Symbol Shaping for Barker Spread Wi-Fi Communications

1
Symbol Shaping for Barker Spread Wi-Fi
Communications

• Tanim M. Taher
• Matthew J. Misurac
• Donald R. Ucci
• Joseph L. LoCicero

Presented by Tanim M. Taher
2
Outline

• Background
• Motivation
• Simulation and Experimental Methodology
• Pulse Shapes Used and Results
• Performance Results
• Line Coding
• Conclusions

3
Spectral Mask Background

• The Industrial, Scientific and Medical (ISM)
  bands are overcrowded.
• The Federal Communications Commission (FCC)
  limits the output power to 1 watt.
• FCC regulates out-of-band Power in Wi-Fi systems
  using a rigid Spectral Mask.
• Most modulation schemes require high order
  filters to achieve spectral mask.

4
IEEE 802.11 Spectral Mask
5
Why Pulse Shaping?

• Filters add to hardware cost, and introduce
  Inter-Symbol-Interference (ISI) that lowers the
  Bit Error Rate (BER) vs Signal to Noise Ratio
  (SNR) performance.
• Shaping the transmitted symbols as opposed to
  filtering prevents ISI, while lowering
  out-of-band interference power.

6
The Barker spread 1 Mbps 802.11 signal

• Access Points and laptops use a spreading code
  called the 11-chip Barker to expand the bandwidth
  of 1 Mbps data signals.
• The spread spectrum system more robust to noise,
  multi-path fading, and narrowband interference.
• This 1 Mbps communication system is used for
  transmitting all the Packet headers and Physical
  Layer Convergence Protocols.
• At higher noise levels, this system is used for
  transferring all data.

7
More about the Barker Sequence

• The Barker chip sequence used in the 802.11
  standard is
• B 1,-1,1,1,-1,1,1,1,-1,-1,-1
• where rectangular pulses are used to represent
  each chip (polarities varied according to B) in
  the sequence.
• The Barker sequence (B) has very good
  auto-correlation properties and this is what
  minimizes multipath effects.

8
Motivation for Project

• The problem is that the Barker spread data
  waveform does not adhere to the spectral mask.
• The rectangular Barker waveform was modified by
  pulse shaping to achieve better spectral
  performance in relation to the spectral mask.
• The resulting modulation system was studied by
  simulation and experimentation. The PSD and BER
  performance were examined.

Simulated PSD of rectangular unfiltered
rectangular pulse Barker waveform.
9
MATLAB Simulation Methodology for each Pulse
Shape
Generate random bit sequence and spread each bit
by pulse shape to obtain data waveform.
1) Design Pulse Shape adhering to Barker
Sequence. 2) Examine its Auto-correlation
properties.
Obtain the PSD of data waveform using the Welch
method.
Add Additive White Gaussian Noise (AWGN).
10010110111010
Use Correlator to obtain timing information from
the received signal
Examine Bit Error Rate
Use Correlator to decode the received bits.
10010110101010
10
Experimental Emulation

• Comblock Devices were used to transmit and
  receive the waveforms experimentally. MATLAB
  software was used to do the coding and decoding
  in a workstation.

The Comblock transmitter
The Comblock receiver.
11
Experimentation Methodology for each Pulse Shape
Generate random bit sequence and spread each bit
by pulse shape to obtain data waveform.
Upload the data waveform to the Comblock
transmitter.
Design Pulse Shape adhering to Barker Sequence in
Matlab.
Transmit over the Air.
10010110111010
Comblock receiver captures the received data
waveform for computer download.
Examine Bit Error Rate
Use Correlator to decode the received bits.
Use Correlator to obtain timing information
10010110101010
12
Logarithmic Symbol Shape

• Practical devices to inexpensively generate these
  symbols can be manufactured using discrete-time
  analog memory devices (like Pulse Amplitude
  Modulation, PAM, chips)

13
Sinc Symbol Shape
14
Sinusoidal Symbol Shape
15
More of the Sinusoidal Symbol Shape
Simulated PSD
Experimental PSD
16
More Sinusoidal Material
Oscilloscope plot of Experimental Data Waveform
Time Auto-correlation
17
Performance Results

• Bit Error Rate (BER) dropped as the filter order
  dropped.

Table 2. Experimental BER measurements at
receiver-to-transmitter distance of 1 meter.
Table. 1. Simulated BER measurements.
Pulse Shape Used Filter Order Bit Error Rate at SNR levels Bit Error Rate at SNR levels Bit Error Rate at SNR levels
Pulse Shape Used Filter Order 11.5 dB 11 dB 10 dB
Rectangular 5 3.70E-03 2.74E-03 9.00E-04
Logarithmic 3 2.48E-03 1.40E-03 5.60E-04
Sinusoidal 2 2.62E-03 1.36E-03 3.80E-04
Sinc-function 2 2.80E-03 1.98E-03 3.80E-04
Pulse Shape Used Experimental BER
Rectangular 9.99E-03
Logarithmic 6.22E-03
Sinusoidal 3.71E-03
Sinc-function 5.84E-03
18
Line Coding and Pulse Shaping

• A modified Barker spread system was examined that
  buffered 2 bits.
• A line coding involving a total of 8 pulse shapes
  was developed and tested.
• The idea was to eliminate discontinuities that
  arises when a bit transition occurs from 1 to 0
  or vice versa.
• The system buffers 3 bits in order to eliminate
  discontinuities and by selecting the appropriate
  pulse shape (from pool of 8) to transmit.

19
Plots of 3 bit buffer system.
1 to 1 bit transition
-1 to 1 to -1 bit transition
1 bit 1 -1 bit 0

• However, the results showed no significant
  spectral improvement.
• The unbuffered sinusoidal system gives best BER
  performance.

20
Conclusions

• Pulse Shaping was thoroughly applied to 802.11
  Barker Spread Signal.
• Complete simulation and experimental studies were
  performed to examine performance. Analytical
  study was performed for Sinusoidal pulse shape.
• The spectral performance was improved and BER
  reduced.
• Future Work will look at 802.11 CCK signals.

21
Thank you!Questions?Tanim Taher
(tahetan_at_iit.edu)