A Modified Methodology for Generating Indoor Navigation Models

Round 1

Reviewer 1 Report

Navigation networks are fundamental questions for indoor navigation. This paper presents an interesting method to generate navigation networks considering both corridors and rooms. The method uses the vertices of walls to create a triangulated irregular network, then selected the midpoints of the edges within corridors as base points and generates Voronoi diagram based on those midpoints. The method generates networks that well adjust the shape of corridors. While considering rooms, the method adds link room-entrance-midpoint, where the midpoint belongs to an adjacent cell of the room. In general, the method has been clearly described, and the presented examples show that the method works well and generate navigation routes while avoiding acute angles. The paper does have a lot of texting issues.

In the abstract, line 18, does the paper mean ‘based on’ rather than ‘based’?

Line 54, literature [7] has not used GRM (RGM?) as the abbreviation of regular-grid model.

Line 62, ‘[4,8] developed a method’, the sentence should start with the papers ‘Lee [4,8] developed’. The same problem applies for line 96, where ‘had been developed by [39]’ should be ‘had been developed by Kodzik [39]’. Line 238, the paper used ‘Lewandowicz and Lisowski (2018) [37] observed’. The ‘(2018)’ is not expected here. There are a lot other places like this. The paper should check all the places where names should be explicitly referred.

The captions in Figure 1 is confusing. In particular, the ‘(c)’ in ‘and NRSE (corridor-entrance-room) (c) models’ is not in the correct place.

Line 134, 135, 150, 154, 178, 181, 183 and many more, I observed ‘PMid’, ‘P_{Mid}’, and ‘P_{mid}’. I assume they refer to the same thing. If so, they should be written in the same way.

Line 172, I consider the paper mean ‘Dijkstra algorithm’ rather than ‘Distkri algorithm’?

Line 192, ‘Fig. b’ should be ‘Fig. 5b’. Similarly, Line 220, ‘Fig. 8a’ should be ‘Fig. 8b’.

In Figure 2, ‘Midpoints of inner triangle edges’ and ‘mid-line’ is too small to be identified in printed A4 version. The legends in Figure 5, 6, 7 are also too small.

In Figure 6a, why the legend for ‘Topology of MPRSSE network’ is just empty. Also, in this legend, the first item ‘Shortes’ should be ‘Shortest’. The legend for ‘Topology of the MPRSSEM network’ becomes invisible as well in printed A4 paper.

In Figure 7, the caption ‘The second shortest route’ is inappropriate. The paper means ‘Another example of the shortest route’.

In Figure 11, 12, 13 and 14, the ‘b)’, ‘c)’, and ‘d)’ should not overlap with the images.

Line 248, 250, there is a new term 'MP'. Does the paper mean 'midpoint'? Is 'MP' necessary here.

Line 281, ‘section 5’ should be ‘Section 5’.

Line 280, ‘chapter 4’? Does the paper mean ‘Section 4’?

In References 9, the paper referred an article ‘Afyouni, I.; Cyril, R.; Christophe, C. Spatial models for context-aware indoor navigation systems: A survey. 318 Journal of Spatial Information, 2012, Science 1.4, pp. 85–123’, but the text ‘Seamless Pedestrian Navigation in 319 Indoor/Outdoor Large Spaces with No Clear Patterns for Movement. Available from: 320 https://www.researchgate.net/publication/322569712_Seamless_Pedestrian_Navigation_in_IndoorOutdoo321 r_Large_Spaces_with_No_Clear_Patterns_for_Movement [accessed Jan 22 2018].’  Is not supposed to be here.

For reference 10, line 324, the year ‘20176’ is not correct.

Line 362, the ‘220)’ is incorrect.

Please check all the references for capital letters and ‘.:,)’ marks, etc.

For the methodology, it seems the paper has not considered more than one entrances (e.g., doors) in one room, but may need a sentence to discuss this. As a suggestion of the method, from Figure 5c, the modification can also be adding an edge between each entrance and its adjacent segment of corridor. In this case there will be no such edges as marked in white in Figure 5c.

Author Response

We would like to thank Reviewer 1 for a thorough perusal of our manuscript. The Reviewer's valuable remarks have been taken into consideration during the revision process, and the manuscript has been revised accordingly. The authors would also like to thank the Reviewer for valuable editing suggestions and apologize for the errors, in particular those relating to inconsistent use of abbreviations. Figure 14 has been modified, and a new paragraph has been added to improve the Results section.

Author Response File: Author Response.pdf

Reviewer 2 Report

As the authors declare -The aim of this study was to propose a new method for the automatic generation of indoor navigation networks based on the floor plan. The proposed method is easy to use, and it involves simple transformations of geometric data from floor plans- but  the whole paper deals with the progress of a specific previous research   conducted in 2006-2015. As the topic developed  a lot all over the word a wider frame would have been appreciated, explaining  better differences when exhisting from the specific research of the paper. Wider means with more illustration for the algorithms for exhample inthe paraghaph related to previous theories and application.

Author Response

We would like to thank the Reviewer for the positive comments. This study modifies the algorithm which had been described in detail in 2018 in an article by Elżbieta Lewandowicz and Przemysław Lisowski entitled "Methodology to generate navigation models in buildings", Journal of Civil Engineering and Management, 2018, 24(8), pp. 619-629. https://doi.org/10.3846/jcem.2018.6599.

In this study, the algorithm has been modified with the use of midpoints rather than segment centroids. The algorithm was not described because it had been discussed in detail in a recent study (2018) based on centroids. The description of the algorithm would be nearly identical, and the authors were concerned about allegations of autoplagiarism. In this study, the methodology was described based on graphic presentations.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript presents an interesting question to settle the path generation in indoor navigation process. Some geometric algorithm have developed for the axis line generation. The following comments can be referenced.

 1) In the field of geometric shape analysis, there are many methods extracting polygon region  axis line, for example the medial axis transportation (abbreviated as MAT) based on the angular bisector operation, the Delaunay triangulation based method. In the introduction, more words need to be added to summarize MAT question discussing the main methods, advantages/disadvantages,  application conditions. The following reference can be included

____ Aichholzer, O.; D. Alberts, D.; Aurenhammer, F.; and Gärtner, B. A novel type of skeleton for polygons. J. Universal Comput. Sci. 1995, 1(12), 752–761.

 2) In the manuscript , the author mentions “The method proposed in this study relies on a new approach to the segmentation of structures. The applied segmentation method minimizes the irregularity of segments by dividing space across corridors.”, How to prove “minimize”. The zigzag line can be simplified based on the skeleton extraction from TIN. The following paper discusses this question. Suggest you reference it and compare the skeleton extraction

____ Tinghua Ai, Shu Ke, Min Yang & Jingzhong Li (2017) Envelope generation and simplification of polylines using Delaunay triangulation, International Journal of Geographical Information Science, 31:2, 297-319, DOI: 10.1080/13658816.2016.1197399

3) For indoor region analysis, what element and edge points are applied to handle the skeleton line? And what role are them to play among the wall, door, corridor and others. I expect to see a normal definition. The following reference can be feferenced.

___Abdoulaye A. Diakité & Sisi Zlatanova (2018) Spatial subdivision of complex indoor environments for 3D indoor navigation, International Journal of Geographical Information Science, 32:2, 213-235, DOI: 10.1080/13658816.2017.1376066

4) The example presented in the manuscript is too simple, the complex floor data is expected with such as more turns.

5) The current method regards the corridor area even without difference in navigation behavior. If the condition is heterogenous, for example the light condition, the smooth condition is different , how to extract a weighted skeleton?

6) Section 2 can be combined into introduction.

Author Response

We would like to thank the Reviewer for valuable comments which have enabled us to improve the quality of the manuscript. All suggestions have been taken into consideration in the revision process. The Reviewer's remarks have enabled us to adopt a broader perspective on the discussed problem. Figure 14 has been modified, and a new paragraph has been added to improve the Results section. Point-by-point responses to the Reviewer's comments are presented below.

Point 1.  In the field of geometric shape analysis, there are many methods extracting polygon region  axis line, for example the medial axis transportation (abbreviated as MAT) based on the angular bisector operation, the Delaunay triangulation based method. In the introduction, more words need to be added to summarize MAT question discussing the main methods, advantages/disadvantages,  application conditions. The following reference can be included

____ Aichholzer, O.; D. Alberts, D.; Aurenhammer, F.; and Gärtner, B. A novel type of skeleton for polygons. J. Universal Comput. Sci. 1995, 1(12), 752–761.

Response 1: Thank you for this valuable comment and the reference which has expanded our knowledge of extraction methods. The MAT problem appears to have many solutions, many of which are highly complex. We have proposed a simple transformation method with the use of GIS tools. The Reviewer's comment has been taken into consideration in the revision process, and the suggested reference has been added. We have also included an additional reference to Cheng, SW., Vigneron, A., Motorcycle Graphs and Straight Skeletons. Algorithmica (2007) 47: 159. A broader discussion about MAT would require in-depth knowledge and more extensive studies. One of the co-authors will incorporate the relevant information in his doctoral dissertation. Once again, thank you for this valuable remark.

Point 2) In the manuscript , the author mentions “The method proposed in this study relies on a new approach to the segmentation of structures. The applied segmentation method minimizes the irregularity of segments by dividing space across corridors.”, How to prove “minimize”. The zigzag line can be simplified based on the skeleton extraction from TIN. The following paper discusses this question. Suggest you reference it and compare the skeleton extraction

____ Tinghua Ai, Shu Ke, Min Yang & Jingzhong Li (2017) Envelope generation and simplification of polylines using Delaunay triangulation, International Journal of Geographical Information Science, 31:2, 297-319, DOI: 10.1080/13658816.2016.1197399

Response 2: The provided reference has been incorporated into the manuscript. It presents an interesting solution that the authors were not familiar with. In the described methodology, the axis is not a zigzag line because it is developed based on the midpoints of edges that intersect corridor space. The model better fits the potential navigation route and occupies the center of corridor space when a larger number of points are used to generate the network. (Fig. 10-14). The generated paths can be smoothed with additional algorithms. Such possibilities are discussed by the suggested reference. Thank you for bringing it to our attention.

Point 3) For indoor region analysis, what element and edge points are applied to handle the skeleton line? And what role are them to play among the wall, door, corridor and others. I expect to see a normal definition. The following reference can be referenced.

___Abdoulaye A. Diakité & Sisi Zlatanova (2018) Spatial subdivision of complex indoor environments for 3D indoor navigation, International Journal of Geographical Information Science, 32:2, 213-235, DOI: 10.1080/13658816.2017.1376066

Response 3. The skeleton lines in the network model are developed based on the topological relationships between segments in 2D space. These relationships are determined by identifying the left and right polygons of each edge. Entrance points are extensions of topographic data. The choice of midpoints of TIN edges, which are required for segmentation with Voronoi diagrams, is the most important step in the presented methodology. A larger set of points for generating TINs increases the resolution of the network model which then better fits the corridor axis. This is shown in Figures 11-14. The study by Abdoulaye et al. (2018) is worth recommending because it identifies a set of subspaces with dedicated properties in 3D space. The study is cited in the last paragraph of the revised manuscript to set directions for future research.

Point 4. The example presented in the manuscript is too simple, the complex floor data is expected with such as more turns.

Point 5. The current method regards the corridor area even without difference in navigation behavior. If the condition is heterogeneous, for example the light condition, the smooth condition is different , how to extract a weighted skeleton?

Response 4, 5. The aim of simple graphics in the first chapters was to present the methodology. Successive examples are more complex, and they are presented in expanded 2D space (Fig. 10-14).  This methodology was described in greater detail by Elżbieta Lewandowicz and Przemysław Lisowski in an article entitled "Methodology to generate navigation models in buildings", Journal of Civil Engineering and Management, 2018, 24(8), pp. 619-629. https://doi.org/10.3846/jcem.2018.6599. The algorithm was not described because it had been discussed in detail in the above recent (2018) study based on centroids. The description of the algorithm would be nearly identical, and the authors were concerned about allegations of autoplagiarism. In this study, the methodology was described based on graphic presentations.

Thank you for pointing out that weights should be considered in the network model. The weights should describe the conditions in variable segments of the network. This approach should be adopted during the development of a navigation system. We will consider this option in future research to expand the proposed methodology. Thank you for this valuable remark.

 Point 6. Section 2 can be combined into introduction.

Response 6. Section 1 (Introduction) has been combined with section 2 (Research object).

Author Response File: Author Response.pdf