Proc. IEEE Int. Conf. on Communications, Work. Integrating Comm., Control, Comp. Tech. for Smart Grid, May 20-24, 2018, Kansas City, MO, USA.

Real-Time Testbed for Diversity in Powerline and Wireless Smart Grid Communications

Junmo Sung and Brian L. Evans

Department of Electrical and Computer Engineering, Wireless Networking and Communications Group, The University of Texas at Austin, Austin, TX 78712 USA -

Paper Draft - Presentation Slides - Software Release

Our Magazine Article on Smart Grid Communications Diversity

Smart Grid Communications Project


Two-way communication is a key feature in a smart grid. It is enabled by either powerline communication or wireless communication technologies. Utilizing both technologies can potentially enhance communication reliability, and many diversity combining schemes have been proposed. In this paper, we propose a flexible real-time testbed to evaluate diversity combining schemes over physical channels. The testbed provides essential parts of physical layers on which both powerline and wireless communications operate. The contributions of this paper are
  1. design and implementation of a real-time testbed for diversity of simultaneous powerline and wireless communications,
  2. release of the setup information and complete source code for the testbed, and
  3. performance evaluation of maximal ratio combining (MRC) on the testbed.
As initial results, we show that performance of MRC from measurements obtained on the testbed over physical channels is very close to that in simulations in various test cases under a controlled laboratory environment.

Questions & Answers

Q1. Is it practical to have a data aggregators support both powerline and wireless communications
A1. If a wireless communications network were initially deployed for the connection between a smart meter and the data aggregator, then the data aggregator will likely not be located on a power line and hence adding a powerline connection would be impractical. However, if a powerline communications network already exists, then it might be practical to add a wireless communications link at the data aggregator.
This particular project was funded by NXP Semiconductor and Texas Instruments through the Semiconductor Research Corporation. In their reference designs for smart meters, NXP and TI supported both powerline communication and wireless communications over the 900 MHz unlicensed band. Their idea was to allow power utilites to decide which communication modality to enable when a power meter is deployed in the field. In the project, NXP and TI asked us to look at what improvements in reliability could be possible for joint transmission over the powerline and wireless channels. Our testbed has shown 3 dB of SNR gain in the average bit error rate (BER) vs. SNR, or equivalently a reduction by three orders of magnitude in average BER for the same SNR setting.

Q2. The average coded BER vs. SNR plots show a shift in the shape of the curve for the MRC combiner of the PLC/wireless channels vs. separate decoding. However, the slope of the curve is the same, and this does not indicate an improvement due to diversity.
A2. Due to the independence of the PLC and wireless links, we would expect to see twice the diversity order play out in the communication performance. At high SNR, we would expect to see a doubling in the slope in the average BER on a log scale vs. SNR in dB curves for the hybrid system vs. that of a single link. In the paper, we used convolutional coding and Viterbi decoding, and swept SNR values from -6 dB to 4 dB for binary phase shift keying (BPSK) and -4 dB to 8 dB for differential BPSK using the testbed. For the average coded BER on a log scale vs. SNR in dB for BPSK, the slopes for the two data points with the highest SNR values are -1.1835 for the PLC link, -1.1896 for the Wireless Communications link, and -1.4652 for the MRC. The SNR range that we considered is too low to see the dramatic increase in the slope. We will seek to verify that as SNR increases, the average BER of the hybrid system decreases more rapidly (i.e. higher slope) than the BER of a single link. (Note: Prof. Naofal Al-Dhahir at The University of Texas at Dallas contributed to this answer.)

Q3. Did you consider simply selecting the link with the higher SNR?
A3. We didn't try this approach. From eyeballing the average coded BER rate vs. SNR curves, choosing the link with the higher SNR would not perform as well as the proposed MRC approach, except in the case of an exteme difference in SNRs in which selecting the link with the higher SNR would give the same result as the proposed MRC approach.

COPYRIGHT NOTICE: All the documents on this server have been submitted by their authors to scholarly journals or conferences as indicated, for the purpose of non-commercial dissemination of scientific work. The manuscripts are put on-line to facilitate this purpose. These manuscripts are copyrighted by the authors or the journals in which they were published. You may copy a manuscript for scholarly, non-commercial purposes, such as research or instruction, provided that you agree to respect these copyrights.

Last Updated 05/29/18.