kumoh national institute of technology
Networked Systems Lab.

Review Comment

NSL > Education> Review Comment
Royyan - [Wireless Days 2017]
By :
Date : 2017-01-25
Views : 81

======= Review 1 =======

> *** Comments to the author: Summarize the strengths and weaknesses of the paper. Provide a rationale for your rating, and suggested improvements (if appropriate).

This paper considers the problem of faulty node detection in wireless sensor networks. To detect faulty nodes, the authors argue for the use of a Markov chain model. Thus, an a sound analytic modeling is provided and extensive simulations are conducted to demonstrate the efficacy of their approach.

First of all, the paper is very well written. Both grammar and the style of included figures are convincing and the story is easy to follow. I also like to mention that the problem statement and the discussion of related approaches is straightforward and helps the reader to get into the paper.

The topic addressed by the paper is relevant to the WD conference. Moreover, the performance evaluation seems to show better results in terms of faulty node detection and miss-detection rate.

Despite of that, the degree of novelty of the paper is quite limited. Indeed, such topics have been well studied in the literature, over which many key papers are missing from the citation list. More importantly, in the existing results much more generalized optimization formulation exists. Just Google-Scholar search WSN, fault node detection, the list of relevant papers will be there.

I would suggest comparing the performance of the proposed design with more recent works. Indeed, the authors quote many papers, but are not able to relate to them, yet they assume that their strategy outperforms others.

It would be much more convincing to add the confidence intervals for the curves. I also think that averaged results of 50 or 100 simulation runs (for e.g.) presented with confidence intervals will improve some conclusions in the paper.

> *** Familiarity: Rate your familiarity with the topic of the paper.
Truly expert in this area of research (1)

> *** Relevance to the track and timeliness: Rate the importance and timeliness of the topic addressed in the paper within its area of research.
Acceptable (3)

> *** Technical content and scientific rigour: Rate the technical content of the paper, its scientific rigour and novelty.
Valid work but limited contribution. (3)

> *** Quality of presentation: Rate the paper organization, the clearness of text and figures, the completeness and accuracy of references.
Readable, but revision is needed in some parts. (3)

> *** Overall evaluation: Please judge whether the paper should be accepted or rejected
Borderline paper (3)

======= Review 2 =======

> *** Comments to the author: Summarize the strengths and weaknesses of the paper. Provide a rationale for your rating, and suggested improvements (if appropriate).

In this paper, the authors address the reliability of sensors by proposing a Markov Chain Model to formulate the problem. The authors classify each node in one of the following state: i) Good, ii) Warning and iii) Bad as defined in Section II.B.
Then, they propose two algorithms to examine the state of each node. The authors validate their proposal with simulation based on Matlab. The simulations results show that the proposal achieves good performance.

Unfortunately, many comments can be formulated for this paper:

1. The text must be improved and the reading is really difficult. I believe that the authors do not revise the paper before submission.
2. The confidence interval of the simulation results must be calculated.
3. The authors should validate their solution by assuming a whole protocol stack with using network simulator such as NS3, OPNET, QualNet, etc. In fact, the numerical validation with Matlab is really poor.

> *** Familiarity: Rate your familiarity with the topic of the paper.
Working in this area of research (2)

> *** Relevance to the track and timeliness: Rate the importance and timeliness of the topic addressed in the paper within its area of research.
Little (4)

> *** Technical content and scientific rigour: Rate the technical content of the paper, its scientific rigour and novelty.
Marginal work and simple contribution. (4)

> *** Quality of presentation: Rate the paper organization, the clearness of text and figures, the completeness and accuracy of references.
Substantial revision work is needed. (4)

> *** Overall evaluation: Please judge whether the paper should be accepted or rejected
Weak reject (4)

======= Review 3 =======

> *** Comments to the author: Summarize the strengths and weaknesses of the paper. Provide a rationale for your rating, and suggested improvements (if appropriate).

The paper proposes a faulty node detection scheme with a hybrid algorithm using a Markov chain that gives each node three different states (good, warning and bad), the master node can analyze these states
- What is the difference between warning and bad states?
- The hidden markov model is not used in the paper; please check the keywords
- The authors need to revise the figures and move the legends such that they dont cover the plotted results (see figure 5 and 6), and revise the other plots
- In the Fig.5, Why the miss-detection rate increases in the "1st scheme in only CRC"?
- The references format needs revision

> *** Familiarity: Rate your familiarity with the topic of the paper.
Familiar with this area of research (3)

> *** Relevance to the track and timeliness: Rate the importance and timeliness of the topic addressed in the paper within its area of research.
Good (2)

> *** Technical content and scientific rigour: Rate the technical content of the paper, its scientific rigour and novelty.
Valid work but limited contribution. (3)

> *** Quality of presentation: Rate the paper organization, the clearness of text and figures, the completeness and accuracy of references.
Well written. (2)

> *** Overall evaluation: Please judge whether the paper should be accepted or rejected
Weak accept (2)

======= Review 4 =======

> *** Comments to the author: Summarize the strengths and weaknesses of the paper. Provide a rationale for your rating, and suggested improvements (if appropriate).

Strengths:
- the topic addressed in the paper is timely and relevant
- the proposed scheme is interesting and has relevant applications
- the authors evaluate their scheme against state of the art counterparts

Weaknesses:
- the paper could be better written. Some sentences are gramatically wrong and hard to understand.
- the scheme proposed could be better explained. It is hard to understand the main underlying idea.
- the authors should explain early in the paper what BCH codes and possibility distribution are. The average reader will probably don't know anything about it.
- the number of Table I is missing in Section II.B

> *** Familiarity: Rate your familiarity with the topic of the paper.
Familiar with this area of research (3)

> *** Relevance to the track and timeliness: Rate the importance and timeliness of the topic addressed in the paper within its area of research.
Excellent (1)

> *** Technical content and scientific rigour: Rate the technical content of the paper, its scientific rigour and novelty.
Valid work but limited contribution. (3)

> *** Quality of presentation: Rate the paper organization, the clearness of text and figures, the completeness and accuracy of references.
Substantial revision work is needed. (4)

> *** Overall evaluation: Please judge whether the paper should be accepted or rejected
Borderline paper (3)