Analysis of the Radio Coverage for a Mobile Private Network Implemented Using Software Defined Radio Platforms

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper investigates the radio coverage of a 4G mobile private network (MPN) implemented with an SDR platform as radio front-end and an open-source software suite (srsRAN 4G). A description of the equipment, the configuration parameters for the hardware and software components and the site of deployment and test measurements is given. Seven well-established propagation models are reviewed and applied to estimate the radio coverage of the implemented MPN via a commercial software suite (HTZ Communication). The resulting estimates are compared with field-test measurements for three LTE frequency bands, taken by a commercial mobile device using an Android app (G-MoN Pro) to log values of received signal strength. A detailed comparison is carried out and some performance indices are calculated for the results of each model considered. Based on this, a comparative assessment and ranking of these models is discussed, with an advantage of the the ITU-R 1225 model found.
In all, the assessment of the models studied is clearly and concisely presented, with findings consistent with the conclusions discussed in the last Sections. However, I suggest that some questions be clarified before publication, as follows:

Current interest for MPN implementations is mentioned in the Introduction, supported by two relevant review articles cited, and offering justification for the authors' focus on radio coverage of a MPN setup. (I also suppose that this setup offers advantages facilitating the testing procedure, e.g. knowledge and control of all network parameters.) On the other hand, are the results obtained here comparable with previous results for usual commercial networks? E.g. are the authors aware of similar comparative studies and, if so, are similar advantages / disadvantages of the same models supported by evidence therein? Some discussion would be very interesting.

Experimental test results are obtained using a COTS mobile device by means of an app. Notwithstanding obvious accuracy and repeatability concerns, I think that this practical approach is not without merit, since achieving adequate performance on a mobile user device is, ultimately, the goal of the network. However, a discussion on the reliability of the results would be welcome. Did the authors perform some check against measurements by standard instrumentation, or perhaps test with more than one commercial mobile devices? Further on, did they get a single measurement for each frequency examined on each "step" along the route of Fig. 4, or more, and in the former case do they consider it adequate?

According to Fig.4, the route chosen for measurements corresponds to a rather limited part of the coverage footprint areas shown in Fig. 3. Is there a specific reason for this choice? I expect that convenience / practicability considerations were important, but are there also others? Do the authors think it is indicative of the overall merits of the models considered?

A minor issue:
In lines 44-46 it is stated : "The authors propose an automation of the radio planning process of 5G networks in urban scenarios by developing a software solution capable of collecting data from a real map and extracting 3D terrain and buildings heigh information" (I note in passing that "heigh" should read "height"). However, this does not appear to refer to the subject of the present paper. Does this statement refers to work cited here, probably Ref. [5] which seems to be missing in the text? Please clarify.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The material in the manuscript is clear, presented logically and consistently.

The topic is relevant, the results of coverage area assessment obtained by modeling tools are compared with experimental results and this is a great advantage of this study. The conclusion corresponds to the content.

The only recommendation for the manuscript before publication is following.

The material in Section 2, according to the reviewer opinion, should be systematized into a table with respective conclusions. In section 4 authors summarize results in tables and this is what should be reproduced in section 2.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The paper describes a measurement campaign for an MPN and compares the measurement with predictions from different standard propagation models. I think that the paper could have been much more significant if it had discussed more in depth practical issues with the deployment of the MPN and the environment for taking measurements. 

Regarding Section 2, it is not very clear what is the interest of describing superficially the used models. Maybe a more in-depth discussion on each one, where are the models used and, predictably, which ones will be better for different scenarios. Moreover, a clear introduction should be written for the Section, indicating what is the interest to the overall work, what specific aspects are the most important for the study and what will be done to study such aspects. In other words, what is the purpose of comparing models.

The paper lacks a "problem formulation" section, which clearly describes what exactly are we aiming at (is it just correlating models with a specific campaign? or are we creating a framework to test different models in different scenarios to produce maybe a model that can, for instance, extrapolate missing points in a campaign?). I think such section would clear up the exact purpose of the paper.

A practical measurement campaign description does have its interest for a publication, but I think that the most important aspects are not discussed in this paper. Only choosing which model is the closest to the measurments is not a contribution per se, we expect the models to somehow predict what is measured. A framework for putting this into testm, systematically, over different setups, on the other hand, is more relevant.

Hopefully my comments can help the authors build a more ambitious second version.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thanks to the authors for their clear reply. My comments having been adequately addressed, I recommend publication of the paper.

Reviewer 3 Report

Comments and Suggestions for Authors

I find the paper has improved very much. Just to point out one detail, the new "Problem Formulation" subsection would more likely be the "Objectives" subsection. A "Problem Formulation" section is normally used to describe, normally in mathematical terms, what exactly is what the problem that the proposed method/algorithm/prototype solves. In the current form, I find that the Subsection 1.2 describes enough of what needs to be acheived. I would still suggest to include a proper "Problem formulation" Section, describing, for instance the metrics for coverage, the format of the measurements taken, etc. I think this will add scientific soundness to the paper; although it can be well understood in its current form.