System, method and apparatus for recovering polarization radar data
View Patent ↗A method and apparatus that measure M coupled channels of radar polarization data ({tilde over (y)}), that is a subset, which is less than the full set of received coupled channels. Sparse recovery operations are performed on the measured M coupled channels of radar polarization data ({tilde over (y)}) and a representation of a set of more than M channels of radar polarization data is generated from the sparse recovery performed on the subset of coupled channels.
1. A method comprising:
transmitting radar signals using one or more transmit antenna;
receiving the radar signals that are reflected using one or more receive antennas that are configured to receive at least two polarizations;
measuring M coupled channels of radar polarization data from the received radar signals using a radar receiver;
performing sparse recovery operations on the measured M coupled channels of radar polarization data utilizing mixing of channel information caused by cross-channel coupling that allows a reduction in a number of polarization channels that the radar receiver needs to process; and
estimating a representation of M′ channels of radar polarization data when M′ is greater than M.
2. The method as claimed in claim 1 , where the M coupled channels of radar polarization data is a subset of the M′ channels of radar polarization data.
3. The method as claimed in claim 1 , where the M coupled channels of radar polarization data ({tilde over (y)}) is modeled by {tilde over (y)}=Ab+{tilde over (w)} where A is a matrix containing a dictionary matrix (D), b is a complex-valued data coefficient vector, and {tilde over (w)} is additive noise.
4. The method as claimed in claim 3 , where elements of the dictionary matrix (D) allow a sparse representation (b) of the M′ channels of radar polarization data.
5. The method as claimed in claim 1 , where the M coupled channels of radar polarization data ({tilde over (y)}) are comprised of radar spatial frequency domain data.
6. The method as claimed in claim 1 , where the M coupled channels of radar polarization data ({tilde over (y)}) are comprised of radar spatial domain data.
7. The method as claimed in claim 1 , where the estimate of the representation of the M′ channels of radar polarization data is a set of polarimetric radar channel data having reduced crosstalk relative to the M coupled channels of radar polarization data ({tilde over (y)}).
8. The method as claimed in claim 1 , further comprising measuring M coupled channels of radar polarization data ({tilde over (y)}) that is sub-sampled in either fast time or slow-time.
9. The method as claimed in claim 1 , where the representation of the M′ channels of radar polarization data is an estimate of the dictionary coefficients, ({circumflex over (b)}).
10. The method as claimed in claim 9 , further comprising a series of transforms to generate a second representation of the M′ channels of radar polarization data.
11. The method as claimed in claim 10 , where the transform is a dictionary matrix (D) such that the second representation is an estimate of the M′ channels of radar polarization data comprising reflectivity data.
12. The method as claimed in claim 10 , where the transform is the operator à M′ D such that the second representation is an estimate of the M′ channels of radar polarization data in a measurement domain.
13. The method as claimed in claim 1 , where the estimated representation of M′ channels of radar polarization data is used for data exploitation tasks.
14. The method as claimed in claim 1 , where the sparse recovery operations produce an estimate of a measurement matrix (A).
15. The method as claimed in claim 1 , further comprising:
estimating a clutter plus noise level from the M coupled channels of radar polarization data; and
using the clutter plus noise level to set a stopping criterion for the sparse recovery operations.
16. The method as claimed in claim 1 , where the sparse recovery includes minimization:
{circumflex over (b)} =argmin b ∥b∥ 1 s.t.∥{tilde over (y)}−Ab∥ 2 ∈,
where ∈ is non-negative.
17. The method as claimed in claim 1 , where the sparse recovery includes minimization:
{circumflex over (b)} =argmin b ∥Tb∥ u u′ s.t.∥{tilde over (y)}=Ab∥ v v′ ≤∈,
where 0≤u<2, u′ is positive, v is non-negative, v′ is positive, ∈ is non-negative, and T is a transform matrix.
18. The method as claimed in claim 1 , where the sparse recovery includes minimization:
{circumflex over (b)} =argmin b ∥{tilde over (y)}−Ab∥ v v′ s.t.∥Tb∥ u u′ ≤τ,
where 0≤u<2, u′ is positive, v is non-negative, v′ is positive, τ is non-negative, and T is a transform matrix.
19. The method as claimed in claim 1 , where the sparse recovery includes minimization:
{circumflex over (b)} =argmin b ∥{tilde over (y)}−Ab∥ v v′ +λ∥Tb∥ u u′,
where 0≤u<2, u′ is positive, v is non-negative, v′ is positive, λ is non-negative and T is a transform matrix.
20. The method as claimed in claim 1 , where the sparse recovery includes minimization:
b
^
=
arg
min
b
y
˜
-
Ab
v
v
′
+
∑
i
λ
i
T
i
b
u
i
u
i
′
where v is non-negative, v′ is positive, u i are non-negative, u i ′ are positive, λ i are non-negative, and T i are transform matrices.
21. The method as claimed in claim 1 , where the M coupled channels of radar polarization data ({tilde over (y)}) are reshaped into a matrix {tilde over (Y)} and modeled by {tilde over (Y)}=ΘB+{tilde over (W)}, where Θ is a matrix containing a dictionary matrix (D), B is a complex-valued dictionary coefficient matrix, and {tilde over (W)} is a matrix of additive noise.
22. The method of claim 1 , wherein estimating the representation of M′ channels of radar polarization data when M′ is greater than M to support a synthetic aperture radar (SAR).
23. An apparatus comprising:
one or more transmit antennas, operatively coupled to the one or more processing devices, that are configured to transmit radar signals,
receiving the radar signals that are reflected using one or more receive antennas that are configured to receive at least two polarizations;
one or more receive antennas that receive the radar signals that are reflected;
a radar receiver that measures M coupled channels of radar polarization data; and
one or more processing devices, operatively coupled to the one or more receive antennas, configured to perform sparse recovery operations on the measured M coupled channels of radar polarization data ({tilde over (y)}) and estimate a representation of M′ channels of radar polarization data when M′ is greater than M, utilizing mixing of channel information caused by cross-channel coupling that allows a reduction in a number of polarization channels that the radar receiver needs to process
wherein each of the one or more processing devices comprise a sparse recovery module having a computer processor and a physical memory operatively coupled to the computer processor.
24. The apparatus as claimed in claim 23 ,
where the number of recovered radar polarization channels is based on the number of receive antennas and the number of transmit antennas.
25. The apparatus of claim 23 , wherein the one or more transmit antennas, the one or more receive antennas, the radar receiver, and the one or more processing devices comprise a synthetic aperture radar (SAR).