2-Bordoni_IGARSS11_A..
Download
Report
Transcript 2-Bordoni_IGARSS11_A..
Ambiguity Suppression by Azimuth Phase
Coding in Multichannel SAR Systems
F. Bordoni, M. Younis, G. Krieger
DLR - Institut für Hochfrequenztechnik und Radarsysteme
IGARSS 2011, 24-29 July, Vancouver,Folie
Canada
1
Outline
o
Introduction
o
APC (Azimuth Phase Coding) technique
o
APC in multichannel SAR (Synthetic Aperture Radar) systems
o
Figure of merit
o
Numerical analysis
o
o
APC performance versus system parameters
o
Example: two multichannel systems for high resolution wide swath imaging
Conclusions
Microwaves and Radar Institute
2
Introduction
Current spaceborne SAR systems limitation:
trade-off spatial resolution v.s. swath width
Research in two main directions:
New, more flexible SAR systems
- Multichannel systems
-Digital Beamforming (DBF) on receive
- Multichannel processing
Processing methods for
removing the ambiguities
APC
- low implementation complexity
- effectiveness for
point and distributed ambiguities
APC is conceived for conventional SAR systems:
APC in multichannel systems based on DBF on receive?
Microwaves and Radar Institute
3
Review of the APC Technique
APC is a technique for range ambiguity suppression, conceived for
conventional (1 Tx and 1 Rx) SAR systems [Dall, Kusk 2004]
[Dal04] J. Dall, A. Kusk, “Azimuth Phase Coding for Range Ambiguity Suppression in SAR”, IGARSS 2004.
APC is based on three main steps:
1) Azimuth, i.e. pulse to pulse, phase modulation on Tx
APC modulation phase
mod (l )
Tx pulse number
2) Azimuth phase demodulation on Rx
APC demodulation phase
dem (n) mod (n m) @ round-trip delay
APC residual phase
res (n, k , M )
azimuth sample number, order of range ambiguity,
APC shift-factor
3) Azimuth filtering over the processing bandwidth
Microwaves and Radar Institute
4
APC residual phase Doppler shift
Time domain: linear phase
Frequency domain: Doppler shift
xk ,apc (n) xk (n) exp jres (n, k , M )
X k ,apc ( f ) X k ( f f )
2
kn
M
order of range ambiguity (0 useful signal)
PRF
f f (k , M ) mod k
M PRF / 2
res (n, k , M )
x
+
0
1
x
…
2
k
M
…
k
x
M
=
…
1
n
(t n / PRF )
Az. FILTER
x
1
f (2, M)
k=2
k=
…
+
2
M
0
k
=
k=1
2
2
Az. FILTER
res
k=0
f (1, M)
0
PRF
M
2
PRF
M
f
Bp
PRF
M=2 maximum Doppler shift of the 1st order range ambiguity
Larger oversampling PRF / Bp Larger ambiguity suppression
Microwaves and Radar Institute
5
Application to Multichannel Systems
Multichannel SAR system: 1 transmitter, N receivers
1
N
2
N Rx az. signals sampled at PRF
PRF << Bp
mod (l, M )
dem (n, M )
dem (n, M )
APC residual phase: res (n, k, M )
X k ,apc ( f ) X k f f (k )
X kr,apc ( f )
MULTICHANNEL PROCESSING
reconstructed multichannel signal sampled at PRFeff =N PRF:
X kr,apc ( f )
mc mc
APC residual phase: res (n , k , M )
The behavior of the APC changes when applied to a multichannel system
Microwaves and Radar Institute
6
APC & Reconstructed Multichannel Signal
The APC residual phase has no more a linear trend versus the azimuth
sample (pulse) number no shift of the Spectrum
(uniform PRF*)
mc
res
N M 2
2
x x
,
,
0
x x
0
PRF
x k= 1, 3, …
x x
nmc
2
4
t n
mc
/( N PRF )
0
f
Bp
PRF
PRFeff = 2 PRF
mc
2
n
k, M )
k int
The residual phase a “stair” shape (<≠> Doppler shift):
M
N
r
r
mc
The ambiguity spectrum: X k ,apc ( f ) X k ( f ) FT exp jres (n, k , M )
mc mc
res
(n ,
*PRF matched to the antenna length and No. of apertures > regular sampling in azimuth results
Microwaves and Radar Institute
7
Figure of Merit
Measurement of the ambiguity suppression induced by APC
APC Gain:
Computed on the SAR signal after multichannel processing
PSD (Power Spectral Density) range ambiguity of 1st order
if APC is not applied
Bp / 2
2
r
X
(
f
)
df
1
B
/
2
p
Gapc
Bp / 2
2
r
X
( f ) df
1,apc
processed bandwidth
B
/
2
p
r
X 1r ( f ) X 0,apc
(f)
useful signal after multichannel
reconstruction (neglect. elev.)
PSD range ambiguity of 1st order if APC is applied
Note: the Gapc depends on the azimuth pattern shape
Microwaves and Radar Institute
8
APC Performance Analysis
Reference Multichannel Planar Systems
Parameter
System #
1
2
3(Ref.)
Orbit height [km]
4
520
Carrier frequency [GHz]
9.600
Rx antenna total length [m]
3
6
12
Tx antenna length [m]
(and Rx subapert. length)
24
3
Bp
No. of az. Rx channels
1
2
4
8
PRF [Hz]
5068
2534
1267
633.5
(uniform)
PRFeff [Hz]
The systems have the
same azimuth patterns
PRFeff
5068
Processing bandwidth 2316 Hz ≤ Bp ≤ 4168 Hz
Investigation:
Behavior of APC versus the number of Rx channels, N
Effect of the Doppler oversampling N PRF Bp
The effect of the pattern shape is not evident
Microwaves and Radar Institute
9
Numerical Results: Gapc
APC Gain v.s. oversampling factor
N=1
N=2
N=4
N=8
For the considered systems, for M=2:
0.1dB ≤ Gapc ≤ 3.13dB
for a given N, the Gapc increases with the oversampling factor,
the Gapc decreases for increasing number of channels, N
the sensitivity of Gapc to decreases with increasing N
Microwaves and Radar Institute
10
Numerical Results: PSD v.s. N
Normalized PSD 1st range ambiguity after multichannel reconstruction
N = 1, 2, 8
without APC
The thickness of the curves is a
fast variation of the spectrum,
due to aliasing
Bp
with APC
N=1
Bp
N=2
Bp
N=8
Bp
larger N, the upper profile PSD with or without APC are similar and Gapc reduces
Microwaves and Radar Institute
11
HRWS SAR Multichannel Systems
HRWS (High-Resolution Wide-Swath) SAR System
promoted by the German Aerospace Centre (DLR)
conceived to obtain high resolution and wide swaths
Parameter
Planar
Reflector
Orbit height [km]
520
745
Carrier frequency [GHz]
9.600
9.650
Tx/Rx antenna total length [m]
8.75
Paraboloid diameter (elev., az.) [m]
10, 12
Total number of feeds (elev., az.)
60, 10
No. of az. Rx channels
7
10
PRF [Hz]
1750
2792
Processed bandwidth [Hz]
6252
5946
Oversampling factor
1.960
4.696
Planar system:
currently adopted design
Reflector system:
alternative design option,
studied in DLR
(1 m resolution, 70 km swath width in stripmap mode)
Different Rx azimuth patterns & multichannel reconstruction
Microwaves and Radar Institute
12
Peculiarities HRWS Systems
Planar system
Reflector system
Bp
Bp
Be
PRF
Be Bp / N
N PRF
The pattern of each Rx channel covers Bp
The pattern of each Rx channel covers 1/N of Bp
Multichannel processing: Multi-Aperture Reconstr.
Multichannel processing: Spectral decomposition
The patters do not change along the swath
The patters change along the swath
Evidence of the dependence of the APC performance on the pattern shape
Microwaves and Radar Institute
13
Numerical Results: Planar HRWS System
Normalized PSD 1st range ambiguity used to compute the Gapc
(after multichannel reconstruction)
without APC
with APC
Bp
Bp
For M=2, Gapc = 0.69 dB
The high number of channels (7) and the small oversampling (1.96) associated low Gapc
Microwaves and Radar Institute
14
Numerical Results: Reflector HRWS System
Normalized PSD 1st range ambiguity used to compute the Gapc
(before multichannel reconstruction, single Rx channel)
without APC
with APC
Be
Be
For M=2, 3.2 dB ≤ Gapc ≤ 8.6 dB over the swath, depending on the azimuth pattern
The azimuth pattern strongly affects the APC performance
The reflector based system, characterized by a higher oversampling factor (4), takes
better advantage from the application of APC
Microwaves and Radar Institute
15
Conclusions
o
In multichannel systems, the APC effect is no more a frequency shift of
the range ambiguity.
o
Also in multichannel systems, the APC allows for improved ambiguity
suppression.
o
The azimuth pattern strongly affects the APC performance.
o
For a given azimuth pattern, the suppression is directly proportional to
the oversampling factor and inversely proportional to the number of
receive channels.
o
In a conventional SAR system with = 2, the achievable suppression of
each ambiguity of odd order is about 3 dB. In multichannel systems
based on planar antenna architectures, the suppression is generally
poorer.
Reflector based systems reach better performance, because of the higher
oversampling.
o
In the planar and reflector based HRWS systems the APC suppression is
about 0.7 dB and between 3 and 8 dB, respectively.
Microwaves and Radar Institute
16