kumoh national institute of technology
Networked Systems Lab.

Long - IEEE Communications Letter
By :
Date : 2017-07-24
Views : 678

Reviewer: 1

Comments to the Author
This work presents a data rate-aware routing scheme for underwater cognitive acoustic sensor networks.
The proposed routing scheme exploits data rate and hop counts to construct a route that can maximize throughput and reduce latency.
Very interesting simulation results have been obtained.
The paper is easily readable and the structure is appropriate.
Authors should provide more details about the developed models and its validation.
Several typos should be corrected.
Authors should increase the text size of fig.1 and table 1.

Reviewer: 2

Comments to the Author
Minor changes

1). Page 1, line 14, It is difficult to acknowledge the communication mechanism of primary users.

To whom this acknowledgment is required? This sentence is vague, please, clarify the object in the sentence.

2). Page 1, line 16, as an the ON-OFF model, "the" is wrongly used. 3). Page 2, line 7 to 11, The PDF of on off in equation 1 and 2 can be described by a single equation. Similarly, the explanation of the lambda ON and OFF can be explained by the single variable, perhaps x={ON, OFF}.

4). Page 3, line 40, Problem Formulation, please cite the relevant article(s) about the challenge faced by the authors described in the text.

5). Page 3, line 10, Replace "an" with "a" in "an lower bound" on page 2 line 9 right column.

6). Page 3, line 51, left column, the word "so" is unnecessary.

7). Page 4, line 30, "Moreover, only sensing the links which have the maximal bandwidth to construct a data route, EFDA also considers hop counts as a key metric to select the next-hop". This sentence is hardly making any sense, please, rephrase to convey the intended meaning.

Major Comments.

The plots and their reasoning need major improvements.

1). The authors add no apparent reason for the fluctuating behavior of the throughput when plotted with respect to the number of CA users in Fig 1a.

2). In Fig(2b), the throughput of the proposed scheme is not changing with respect to the number of hops, however, the reason for this behavior is not given in the result discussion. Similar behavior in Fig(2c) is also needed to be explained with reasons.

3). In Fig(2c), the delay of the proposed scheme seems equal to zero (It may be very close to zero), The y-axis limit should be adjusted to show the actual values of delay.

4) The performance improvements or tradeoff should be described as the percentage increase or decrease.
Suggestion: if less number of markers are used in Fig2, b, c, d, figures would be better to understand.

Reviewer: 3

Comments to the Author
My main concerns with this paper are the following:
- It is unclear how the collaborative sensing strategy is carried out.
- Dividing the spectrum in several bands is inefficient. TDMA schemes are more efficient for multiple users. Even more, how is the bandwidth shared among transmitters (unequal bandwidth assigned to each)? How does each node now its allocated bandwidth? This requires some kind of management/control mechanism.
- The simulation results are based on pre-assigned length of duty cycles. In a deployed network these parameters should be measured. Besides, from paragraph above Section V, it should be explained how the environment condition (pn) can be used to avoid collisions. If it is an average value, nothing is known of its instantaneous value.
- The theoretical problem is stated, but the solution is not explained, only a likely procedure is briefly indicated.
- Some clarifications would be required for figure 2:
- If end-to-end delay is linear with the number of hops (equation 9), the line in figure 2.c for EFDA should linearly increase with the number of hops. It can't remain constant.
- Does Figure 2.b take into consideration the transmission of beacons, and the transmission at different hops? How are these 5000 bits split between transmitters? Is it bits per second?