A Master Course Student, Ms. Song published a Journal Paper in IEEE Sensors Journal
A Master Course Student, Ms. Shulin Song(2nd year master's student in Matsumoto Lab. of Security and Network Area) published a Journal Paper in IEEE Sensors Journal.
■Short Message from her Supervisor, Prof. Tad Matsumoto
We, Information Theory and Signal Processing Laboratory members, are all very much honored to the extraordinary achievement by Ms. Shulin Song. I hereby also express my appreciation to the hard work and diligence to the scientific work of all our laboratory members.
■Journal name & date of publication
IEEE Sensors Journal (Impact Factor 3.78)
This paper has been published in IEEE Explore, August 25, 2020.
※JAIST repository
■Authors
Shulin Song, Meng Cheng (graduated from the JAIST PhD course at March 2014, Matsumoto Lab., Security and Network Area), Jiguang He (Researcher in University of Oulu, PhD), Xiaobo Zhou (graduated from the JAIST PhD course at September 2013, Matsumoto Lab. , Security and Network Area), Tad Matsumoto (Professor in JAIST, main supervisor)
■Title
Outage Probability of One-Source-with-One-Helper Sensor Systems in Block Rayleigh Fading Multiple Access Channels
■Summary
In Information Theory and Signal Processing Laboratory, research on Lossy-Forwarding (LF)wireless networks has long been a main research topic, and the laboratory has publisheda lot of publications on this and its related topics at high Impact Factor journals. This is because the laboratory found for the first time in the world that the related lossy network problems can be formulated and solved in the theoretical framework of Distributed Multi-terminal Source Coding in Network Information Theory. In LF, the relay nodes continue forwarding the decoded information sequences, even though the decoded information part is found to have some errors. According to the Slepian-Wolf correlated source coding theorem, the original source information can be recovered losslessly by utilizing the correlation knowledge between the original and relayed sequences in the joint detectionprocess.
The paper published in IEEE Sensors Journal utilized the Multiple Access channel (MAC)region theorem, which states: the multiple signals "collide" in the wireless propagation process, the signals can be losslessly recovered by the joint detection so far as the ratesof the sources are within the MAC rate region.
However, when applying this theorem to the sensor network, the captured source information has correlation. Therefore, the MAC and Helper-assisted Slepian-Wolf (h-SW) rate regions have to have an intersection area. To identify the necessary-and-sufficient condition for the tworegions to have an intersection has long been an open and difficult question in the NetworkInformation Theory.
The authors used a very simple example scenario, one-source-with-one-helper network,andintroduced a very simple sufficient, yet accurate, condition. Then, they identified the conditions for the MAC and h-SW regions to have an intersection by characterizing the region structures into several distinctive cases, and calculated their occurrence probabilities.
It should be noted that the identified condition is a sufficient condition, it provides systemdesigners with an upper bound of outage curve which is a safer side of coverage definition, and hence, very important in practice.
By utilizing this result, it is made possible to calculate the outage probability in time-varying channels such as Rayleigh fading channels. Furthermore, in the paper, it is shown that the theoretical and simulation results are consistent, and hence the accuracy of the theoretical work has been proven to be correct.
Furthermore, the authors further extends the results to the case where the fading variations of the source and the helper links are time-correlated, and identified the impact of the fadingvariations. It has been shown that in the low average signal-to-noise power ratio (SNR) region, a second-order diversity curve can be achieved on the outage versus average SNR performance,however, the diversity order converges into one when the average SNR becomes large.
Finally, the authors also analyzed the outage performance where instead of MAC transmission,orthogonal transmission is used so that the two signals do not collide. The most significantoutcome of this paper is that the outage performance curves, including diversity order, are almost the same between the two cases, however, the MAC transmission does not two transmission phases, and hence significant throughput improvement can be expected.
August 27, 2020