kjli

1
UWB Synchronization

• Hui-Min Yeh

2
Outline

• Introduction
• Code synchronization
• Typical acquisition scheme
• Typical search strategies
• Definition of Hit set
• Transmitter design
• Hybrid TH/DS signal format
• Receiver design
• Proposed search algorithms
• Joint BRS-NCS search
• Joint RPS-NCS search
• Proposed acquisition scheme
• Simulation results
• Conclusions

3
Synchronization ?
4
Introduction

• Code synchronization (TH code DS code)
• To enable the lower BER of data demodulation, the
  template signal should be aligned with the
  received signal, the code alignment is so-called
  code synchronization process.
• Code synchronization can be spilt into two parts
• Code acquisition
• It is a coarse code acquisition, to resolve the
  code phase error within certain range.
• Code tracking
• It is a fine tuning process, to guarantee the
  timing error below an acceptable level.

5

• Acquisition strategy The search for acquisition
  is based on the auto-correlation properties of
  the applied the codes, the auto-correlation is
  high if the receiver is synchronized and low in
  other situations 5.
• MAX criterion (Maximum selective)
• TC criterion (Threshold Crossing)
• MAX/TC criterion

6

• Before any data can be received in an UWB
  communication, the receiver must synchronize on
  the transmission.
• Receiver structures
• Noise template signal
• Reference template signal
• Packet data
• Synchronization on (known) code
• Reception and processing data

Data can have specific coding-gt communication can
not be tapped
7
Code synchronization

• Find the best fit between the code and the
  received signal

8
Typical acquisition system

• The receivers reference signal generator will
  shift and guess the code boundary of received
  signal, and the AWGN noise will effect the
  performance.

Correlator
4 5
3
9

• Why use RAKE receiver?
• More than one hypothesized phase can be
  considered as an estimated timing delay for a
  coarse acquisition

10
Definition of Hit set

• Hit set definition
• The definition of Hit set is a critical issue .
  When the threshold setting is low, the noise
  effect on the false alarm probability will
  increase. Oppositely, when the threshold setting
  is high, the mean acquisition time will increase.
  

11

• Ex The periodicity of TH code 4, frame duration
  , that is Nh128, Tc 1
  nsec, normalized pulse energy 1,CM1(noise free),
  
• search space 512 cells, Th2, EGC receiver
  (20 fingers)

Peak value
Partial correlation energy
Hit set
N
1
2
True phase
12

• Rake fingers 20, search space 512 cells, CM1

13
Search strategies

• What is the efficient search algorithm?
• Serial or Parallel
• Typical serial search algorithms 1
• Linear Search (LS)
• Random Permutation Search (RPS)
• Bit Reversal Search (BRS)
• Ideal mean stopping time

Linear Search (LS)
Random Permutation Search (RPS)
Bit Reversal Search (BRS)
14

• Bit Reversal Search
• The search algorithm is described by assuming
  that N is a power of 2.
• The order of search positions in the bit reversal
  search
• algorithm is decided by bit reversing
• For example Let the integer of N is equal to 8
  (23)

Search locations for N 23
Decimal 0 1 2 3 4 5 6 7
Binary 000 001 010 011 100 101 110 111

Bit Reversing
Reorder search locations by bit reversing
Decimal 0 4 2 6 1 5 3 7
Binary 000 100 010 110 001 101 011 111
15
The ideal and normalized mean stopping time vs.
the parameter H/N for three serial search
algorithms (N512 cells)
16

• Bit Reversal Search algorithm , code length4,
  Nh128,
• CM1(noise free), search space 512 cells,
  EGC receiver (20 fingers)

17
Outline

• Introduction
• Code synchronization
• Typical acquisition scheme
• Hit definition
• Typical search strategies
• Transmitter design
• Hybrid TH/DS signal format
• Receiver design
• Proposed search algorithms
• Joint BRS-NCS search
• Joint RPS-NCS search
• Proposed acquisition scheme
• Simulation result
• Conclusions

18
Transmitter design Hybrid TH/DS signal format
Next symbol
1s Symbol (NdsTf)
(TH-sequence)
(Repeat NT times)
Frame (CNth-1)
Frame (C0)
Frame (C1)
Tf
19

• The variation of power spectral density (PSD)
  between the pure time hopping sequences and
  hybrid TH/DS signals
• No TH coding
• The periodicity of TH code is equal to
  the number of pulses per symbol
• Hybrid TH/DS signal
• The hybrid TH/DS signal format method not only
  speeds up the code acquisition, but also improves
  the interference to other co-existing systems 7.

20
UWB signals with no Time-Hopping coding

• Parameters 10 symbols number of pulses per
  symbol ( ) 64 pulse repetition time ( ) 10
  nsec chip duration ( ) 1 nsec
• The transmitter power is concentrated at
  multiples of pulse repetition frequency.

Amplitude
Time (ns)
21
UWB signals with Time-Hopping coding

• Consider the same parameters
• The periodicity of TH code ( ) is equal to the
  number of pulses per symbol, that is ,
• The PSD of the signals is composed of spectral
  lines at the distances of
• The transmitted power is concentrated at
  multiples of symbol repetition frequency.

22
UWB signals with hybrid TH/DS signals

• Consider the same parameters
• a) Short TH code with repeat NT 8
• It will increase a lot of spectral lines due to
  the repeat operation
• b) Short TH code with repeat NT 8
• The polarity of pulses per symbol is scrambles by
  DS code

23
Outline

• Introduction
• Code synchronization
• Typical acquisition scheme
• Hit definition
• Typical search strategies
• Transmitter design
• Hybrid TH/DS signal format
• Receiver design
• Proposed search algorithms
• Joint BRS-NCS search
• Joint RPS-NCS search
• Proposed acquisition scheme
• Simulation result
• Conclusions

24

• Proposed Joint Search Algorithms
• Joint BRS-NCS Search
• Joint RPS-NCS Search
• Non consecutive search (NCS) criterion
• The non consecutive search strategy is proposed
  owing to more than one HIT phase in multi-path
  channel models.
• The non consecutive search can achieve a rapid
  acquisition by testing search locations with a
  step size D which is greater than one cell.
• The cells in total search space are so
  called uncertainty region that can be divided
  into blocks
• This will reduce the search time required to
  acquire the location of hit set from an initial
  search cell, but it could also increase the miss
  probability.

25
LI-NCS (Linear Increasing NCS)

• In order to avoid the miss search, we propose a
  linear increasing NCS (LI-NCS) approach.
• The search equation can be expressed as
• The LI-NCS approach can be described as
  following
• Set initial loop 1, and check blocks. The
  number of test cells are
• If the correlation output of each block in this
  loop is lower than threshold, then go to the
  step (3). If the correlation output of one block
  in this loop is higher than threshold. then the
  search procedure is DONE.
• Set loop loop 1, and also check blocks,
  the number of cells are
• ,
  then go to step (2)

26
An example of the LI-NCS search approach

• Assume that search space N12, hit set 7 8
  9 10 11, block size D 4,

0
1
2
3
4
5
6
7
8
9
10
11
27
Joint BRS-NCS search procedure a) search space
N48, hit set H8, b) divided into 8 blocks,
each block with D 6
Search space N
Correct Symbol Boundary
(a)
New search location
D cells
(0)
(4)
(6)
(3)
(1)
(5)
(7)
(2)
(b)
28
Receiver design Acquisition scheme

• Two-stage acquisition scheme
• TH code acquisition
• DS code acquisition

Find DS-boundary (get symbol boundary)
ACQ2
(Second stage)
ACQ1
Coarse TH-boundary acquisition
Fine acquisition
Received signal
(First stage)
29
First Stage
(RAKE receiver with parallel correlators)
Hit
Z
Roughly find TH-boundary
TC 1
No Hit
Set new search location based on Joint Search
Algorithm (new shift )
Template signal generator
(Coarse Acquisition)
Maximum path amplitude selective
One branch correlator
Second Stage
(Fine Acquisition)
30
Fine Acquisition Procedure

• Step1 The terminating point of coarse
  acquisition process is the starting
  point of fine acquisition.
• Step2 The new search window in fine acquisition
  process is equal to the double Hit set size.
• Step3 The fine acquisition process searches both
  right and left directions from the starting point
  of fine acquisition with one branch correlator to
  acquire the strongest path. The search range of
  each direction is equal to the size of hit set.
• Step4 We utilize the maximum selective criterion
  to select the strongest path amplitude
• Step5 Finally, the fine acquisition process
  passes the parameter on to the second stage.

Left direction
Right direction
Starting point
31
Second Stage
(RAKE receiver with parallel correlators)
Found DS-boundary (Found symbol boundary
simultaneously)
Z
Hit
TC 2
No Hit
Set step size time-hopping code length and new
search location based on Joint Search Algorithm
location (new shift )
Template signal generator
32
Outline

• Introduction
• Code synchronization
• Typical acquisition scheme
• Hit definition
• Typical search strategies
• Transmitter design
• Hybrid TH/DS signal format
• Power spectral density (PSD)
• Receiver design
• Proposed search algorithms
• Joint BRS-NCS search
• Joint RPS-NCS search
• Proposed acquisition scheme
• Simulation results

33
Simulation cases

• Two-stage vs. one stage
• Comparison of search strategies
• Two-step approach
• MAI case

34
Channel model 14

• The first path of CM1 we consider has the
  strongest amplitude gain, and the tenth path of
  CM4 has the strongest amplitude gain in our
  simulation.

The strongest path
The strongest path
35
SNR vs. Detection probability for the second
stage in CM1

• Assume that the first stage finds the correct
  boundary.

36
Mean stopping time of two acquisition schemes as
a function of threshold parameter A

• Simulation parameters

37

• The mean stopping time of Joint BRS-NCS /Joint
  RPS-NCS method versus threshold setting for step
  size D4, D8, and D is approximated to hit set.

38

• Comparison of performance in terms of Mean
  stopping time versus threshold setting for
  proposed search algorithms and bit reversal
  search algorithm.

39
A statistical number as function of offset value
for 5000 trials, a) SNR 10 dB, coarse
acquisition result, b) considering fine
acquisition result, in CM1
Single-user
Multi-user
40
SNR 20 dB, 5000 trials, CM4
Single-user
Multi-user
41
MAI vs Detection probability (CM1)

• Simulation results for detection probability of
  the UWB channel model (CM1), as a function of the
  number of interfering users Nu-1

42
MAI vs Detection probability (CM4)

• Simulation results for detection probability of
  the UWB channel model (CM4), as a function of
  number of interfering users Nu-1

43

• Simulation results of detection probability for
  different lengths of in UWB channel model
  (CM1), as a function of interfering users (Nu-1),
  SNR20dB

44
Conclusions

• The mean stopping time of two-stage acquisition
  scheme is simulated and it is evident to see that
  the two-stage acquisition scheme outperforms the
  conventional one stage acquisition scheme.
• By simulation results, we can see that the Joint
  Search Algorithms have lower mean stopping time.
• We propose a two-step approach which is composed
  of coarse step and fine step to achieve a fine
  code acquisition in first stage.
• Adding more users to the systems is investigated
  by computer simulations.

45
References

• 1 E. A. Homier and R. A. Scholtz, Rapid
  acquisition of ultra-wideband signals in the
  dense multipath channel, Proceeding of 2002 IEEE
  Conference Ultra Wideband Sys. Tech, (Baltimore,
  MD), pp. 105-109, 2002.
• 2 J. Oh, S. Yang, and Y. Shin, A Rapid
  Acquisition Scheme for UWB Signals in indoor
  wireless channels, IEEE WCNC 04, Wireless
  Commun. and Networking Conference, vol. 2, 2004.
• 3 S. Gexici, E. Fishler, F. Kobayashi, H. V.
  Poor, and A. F. Molisch, A rapid Acquisition
  Technique for Impulse Radio, in Proc. IEEE
  Pacific Rim Conference Commun., Comput., Signal
  Process., pp. 627-630, Aug 2003.
• 4 L. Reggiani, G. M. Maggio,Rapid Search
  Algorithms for Code Acquisition in UWB Impulse
  Radio Communications, IEEE Commun. Journal, vol.
  23, pp. 898-908, May 2002.
• 5 G. E. Corazza, On the MAX/TC criterion for
  code acquisition and its application to DS-SSMA
  systems, IEEE Trans. on Commun., vol. 44, pp.
  1173-1182, Sept. 1996.
• 6 M. Z. Win, R. A. Scholtz, Impulse Radio
  How It Works, IEEE Commun. Letters, vol. 2, No.
  1, pp. 36-38, Jan. 1998.

46

• 7 S. Aedudodla, S. Vijayakumaran, T. F. Wong,
  Rapid Ultra-wideband Signal Acquisition, IEEE
  Wireless Commun. and Networking Conference, vol.
  2, pp 1148-1153, March 2004.
• 8 O. S. Shin and K. B. Lee, Utilization of
  multi-paths for spread-spectrum code acquisition
  in frequency-selective rayleigh fading channels
  IEEE Journal Trans. on Commun., Apr. 2001.
• 9 S. L. Han, S. O. Hyun, E. K. Chang, Code
  acquisition for the DS-CDMA RAKE receiver in a
  multi-path fading channel, Proceeding of IEEE
  Singapore International Conference, pp. 215-219,
  July 1995.
• 10 V. Saravanan, F. W. Tan,Equal gain
  combining for acquisition of UWB signals, MILCOM
  03, IEEE Military Commun. Conference, vol. 2,
  pp.880-885, Oct. 2003.