1. Diversity and Coexistence within Smart Grid Communications
Prof. Brian L. Evans, Wireless Networking and Communications Group, The University of Texas at Austin
Students: Ms. Ghadi Sebaali and Mr. Junmo Sung Current Collaborators: Prof. Naofal Al-Dhahir (UT Dallas) and Mr. Mostafa Sayed (UT Dallas) Past Collaborators: Dr. Jing Lin (Qualcomm), Dr. Marcel Nassar (Samsung), Dr. Aditya Chopra (NI)
Sponsors: Semiconductor Research Corporation GRC under Task Id (from Freescale and Texas Instruments)
Objective: Improve reliability of smart grid communications
(1. Project Overview)
(2.1 Wireless/PLC Diversity) 
Non-identical channel, noise and interference statistics
Maximal Ratio Combining is a maximum likelihood optimal technique for AWGN
OFDM transmission with 256 sub-channels and BPSK modulation
0.4 MHz sampling rate
Wireless Link noise model: Gaussian Mixture with 2 components
PLC Noise Model : cyclostationary noise
Legend
Wireless Link
PLC Link
Focus
Neighborhood-area smart utility network between a data concentrator and smart meters along two paths
 Instantaneous SNR Combining gives 7dB gain over PLC alone
(2.2 Coexistence Mechanisms)
Powerline (PLC) Link
Wireless Link
low-voltage power lines in kHz band
unlicensed wireless MHz band
ah and g standards share the 900 MHz ISM band
Path Loss (PL) Model used is outdoor large-zone model
Interference Model is used as follows: d(Rxv, Txv) = dD and d(Rxi, Txi) = dU
Metric used is desired/undesired signal ratio 
Goal Integrating customers in the grid
Scale voltage with energy demand
Bill customer using real-time rates
Analyze customer load profiles
Improve system reliability
Interferer
Victim
Victim Required FER (%)
Critical distance (m) g
802.11ah 1
5 – 6.5
802.11ah, g,
Project Components
Interference Mitigation
PLC/wireless Diversity
Coexistence Mechanisms
PLC/wireless Testbed
Limitations
Signal attenuation in propagation channel
Reflections at impedance discontinuities
Noise or interference
Frequency selectivity
f selective or wideband:
if time difference between the received copies of the transmitted signal (DELAY SPREAD) is significant
Nassar
Figure 3.11:
Part 1:
Home area networks (HAN) that
connect smart appliances and sensors on indoor power lines with smart meters;
Neighborhood area networks (NAN) that
connect smart meters to data concentrators that are deployed by local utilities on medium-voltage (MV) lines (in the US) or low-voltage (LV) lines (in Europe);
Communication backhaul that carries trac between data concentrators and local utilities. 

BERavg performance is more sensitive to the accuracy of SNRinst estimates (given same SNRavg for both links)
combined average BER that is better than the individual average BER of each link over the whole Eb/No range is the case where the pilot spacing is equal to one, i.e. all the subchannels are pilots.
(2.3 Testbed)
( Current Work )
A real-time hardware/software testbed for wired MIMO OFDM communication
Algorithms evaluated: bit allocation, time equalization, far-end crosstalk cancellation ( zero-forcing & successive interference)
 Add simultaneous wireless communication to the PLC testbed
 Evaluate communication performance vs. complexity tradeoffs for wireless diversity methods
Project Web Page:
Software
Hardware
Graphical User Interface (GUI) runs in LabVIEW
Real-time operating system running on embedded processors:
NI PXI 1045 Embedded Controller
NI PXI-5122 for analog-to-digital (A/D) conversion
NI PXI-5421 for digital-to-analog (D/A) conversion