Indoor 3-D RT Radio Wave Propagation Prediction Method: PL and RSSI Modeling Validation by Measurement at 4.5 GHz

1

The authors carefully answered to all my previous questions   and comments. I have just a minor remark: references 18 and 20 are identical.

Thanks a lot for your remark, reference 20 has been rectified to make 18 and 20 differentiated.

1

The RSSI (received power indicator) is mainly affected by   the PL (indoor environment model), we can think that if PL is accurate, RSSI   will be also. Discussing both is useless.

As per your valued suggestion, the discussion about RSSI has been significantly reduced. We removed old table 3, figure 6 and lines 424-426,435-437, and 441-468 (tracking changes view).

2

I do not understand why the authors compare the accuracy of   the proposed model point by point, whereas it would be more judicious to   compute the error between the models (3D-RT and SBRT) and  the measurements. Furthermore,  an overall    comparison should be considered for both models , for example using   the MMSE, to show how 3D-RT is more accurate than SBRT. The idea is to   quantify 3D-RT accuracy improvement as in figure 6, for the RSSI modeling,   both 3D-RT and SBRT show really no significant differences in accuracy

We used the point to point comparison to express the accuracy level in details with respect of zoom-in view. As per your valued suggestions, the overall error have been incorporated in the manuscript for the SBRT and proposed method with respect to measurement using Standard Deviation (SD). Less SD express the higher   accuracy of simulation. After calculated the SD of the proposed method with respect to measurement, it has been compared with the SBRT method to express the accuracy improvement of the proposed method. Please see the lines 423-424, 433,   444-446, 487-488, 494-495, and 427-529   

3

Tracking changes should be removed from the document for   readability,  I suppose it is an   oversight of the author at the submission of the article.

The paper has been resubmitted version and we are bounded to use tracking changes as per editor direction.

1

In this paper, the authors proposed an indoor   "3D" RT Radio Wave Propagation Prediction Method, which has been validated   with PL and RSSI with measurement data. The paper is interesting and   well-organized. This reviewer has only one comment: different height of RX is   not discussed; the model seems like only 2D model.

We used the same height for all Rx as per the measurement. In measurement, the same mobile station placed one by one in every point and re-run the measurement which is mentioned in lines   151-154 (Track changes view). Only for visualization of 2-D layout and ray tracing simulation, we have provided the figures 3, 4(a), and 5(a) in the manuscript. Moreover, to visualize ray tracing simulation 3-D model we have provided the figures 4(b), and 5(b), respectively.

1

The improvements made are very interesting. A last question   arises of itself, is the reduction of the SD from 4.91 dB for SBRT to 2.69 dB   for the proposed method is significant and justifies the complexity increase   of 3D proposed method?

Thanks a lot for your question, as compared to 3D SBRT method, our method can achieve similar or slightly better accuracy, but with reduce complexity. The complexity of the proposed 3-D RT is low because only pre-define rays are finally launched.   The lower number of launching rays give better accuracy even under a higher level of interactions. Moreover, less computational resources are needed to handle this method. In conventional methods, more rays are shot in all directions, so the complexity of the calculation increases massively.   However, the shooting of massive ray's object touching calculations is needed to perform blindly without knowing whether it contributes to Rx or not. Our method uses less time and, therefore, less computational complexity.   In section 5.1 we discussed details about the complexity Analysis of proposed method including mathematical analysis.

2

A conclusion statement should be made on this point

A final statement of overall PL SD, for SBRT and the proposed method, has been mentioned at the conclusion in lines 547-549.