Efficient Diversity Technique for Hybrid Narrowband-Powerline/Wireless Smart Grid Communications Mostafa Sayed, and Naofal Al-Dhahir University of Texas at Dallas Ghadi Sebaali, and Brian L. Evans, University of Texas at Austin
Smart Grid Communications Data Concentrator Smart Meter Focus: neighborhood-area smart utility network between a data concentrator and smart meters along two paths: 1) Low-voltage power lines in 3-500 khz band 2) Unlicensed 902-928 MHz wireless band
PLC/Wireless Diversity 902-928 MHz Modulator Encoder Input DAC UP- 3-500 KHz PLC Channel Down- Combiner Soft Output Decoder M Smart Meter D Data Concentrator Simultaneous PLC/wireless transmissions using low-voltage power lines in 3-500 khz band and unlicensed 902-928 MHz wireless band Goal: Improve reliability of smart grid communications using PLC/wireless receive diversity combining methods
Symmetric Diversity Combining Old: Combining of two wireless links Receiver Transmitter Down- g 1 X Maximal Ratio Combining (MRC) Output Modulator DAC UP- Soft Decoder Encoder Down- X Input g 2 Same channel, noise, and interference statistics Same Average SNR
Asymmetric Diversity Combining New : PLC/Wireless combining for Smart Grid Comm. 902-928 MHz Modulator Encoder Input DAC UP- 3-500 KHz PLC Channel Down- Combiner Soft Output Decoder M Smart Meter D Data Concentrator Different channel, noise and interference statistics PLC and wireless might have different average SNR!
Noise Models Cyclo-stationary noise for PLC Period is half AC cycle ~8.3 ms Modeled as colored Gaussian Spectrum varies with time Asynchronous noise in unlicensed wireless bands Modeled as Gaussian mixture PDF ~ α " G(0,σ " % )+α & G(0,σ & % ) One period, 3 regions R1 R2R3 0.2 Gaussian mixture Noise 0 0.2 0 5000 10000 Samples
Applying Conventional MRC Receiver Transmitter Down- g 1 X Maximal Ratio Combining (MRC) Output Modulator DAC UP- Soft Decoder Encoder Down- X Input g 2 Maximal Ratio Combining (MRC) is a maximum likelihoodoptimal technique for white Gaussian noise The log-likelihood (LL) function of MRC is given by 0 0 0 0 0 0 0 LL X / = log p Y *,/ H *,/ X / p Y,,/ H,,/ X / = ; > <,=? > <,= @ > = A B A < ; > C,=? > C,= @ > = A BA C l is the block index and k is the sub-channel index σ * % and σ, % denote average noise powers for PLC and wireless 7
Motivation In symmetric combining, average noise powers σ * % and σ, % are equal, For asymmetric PLC/Wireless combining, the average noise powers σ * % and σ, % are not necessarily equal The instantaneous noise power level on both links shows rapid variations over both time and frequency The instantaneous noise powers have a high peak-toaverage ratio, which is higher on PLC than on wireless link 8
Impulsive Noise in PLC and Wireless Noise power over frequency sub-channels across multiple blocks PLC PAR = 21 db Wireless PAR = 14 db AWGN PAR = 10 db
Proposed PLC/Wireless Combining For PLC/wireless combining, average noise powers σ % * and σ %, don t capture the impulsive noise variations in the PLC and unlicensed wireless links We compare three PLC/Wireless combining metrics Average SNR (noise power averaged over both time and frequency) Instantaneous SNR (no averaging) Noise PSD (noise power time-average per sub-channel) 10
Proposed PLC/Wireless Combining The log-likelihood (LL) function for the instantaneous SNR and the PSD combining can be expressed as 0 LL FGHI X / = Y *,/ 0 0 H *,/ X 0 / % % Y 0 0,,/ H,,/ X 0 / % % LL JKL X / 0 = Y 0 *,/ σd *,0/ 0 H *,/ X 0 / % % Y 0,,/ σe *,0/ σd,,0/ 0 H,,/ % σe,,0/ X / 0 % Where σd 0/ % and σe 0/ % represent the instantaneous noise power and the average noise power per sub-channel (or the noise PSD), respectively σe 0/ % depends on l, the block index, as each block might belong to a different noise region
Proposed PLC/Wireless Combining PSD Combining Wireless link 902-928 MHz 3-500 KHz PLC Channel Down- PSD Estimation PSD Estimation Inverse Soft (LLRs) Soft (LLRs) Inverse X (PSD Inverse) Average SNR per subchannel X Combined Soft (PSD Inverse) Average SNR per subchannel Output Decoder Instantaneous SNR Combining Wireless link 902-928 MHz 3-500 KHz PLC Channel Down- Instantaneous Noise Power Estimation Instantaneous Noise Power Estimation Inverse Soft (LLRs) Soft (LLRs) Inverse X X Instantaneous SNR Instantaneous SNR Combined Soft Output Decoder
Instantaneous Noise Power Estimation As a simple technique to estimate the instantaneous noise power, we employ comb-type pilots inserted periodically within the data symbols We estimate the noise power in the pilot locations followed by linear interpolation to compute estimates over all symbols Received Pilots Stored Pilots + - 2 Linear Interpolation ^ P ( k ) Pilot Spacing Pilot Spacing Frequency
Noise PSD Estimation The noise PSD can be estimated by averaging the received signal power σe 0/ % = E Z / 0 % = E Y / 0 % E H / 0 % Estimated PSD and actual PSD vs the active sub-channel indices (36 sub-channels in the CENELEC A band [35-91]kHz PSD estimate 0.18 0.16 0.14 0.12 0.1 0.08 0.06 Actual PSD Estimated PSD Averaging is performed over 512 Symbols 0.04 0.02 0 0 5 10 15 20 25 30 35 40 Active Subchannel Index
Combining Metrics Comparison Noise Power Ratio = PLC Noise Power/ Wireless Noise Power 4 Average Noise Power Power Spectral Density Instantaneous Noise Power 3.5 3 Noise Power Ratio 2.5 2 1.5 1 0.5 0 0 20 40 60 80 100 120 140 160 180 Symbol Index One Block 36 Active Sub-Channels out of 256 Noise over frequency sub-channels across multiple blocks 15
Simulation Parameters transmission with 256 sub-channels and BPSK modulation. 0.4 MHz sampling rate. CENELEC-A frequency band (35 khz to 91 khz). Rate 1/2 Convolutional Coding Wireless Link Noise Model: GM with two states α " = 0.98, α & = 0.02, σ " % = 0 db, and σ & % = 20 db PLC Link Noise Model: Region 1 Region 2 Region 3 Time Percentage 60 % 30 % 10 % Power (db) -6.59 1.93 5.15 16
Performance Results Average BER vs Eb/No of both links (equal Eb/No) - PLC channel from measurements/rayleigh channel for wireless 10 0 Average BER 10-1 10-2 10-3 10-4 10-5 10-6 10-7 Wireless PLC Avg SNR Combining PSD Combining Inst SNR Combining -12-10 -8-6 -4-2 0 2 4 6 8 10 EbNo (db) Performance Gain over one link only at 10 X3 Average SNR Combining 4 db PSD Combining 5.5 db Instantaneous SNR Combining 7 db
Performance Results Average BER vs Eb/No of the PLC link at Eb/No = 2 db for the wireless link 10 0 Average BER 10-1 10-2 10-3 10-4 10-5 10-6 10-7 PLC Wireless Avg SNR Combining PSD Combining Inst SNR Combining -12-10 -8-6 -4-2 0 2 4 6 8 10 EbNo (db) Performance Gain over PLC only at 10 X3 Average SNR Combining 4.5 db PSD Combining 8 db Instantaneous SNR Combining 10 db
Performance Results Average BER vs Eb/No of the Wireless link for Eb/No = 0 db for the PLC link 10 0 Average BER 10-1 10-2 10-3 10-4 10-5 10-6 Wireless PLC Avg SNR Combining PSD Combining Inst SNR Combining -12-10 -8-6 -4-2 0 2 4 6 8 EbNo (db) Performance Gain over Wireless only at 10 X3 Average SNR Combining 2.5 db PSD Combining 4.5 db Instantaneous SNR Combining 8 db
Conclusion PSD combining provides the best performance /complexity tradeoff - better performance than average-snr combining at lower complexity than instantaneous-snr combining Our proposed PSD estimation method does not require pilot overhead while instantaneous-snr combining requires high pilot overhead (resulting in data rate loss) 20
Back-up Slides 150 Impulse Response 100 50 Amplitude (v) 0 50 100 150 200 10 20 30 40 50 60 Time (us) 1 2 3 4 5 6 7 21
Back-up Slides 60 Frequency Response 50 40 Magnitude response (db) 30 20 10 1 2 3 0 4 5 6 7 10 0 50 100 150 200 250 300 350 400 450 500 Frequency (khz) 22