Sparse Preambles for Enhanced SOP Separation

1
Project IEEE P802.15 Working Group for Wireless
Personal Area Networks (WPANs) Submission Title
Sparse Preambles for Enhanced SOP
Separation Date Submitted November 10,
2003 Source Yossi Erlich Company Infineon
Technologies Address P.O.Box 8631, Poleg
Industrial Area, Netanya 42504,
Israel Voice972-9-8924100, FAX
972-9-8658756, E-MailYossi.Erlich_at_infineon.com
Re Abstract A new set of preambles are
suggested for the MB-OFDM. The set enhances the
preamble detection performance under SOP, and
reduces its complexity. Purpose Technical
contribution to the MB-OFDM proposal Notice This
document has been prepared to assist the IEEE
P802.15. It is offered as a basis for discussion
and is not binding on the contributing
individual(s) or organization(s). The material in
this document is subject to change in form and
content after further study. The contributor(s)
reserve(s) the right to add, amend or withdraw
material contained herein. Release The
contributor acknowledges and accepts that this
contribution becomes the property of IEEE and may
be made publicly available by P802.15.
2
Sparse Preambles for Enhanced SOP Separation
3
Drawbacks of Currently Proposed Sequences

• The cross-correlation functions have high peaks
• Increased false alarm, i.e., detecting preamble
  of other channels (piconets) when the channel of
  interest does not exists
• Prevent the detection of the channel of interest
  when two or more SOP exists
• The preamble auto-correlation has a wide main
  lobe
• Results in increased decision threshold, which in
  turn reduces probability of detection.

4
Solution

• It is widely accepted that there are no binary
  sequences that have
• Good cross-correlation properties
• Good auto-correlation properties
• Low complexity

Solution Use Ternary Sequences
5
Binary and Ternary Sequences

• Binary sequences use 1,-1 as their elements
• For example the sequence 1,1,-1,1,-1,1,1,1,-1,-1,
  
• The complexity of correlating the received signal
  with an hierarchy binary sequence is
  proportional to the summation of their respective
  hierarchy lengths. Currently 24 (168).
• Ternary sequences use 1,0,-1 as their elements
• For example the sequence 1,1,-1,0,0,1,-1,-1,0,0,0
  ,1,0,
• The complexity of correlating the received signal
  with a ternary sequence is proportional to the
  number of non-zero elements in the sequences

6
Sequences Construction

• Design sequences with non-equal number zeros
  between any two non-zero elements results in
  sequences with
• Excellent cross-correlation function between
  sequences
• Narrow main lobe in the auto-correlation function
• Flattening the spectrum can be achieved by virtue
  of Sonys contribution (IEEE 802.15-03-0337/r1)

7
The Proposed Sequences
The proposed sequences have the following
canonical structure
Where 0n represents a sequence of n zeros. The 14
non-zero elements corresponding to the four
preambles are
s0 s1 s2 s3 s4 s5 s6 s7 s8 s9 S10 s11 s12 s13
1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1
2 1 1 1 1 -1 -1 -1 -1 -1 -1 1 1 -1 1
3 1 -1 -1 1 1 1 -1 -1 1 1 -1 -1 -1 1
4 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1
Complexity The correlator requires 13 additions
(14 non-zero elements) instead of the 23
additions (hierarchy 168) required by the
current sequences. Low Complexity!
8
The Proposed Vs. Current Preambles Example
The original sequence
The sequence after applying spectrum flattening
9
The Auto-Correlation Function
The proposed sequence after applying spectrum
flattening
The current sequence after applying spectrum
flattening

• The side peaks of the auto correlation functions
  are practically the same
• However the main lobe is narrower in the proposed
  sequences

10
The Cross-Correlation Functions
proposed
Unwanted peaks for the current sequences
current
11
Simulations Signal Only

• Consider a sequence detector that searches
  sequence A within 1mSec of a single band, tuned
  for 10-3 false-alarm, designed for worst channel
  model
• Assume that only sequence A was transmitted (no
  other preambles from other channels)

12
Simulations Interference Only

• Assume that the receiver tries to detect channel
  A, and channel B or C preamble is transmitted
• The following figure shows the squared correlator
  output when Channel C preamble is transmitted,
  CM3

Peak due to the high cross-correlation in the
current preambles
Proposed preamble
Current Preamble
13
Simulations Interference Only

• The following figure depicts the probability of
  detecting wrong channel as channel A using a
  correlation based detector

14
Simulations Signal Interference

• Assume that the receiver tries to detect Channel
  A and both Channel A and C preambles are
  transmitted. The following figure shows the
  squared correlation signal, CM2, SIR-5dB

Current preambles
Interference
Required Signal
Proposed preambles
15
Simulations Signal Interference

• The following figure shows the probability to
  detect Channel A in the presence of Channel B or
  C as an Interference
• Wrong decision is made when the maximum of
  windowed squared correlator output corresponds to
  an interfering signal

16
Summary and Conclusions

• We propose a family of new preambles based on
  ternary sequences
• The proposed preambles offer additional 6dB
  resistance against interference
• The detection complexity of the proposed
  preambles is lower than the complexity of
  detecting the current ones
• The proposed preambles improve the SOP
  performance by enabling acquisition even in the
  presence of strong interferences