ELEC2590: Sminaires dElectronique et Tlcommunications

1
ELEC2590 Séminaires dElectronique et
Télécommunications
Séance III Réseaux dantennes destinés à la
localisation 15 février 2005
Christophe Craeye, UCL/TELE
2
Introduction

• Antenna arrays
• Spatial resolution beamforming and diversity
• Initially military technology.
• Civilian applications
• remote sensing
• security,
• assets management,
• radio-astronomy,
• satellite communications,
• multiple-antennas wireless communications.
• Made possible thankds to progress in microwave
  technology, electronics, signal processing in
  particular real-time processing.
• Challenges low-cost wideband arrays,
  low-profile structures, elimination of multipath.

3
Outline

• Short review of antenna and array theory
• Wideband arrays active element patterns and
  noise models
• Multiple-antenna receiver with digital
  beamforming
• Adaptive beamforming Minimum variance and MUSIC
  algorithms
• Numerical demonstration aircraft tracking

4
Far fields of an antenna
û
r
Zg
Source series impedance
Vt
Current density
2 p / l
Pattern
Multiply by to
obtain the field
5
The receiving antenna
û
r
Zg
Ir
Amplifier
Reciprocity the current obtained on receive is
proportional to the field transmitted in the
direction and polarisation of interest.
6
Antenna arrays examples
7
Antenna arrays
Multiplication of radiation patterns
Hyp the element patterns are all identical !!!
8
Linear array with equi-spaced elements
Grating lobe
Array pattern
Element pattern
9
Pattern in (u,v) space for 3X3 array with l/2
spacing
u sin q cos f
v sin q sin f
v
Visible space
u
10
Dual-pol wideband array
11
Element patterns in a 180-elements array
V
H
Frequency 0.5 GHz
12
Element (5,5)
H
V
T
Infinite array
Exact
Infinite corrected with finite-by-infinite
13
Element (1,1)
H
V
T
14
Noise model one amplifier
To be multiplied by weight w in beamformer
Antenna
Noise sources
Signal quadripole
15
Noise model coupled array
Array impedance matrix
16
Basic analog chain
90
I
Q
I
Q
17
I and Q signals and Rice components
Analytic signal

(narrow-band signal)
Complex envelope
(Hilbert transform)
Rice components
IX2
QX2
18
Basic signals
y1
y2
y3
y4
Steering vector associated with direction û
Signals
19
Minimum variance method
Goal for a given observation direction û, ensure
minimum output power
While garanteeing a given output for a unit
signal from -û
min
Définition correlation matrix
20
Optimization with the help of Lagrange
multiplyers
Solution
Average power
21
Reminder eigenvalues
Matrix with eigenvectors (orthogonal if A
hermitian)
Diagonalisation
22
Decomposition into quadratic forms
23
Signal and noise subspaces
NS sources gt NS main eigenvalues, M-NS noise
eigenvalues. Example 2 non-coherent sources
Steering vectors in subspace subtended by eigen
vectors (steering vector ? eigenvectors)
24
Minimum variance and MUSIC algorithms
Easier to minimize
25
Conclusions and perspectives
Digital technology opened a wide range of new
possibilities, at affordable cost for new
civilian applications.
Big efforts are still necessary for low-cost
easily integretable wideband low-noise RF front
ends. Introduction of the noise-coupling issue in
the analysis techniques.
On the signal processing side, need to improve
wideband schemes and multipath management. More
work necessary in the introduction of a priori
information (like calibration data) in initially
blind dectection schemes.