Cellular Positioning in an NLOS Environment Applying the COPSO-TVAC Algorithm
Round 1
Reviewer 1 Report
In this paper, the authors proposed a COPSO-TVAC algorithm for cellular positioning in NLOS environment. While some valuable works have been conducted, I have the following comments to improve the paper.
1. Are there any advances in recent years? The authors should add some latest references and compare the proposed algorithm with state-of-the-art algorithms.
2. The writting needs to be improved. Please keep the consistency of tense when introducing related works.
3. Some of the equations are too small, for example, equation (2). Please check this issue throughout the paper.
4. The quality of figures needs to be improved. For example, Figs. 1-2 are not with high resolution.
5. The authors compared the positioning results of urban and suburban scenarios. I suggest the authors clearly show what parameters are different in the two scenarios and how to set the values.
6. In Figs. 5-7, how many times of simulations are conducted to obtain the CDF curves? Please clarify it.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The authors proposed a cellular positioning method includes hybridization of PSO with three techniques to create a quality initial PSO population.It is stated that the COPSO-TVAC method can attain better localization performance and the performance is compared over PSO-TVAC,TRR and TSLS. However, there are some flaws that need to be modified:
1. As far as I know, there are so many positioning method in mobile communication industrial, such as Threshold, Event isolation, STA/LTA and Cadence etc. The TOA is one of the classic and commonly used method. It does not mean that it is the most efficient.
2. The principle of COPSO-TVAC in current scenario is not clearly described, why choose CS to generate the initial population for PSO-TVAC.
3. Why the OBL will help the chaotic initial population search for better positions ? Too much content to describe PSO-TVAC and PSO.
4. The simulation results of the proposed method are made with TRR and TSLS, both proposed many years ago (2009 and 2007, respectively). Why did the authors compare with these ones? I think the targets of performance comparison should be reconsidered.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Review of the paper
Cellular Positioning in NLOS Environment applying COPSO-TVAC Algorithm
Stevo Lukić and Mirjana Simić
Determining the coordinates of a radio emission source using Microwave Sensors Networks under Non-Line-Of-Sight conditions (NLOS), when the direct Line-Of-Sight (LOS) path is blocked, is an urgent task of scientific interest and practical significance. When developing a machine learning algorithm for efficient optimization of the objective function of the maximum likelihood estimator, the authors use the chaos search procedure based on chaotic maps.
The text of the paper should be revised.
1. When describing the statistical characteristics of the ultra-wideband (UWB) indoor channel delay profile (CDP), three models are most often used: The Exponential Model, The Cluster Model and The Exponential-Lognormal Model. The authors use the Gaussian distribution to describe the statistical properties of the NLOS channel (see formula (6)). Such an approximation should be substantiated and the limits of its application should be indicated.
2. In the ratios of the standard deviation of a single distance measurement error (see formula (10)) and the average distance offset NLOS (see formula (11)) the coefficients k1 and k2 must be determined. There is also a contradiction: relation (11) was obtained for the exponential approximation of the distribution, although the authors use the Gaussian probability distribution.
3. It is desirable that the authors justify the choice of the nonlinear objective function (see formula (13).
4. The use of iterations in the study of the convergence properties of the proposed metaheuristic algorithms in the scenario usually leads to a bias in the estimate. The authors should explain the choice of the optimal number of iterations they carry out in machine learning.
5. When comparing the variance of estimates obtained by different methods (see graphs in Fig. 3 and Fig. 4), it can be seen that some estimates have a value less than Generalized CRLB. It is desirable that the authors provide an explanation for this result.
Comments for author File: Comments.pdf
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Thanks for the reply by authors. The revised version of the manuscript entitled "Cellular Positioning in NLOS Environment applying COPSO-TVAC Algorithm" has been sufficiently improved. After a little modification of the language, it can be accepted.
