012). Local path planning for an unmanned ground vehicle based on SVM. International Journal of Advanced Robotic Systems, 9(6), 246.

 

Regarding the methodology the technical information provided is very clearly presented.

The only concern that should be better clarified is how the subsequent paths produced from the SVM are connected each other?

Which is the predicted path length? Is it depending on the range of the obtained LIDAR measurements?

Regarding the experimental results, the paper credibility could be further increased by adding some comparisons of the localization accuracy against GPS measurements. Some tables and graphs showing the positioning error would be of great benefit.

 

Minor Typos:

Line 300: Correct the ??

Line 201: The term passive global localization is not very concrete. Please rephrase

Line 143: index, We
Line 129:  “length of point” do you mean distance of point?

Line 2: The phrase “Increase of environment size” is vague, consider rephrasing

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper is based on 3D observational model to define a curve-localizability-index and path for active localization of a robot in outdoor environment.

The paper explores a research field of interest in the era of mobile assistive robotics. Now, the paper presents some gaps in the literature, description of the model and results.

• In related works, the authors have to add the quantitative results of the papers that they cited, also to better decline this work.
• The explanation of the algorithm is not clear and in particular, it is not clear the inputs needed for the processing of the outputs. A detailed paragraph can be added to explain Fig. 1.
• Also referring to Fig. 1, why some text is in red? Probably the authors can add some visualization methods to explain what they developed and what they used as already developed in other works.
• In general I suggest to pay attention in providing not explained information. For example, AMCL algorithm is not defined. Probably, the authors meant Adaptive Monte Carlo Localization.
• In table 1, the measurement units have to be included in brackets or better explained.
• In the paragraph after table 1, the authors have to include the quantitative results in respect to the state of the art. For example, “although the path length is partly increased of X in respect to Y”, etc.
• Line 300. The final results are shown in Tab. ??. The final results are missing.
• Also in the conclusion the quantitative results are missing.

As last remark, I suggest the authors to provide the differencies between the traditional methods and the new method in terms of accuracy as they introduced in the abstract: “The efficiency and accuracy of traditional localization strategies in wild environments are significantly reduced.”

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The authors modified the paper according to the reviewer's comments.