Since the 1970s, the remote sensing synthetic aperture radar (SAR) systems have been developing increasingly. The SIR-C/X SAR mounted on the American Space Shuttle Endeavour firstly completed the high resolution 3D imaging all over the globe in Feb. 2000 . It employs three dual-polarized sub-arrays operating at L-, C- and X-bands, respectively, as shown in Fig.1. The multi-band dual-polarized array receives more target information than that of a single-band one-polarization counterpart, and thus enhances the capability of target detection and identification. The common bands of space-borne SARs are L (center at 1.275 GHz), S (3.0 GHz), C (5.3 GHz) and X (9.6 GHz) bands. To minimize the volume and weight of a SAR antenna, a common aperture configuration for dual or multiple bands are expected, which will share the sub-systems behind the array as well. As a result, a lot of shared-aperture dual-band dual-polarization (DBDP) integrated planar antennas have been proposed in the last two decades .
The typical configurations of the shared-aperture DBDP planar arrays include the perforated structure, the interlaced layout and the overlapped layout. The perforated structure mainly includes perforated-patch/patch [3,4], ring/patch [5-7] and cross-patch/patch. The interlaced layout includes interlace patch with dipole/slot [9-11] and interlaced slot with slot/dipole [12-13],
In this chapter a universal multi-band SAR array synthetic method is introduced to extend the shared-aperture DBDP array to a shared-aperture MBDP SAR array. As a validation, a shared-aperture L/S/X tri-band dual-polarized (TBDP) array prototype was fabricated and measured with the central frequencies of 1.25 GHz, 3.5 GHz, and 10 GHz for L-, S- and X- bands, respectively . The array design and measured results are presented and discussed.
2. Array configuration
2.1. TBDP array configuration
From the system view, a similar beamwidth in the transverse (elevation) direction for each band is desired, which means that the transverse dimension of aperture for each band should be in proportion to its wavelength . As an example, the transverse dimensions of aperture in L-, S- and X- bands for the SIR-C/X-SAR are 2.95 m, 0.75 m, and 0.4 m, respectively, which have the ratio of 7.375:1.875:1. Then for a TBDP array with three independent apertures of L/S/X bands, its corresponding transverse dimension of aperture may be 2.88 m for L-band, 1.08 m for S-band and 0.36 m for X-band with the ratio of 8:3:1, as seen in Fig.2a.
In this chapter a novel design method is proposed, as shown in Fig.2b, where L/S and L/X sub-arrays are set on the two sides, and L-band sub-array is located in the middle. Then the transverse dimensions of aperture is 1.08+1.44+0.36 =2.88m for L-band, 1.08 m for S-band and 0.36 m for X-band, with ratio of 8:3:1 that is close to the central wavelength ratio of L-, S- and X-bands. Thus S-band and X-band elements are physically separated by a large spacing to avoid mutual interference, and a 33% off in aperture size is achieved as compared to its counterpart with independent tri-band aperture of 1.08 + 2.88 + 0.36 = 4.32 m. Actually, this method divides a TBDP shared-aperture array into two DBDP shared-aperture sub-arrays and one single-band DP sub-array so that the foregone works of DBDP shared-aperture arrays can be consulted and eventually the development of a TBDP shared-aperture array will be simplified.
2.2. L/S, L/X & L sub-array configuration
The interlaced configuration of Fig.3 is adopted for the L/S and L/X DBDP shared-aperture sub-array designs, which is more flexible for the odd frequency ratio than the perforated one. Moreover, the “pair-wise anti-phase feeding” technique  is used in both S- and X- bands to improve the cross-polarization performance and the polarization isolation of whole array as well.
The S-band and X-band elements in L/S and L/X DBDP arrays are distributed in square lattices, with the element spacing of
Since this is a phased array with two-dimensional scanning capability, each element is terminated to a T/R module and a beam controller is employed to steer the beam rather than a fixed feed network. Thus final output of each element is connected to a SMA connector.
3. Element design
3.1. S- & X-band element
To meet the requirements of bandwidth and cross-polarization as well as to simplify the design and fabrication, stacked patches are adopted for the S-band (in L/S sub-array) and X-band (in L/X sub-array) elements. To realize the dual-polarized radiation, the element should be symmetric in two dimensions, so that circular and square (rectangular) shapes are the most common ones. In this design, the square patch is adopted for its better cross-polarization performance  and lower fabrication tolerance, as shown in Fig.4.
The feeding approach greatly refers to the isolation and the cross-polarization level of an element. [17, 21] have presented profound studies on the feeding technique of stacked patches. In , the excellent isolation of better than 40dB is measured by using the hybrid excitation and a balanced feed. As the cost, however, these feeding approaches are too complicated as they are applied to the array environment, because that the aperture coupling and the balanced feed are difficult for the vertical connection, while the balanced feed needs a 180 degree feed network. To ensure the fabrication reliability, the dual-probe feed is finally employed, although its isolation is not as good as the aforementioned methods.
To overcome this drawback, an asymmetric slot is etched on the driven patch to improve the port isolation . Fig.5 and Fig.6 are the calculated S parameters and radiation patterns simulated by HFSS 10.0, showing the change of isolation and cross-polarization with the slot loaded. Different slot shapes are used for the S- and X- band driven patches to improve the isolation. In addition, it is found that the “H”-shaped slot in Fig.4b has less impact to S11 than the “cross”-shaped slot in Fig.4a, thus one has to rematch the element of latter while the “H”-shaped needs not. As seen from Fig.7, the field disturbance introduced by the slot deteriorates the element cross-polarization level. Fortunately, this would not be a problem for an array as the “pair-wise anti-phase feed” technique is used in the array configuration. The detailed design of the S/X stacked patches has been introduced in .
3.2. L-band element
For L/S, L/X and L sub-arrays, the structure of L-band microstrip dipole is almost the same. The only difference is the length of L-band dipole in L/S sub-array, which is shortened by 20% than its counterparts in L/X and L sub-arrays. In order to avoid the conflicting in sub-array splitting caused by the odd frequency ratio of about 3:1 for L- to S-band, the capacities are intentionally introduced in its feed network to realize a minimized design (See Fig.4).
The structure of L-band dipole is detailed in Fig.6. The microstrip dipole is printed on the top layer with a height of
This element as well as its feed structure is very slim and thus can be flexibly interlaced in the gap of S- and X-band elements. Although the profile of L-band element is somewhat high, its intrinsic broadband and good cross-polarization performance make it suitable for the TBDP SAR antenna mission.
4. Measurement results
A shared-aperture L/S/X TBDP array prototype was fabricated and measured to validate the design. Figs.8 a and b show the back and the front side of the array, respectively, which is incorporated with the L/X, L and L/S sub-arrays (from up to bottom in Fig.8a). The S-parameters were measured using the Agilent 8722ES and the measured radiation patterns were obtained in an anechoic chamber, while all the losses such as the insert losses of coaxial line, connectors and power dividers were compensated in the process.
The measured L-band S11 and radiation patterns are shown in Fig.9 and Fig.10. It is seen that though the boundary conditions of L-band elements in L, L/S and L/X sub-arrays are quite different, similar return loss performance is measured except for LS4, which perhaps is caused by the solder false. The measured VSWR≤2 bandwidth is 167MHz (1.163-1.330GHz, 13.4%), and the array isolation is better than 37dB over the whole bandwidth (See Fig.9b). The array isolation hereinafter is defined as -10log10|S21|, which is a positive dB value.
Although the whole L-band array is compounded by the elements in three separate sub-arrays, the measured radiation patterns keep in accordance with the theoretical ones, which means that the compound method raises little impact on the resulting radiation patterns. The cross-polarization level remains lower than -30 dB within the main lobe. The measured gain is 13.2dB (H-port) and 14.6dB (V-port), corresponding to the antenna efficiency of 62% and 61.2%, respectively. These lower figures may be related to its long electric-length of the microstrip and two-wire line feeder.
The narrowest VSWR≤2 bandwidth among all elements is defined as the array VSWR≤2 bandwidth. In S-band, an array VSWR≤2 bandwidth of 14.8% (3.25-3.768GHz) is measured according to Fig.11a, while the measured array isolation remains better than 45dB over the bandwidth (Fig.11b).
The measured radiation patterns agree well with the theoretical ones and the cross-polarization level keeps lower than -30dB in the main lobe (See Fig.12). Its measured front-back ratio is 37.8dB. The measured gain of 18.6dB is achieved for the 8×2 elements aperture, which means an antenna efficiency of 92.4%. Moreover, instead of the phase shifter, the coaxial line with custom-made length is applied to realize the scan experiment, as shown in Fig.13. From Fig.14, a scan capability of ±27°is observed.
As shown in Fig.15, an array VSWR≤2 bandwidth of 16.8% (9.098- 10.781GHz) is measured in X-band with the array isolation better than -43dB over the whole bandwidth. From Fig.16, the measured radiation patterns fit well with the theoretical ones with the cross-polarization level better than -35dB in the main lobe. The measured front-back ratio is 42.3dB at the broadside direction, which is more than 4dB better than that of S band. It benefits from the larger electric size of the ground. Measured gains of 22.19dB (16×2 elements, V port) and 21.79dB (16×2 elements, H port) are realized, which mean the efficiencies of 92.7% and 84.5% at V- and H- ports, respectively. The scan radiation patterns of 30° in both ports and both planes were also measured. As shown in Fig. 17a-c,, the grate lobe of a little higher than -10dB appears around 80°-90°.
The method of assembling two DBDP shared-aperture sub-arrays and one single-band DP sub-array to form a MBDP shared-aperture array has been introduced. An array prototype has been fabricated and measured to validate the method. According to the measured results, the array prototype achieves satisfactory results: similar bandwidth of 13.4%~16.7% in three bands, the array isolation of better than 37dB for all bands and the cross-polarization level of lower than -30dB within the main lobe region and the scanning capacity of ±27 degree at S- and X- bands. The array prototype has a 33% off in the aperture size as compared with tri-band independent aperture antenna and exhibits robust characteristics throughout the bandwidths. This array design method can be extended to the shared-aperture arrays with more than three bands.