Finite Memory Structure Filtering and Smoothing for Target Tracking in Wireless Network Environments

Round 1

Reviewer 1 Report

As a general comment, the present paper represents an interesting contribution to filtering/smoothing in wireless environments. However, before publication, it is my opinion that the following major comments should be carefully addressed by the authors:

1)      The authors should explain, since from the abstract, what they mean by “unified algorithm”.

 

2)      The statement of contributions should be rephrased so as to provide a more rigorous exposition of the technical novelty provided, e.g. possibly with a bullet list.

 

 

3)      The following related works on target tracking should be discussed so as to provide a richer introduction to the topic:

[R1] "SOI-KF: Distributed Kalman filtering with low-cost communications using the sign of innovations." IEEE Transactions on signal processing 54.12 (2006): 4782-4795.

 

[R2] "Tracking the tracker from its passive sonar ML-PDA estimates." IEEE Transactions on Aerospace and Electronic Systems 50.1 (2014): 573-590.

 

[R3] "The Kalman like particle filter: Optimal estimation with quantized innovations/measurements." Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference. IEEE, 2009.

 

[R4] "Distributed classification of multiple moving targets with binary wireless sensor networks." 14th International Conference on Information Fusion. IEEE, 2011.

[R5] "A hash-tree based approach for a totally distributed track oriented multi hypothesis tracker." 2012 IEEE Aerospace Conference. IEEE, 2012.

 

4)      Please add a notation paragraph at the end of Sec. I.

 

5)      In Sec. II the authors consider a linear state space model. Could the present approach be generalized to the general non-linear case?

 

 

6)      I would like the authors to provide a discussion on the computational complexity of the proposed tracking approach, e.g. in terms of big O notation.

 

7)      Conclusions should be enriched with what the authors consider to be further avenues of research.

Author Response

The author appreciates the reviewers’ comments which helped the author improve the paper. The author has fully reflected the reviewers’ comments in the revised paper.

 

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper is generally well-written and easy to understand. The unified algorithm of finite memory filtering and smoothing proposed in this paper is novel and has various advantages over conventional algorithms, including intrinsic robustness against model uncertainties. Mathematics in this paper seems to be correct. However, I have some concerns about the simulation as follows.

1. The title indicates that the algorithm is proposed for target tracking application in wireless sensor network environment wireless network environment. The constant velocity motion model does not include a control input term. However, the unified algorithm was derived considering control input. In the simulation, Equations (15) and (16) indicate that the position of a target is directly measurable, which does not consider properties of wireless sensor network environment.

2. In Figures 1-4, RMSEs of the proposed algorithms are not distinguishable from those of the Kalman filter. Compare the algorithms in terms of time averaged value of RMSE.    

Author Response

The author appreciates the reviewers’ comments which helped the author improve the paper. The author has fully reflected the reviewers’ comments in the revised paper.

 

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

As a general comment, the present paper represents an interesting contribution to filtering/smoothing in wireless environments. Additionally, the authors have either adequately responded to my concerns or modified the paper accordingly to reflect them.

Hence, I am glad to recommend the present work for publication.

Author Response

Dear Reviewer

 

The author is very grateful for the positive evaluation of the reviewer.

Reviewer 2 Report

Localization in wireless sensor networks environment uses various measurements, such as time of arrival (TOA), time difference of arrival (TDOA), angle of arrival (AOA), and received signal strength (RSS). These measurements are represented by nonlinear measurement equations. I think the simulation of this paper can be improved by using the measurements mentioned above.

Author Response

Answers to Comments of the Reviewer #2

 

(Overall rating of the Reviewer #2) Localization in wireless sensor networks environment uses various measurements, such as time of arrival (TOA), time difference of arrival (TDOA), angle of arrival (AOA), and received signal strength (RSS). These measurements are represented by nonlinear measurement equations. I think the simulation of this paper can be improved by using the measurements mentioned above.

 

The author fully understands the reviewer’s recommendation. As the reviewer may know, the consideration of the nonlinear system for the proposed unified algorithm has been mentioned as a future work in Conclusion section of the 1st revised paper. However, the author thinks that a clearer explanation is needed at this point. As the reviewer pointed, in target tracking applications using the CV motion model, we can encounter the state-space model where the state equation is linear and only the measurement equation is nonlinear. The reference [23], which the current author worked on as a corresponding author, deals with this type of state model for the FMS filtering. However, the proposed unified algorithm consists of both FMS filtering and smoothing. Thus, in order that the proposed unified algorithm deals with this type of state-space model, researches on the FMS smoother for nonlinear systems should be preceded. Therefore, the author thinks that the application of the nonlinear measurement equation can be a future work of the current paper.   

In order to reflect this, a couple of sections “Application for Target Tracking Problem” and “Conclusion” have been revised.

The author hopes that the reviewer will accept the author’s answer.