Collaborative Indoor Positioning by Localization Comparison at an Encounter Position

Round 1

Reviewer 1 Report

MDPI Applied Sciences Journal (Manuscript ID: applsci-2416821)

 

Comments to the Author

 

This paper proposes an indoor positioning approach that  combines Bluetooth Low Energy (BLE) based encounter communication, Pedestrian Dead Reckoning (PDR), and map matching using particle filters. The technical concepts are explained clearly. However, there are several points need to be addressed to improve the quality of the manuscript.

 

Suggestions to improve the quality of the paper are provided below:

 

1)     Abstract:

 

·       Please provide the problem statement in the abstract to give readers an overview of the field and challenges before mentioning the proposed approach.

 

·       Also, there is a bit of an order problem in the abstract; for an easy reading for readers, findings and implications should be stated at the end, before the methodology. The line that states, “We obtained an accurate localization comparison of localization using a machine-learning model.” should be part of the methodology and mentioned before stating the results. Please restructure the abstract in a more coherent form.

 

·       I also suggest briefly mentioning details of the study setup in the abstract. In line 44, it is mentioned that experiments have been in eight facilities, for example, this detail should be included in the abstract.

 

2) BLE-based indoor localization has been an important domain, and there are many established works that are not mentioned in the manuscript. Very popular applications of BLE-based localization are used in emergency management, occupancy tracking, smart grid and smart energy management in buildings. Please mention these applications with supporting references to provide a broader overview to readers.

 

BLE used in building emergency management

Filippoupolitis, A., Oliff, W., & Loukas, G. (2016, December). Bluetooth low energy based occupancy detection for emergency management. In 2016 15th international conference on ubiquitous computing and communications and 2016 International Symposium on Cyberspace and Security (IUCC-CSS) (pp. 31-38). IEEE.

 

BLE used in occupancy tracking in office spaces

Tekler ZD, Low R, Blessing L. An alternative approach to monitor occupancy using bluetooth low energy technology in an office environment. InJournal of Physics: Conference Series 2019 Nov 1 (Vol. 1343, No. 1, p. 012116). IOP Publishing.

 BLE used for smart grid applications (residential homes)

Collotta, M. and Pau, G., 2015. A novel energy management approach for smart homes using bluetooth low energy. IEEE Journal on selected areas in communications, 33(12), pp.2988-2996.

 

BLE used for smart energy management (office buildings)

Tekler, Zeynep Duygu, et al. "Plug-Mate: An IoT-based occupancy-driven plug load management system in smart buildings." Building and Environment (2022): 109472.

 

3)     Line 294, I am not very clear with how the feature importance scores are calculated in Figure 11. Please provide more details of the calculation and elaborate on the interpretation of the results, other than just highlighting what can be seen in the figure. We also should understand what could be the potential reasoning behind having one feature be found much more important than the rest.

 

4)     In Section 3.3.4, the authors mentioned that they chose the GBM algorithm for localization. Please provide a reasoning for the chosen algorithm.

 

5)     Please improve the conclusion section by mentioning 1) the limitations of the proposed approach and 2) potential future directions that could be explored.

 

 

The English used in the manuscript is understandable. However, my recommendation is to improve the coherence between sentences in the manuscript for readers to follow easier. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper proposed a method that combines BLE, PDR and map matching using particle filter to localize indoor positioning. I find this paper is hard to understand, I think an extensive editing of English is required.  I have some general questions for the authors: Why a particle filter is chosen for this type of tracking? Have you considered using other algorithms such as unbiased ensemble square root filter, Kalman filter, Ensemble Kalman filter? I recommend that the authors adding these references for a fully picture of objective tracking using particle filters:

Particle filters for multiple target tracking 

https://doi.org/10.1016/j.protcy.2016.05.215

A new approach to linear filtering and prediction problems

https://doi.org/10.1115/1.3662552

Bayesian filtering: From Kalman filters to particle filters and beyond

Statistics, 182 (1) (2003), pp. 1-69

A new development in magnetic particle tracking technology and its application in a sheared dense granular flow https://doi.org/10.1063/1.5100739

A Cubic 3-Axis Magnetic Sensor Array for Wirelessly Tracking Magnet Position and Orientation https://doi.org/10.1109/JSEN.2009.2035711

 

 

F.

 

Gustafsson

, 

Adaptive Filtering and Change Detection

 (

Wiley

, 

New York

, 

2000

) 

  Content wise please address these issues mentioned below. 

Line 29: Please specify what is this method based on?  

Line 124: how do the authors determine the particles number? Is 2000 sufficient enough? 

Line 146: Can the authors specify from where to where on figure 4 is considered as encounter phase? 

Line 146-148: I am not sure what the authors' meaning by 'communication is not established for 10 seconds after the previous communication to prevent communications with the same collaborator immediately after the data exchange?' Are you referring to that once the previous connection is established the communication will not be re-connected for 10 seconds? If that is the case, what if the two devices re-encounter right after the first met? 

Line 150: how to compare the localization accuracy of the two devices? 

Line 159: In figure 5 b) Why there is a red line crossing pixels area that are set as zero? How is it possible to cross the Gound Y area in Figure 5a? 

Line 179: Based on my understanding that the disappeared particles means that there are very low possibilities that the device will locate at those certain positions which indicates that the device is very likely passing through an obstacle. I don't think 'particles passed through the obstacle' is a proper description. 

Line 187: Why the ground truth is the ' waling walking path with a fixed interval'? Is it valid to assume that the interval is fixed? 

Line 217: Why the theta at time step i is determined by the average of all previous heading direction? As what the authors showed in Eq 2, turning position has a much larger theta than walking straight, I do not see how this estimation is valid. 

Overall, I found some sentences are difficult to understand and there are many grammar mistakes. There are several typos in this paper, I have pointed out in my comments to the authors. I recommend an extensive editing of English..

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thank you for addressing my comments and concerns. The revised version of the manuscript is now more clear and comprehensive. Great job!

Author Response

We greatly appreciate the reviewer's comments and suggestions. They are helpful and improve the manuscript significantly. We would like to express our gratitude for the excellent service of the reviewer. 

Reviewer 2 Report

Thank you for the authors inputs and feedback. I think you have answered all my questions. Some expressions need to be polished that I have provided in the Quality of English Language section. Other than this minor issue, I am inclined to publish this work. 

One suggestion for the authors is that I notice many expressions 'we did something' which is not very professional. I suggest going through the whole paper once again and try some different expressions. 

Line 278-280 'We performed 100 trials to evaluate the proposed method using each walking data. We used the evaluation metric for the localization described in section 3.3.3. We used two comparison methods for indoor localization using map images.' 

Line 288 'We experimentally determined d = 1.91 and σE = 1.26.'

d = 1.91 and σE = 1.26 are determined experimentally.

Line 289 'We showed the localization results in Table 2.'

Table 2 shows the localization results. 

Author Response

We greatly appreciate the reviewer's suggestions. The manuscript was undergone extensive language correction. The red words are the corrected words. Please see the revised manuscript. 

Round 3

Reviewer 2 Report

Thank you for the authors' effort. I suggest publishing this paper. 

Thank you for the authors' editing, the expression of this paper is significantly improved.