The multiple-input multiple-output (MIMO) technique can improve the high-resolution wide-swath imaging capacity of synthetic aperture radar (SAR) systems. Beamspace MIMO-SAR utilizes multiple subpulses transmitted with different time delays by different transmit beams to obtain more spatial diversities based on the relationship between the time delay and the elevation angle in the side-looking radar imaging geometry. This paper presents a beamspace MIMO-SAR imaging approach, which takes advantage of real time digital beamforming (DBF) with null steering in elevation and azimuth multichannel raw data reconstruction. Echoes corresponding to different subpulses in the same subswath are separated by DBF with null steering onboard, while echoes received and stored by different azimuth channels are reconstructed by multiple Doppler reconstruction filters on the ground. Afterwards, the resulting MIMO-SAR raw data could be equivalent to the raw data of the single-channel burst mode, and classical burst mode imaging algorithms could be adopted to obtain final focused SAR images. Simulation results validate the proposed imaging approach.
High-resolution wide-swath (HRWS) imaging capacity is one of the most important aims of future spaceborne microwave remote sensing [
According to different types of transmitted waveforms, MIMO-SAR could be characterized as orthogonal waveform MIMO-SAR, subband MIMO-SAR, and beamspace MIMO-SAR [
In this paper, a processing approach for beamspace MIMO-SAR data focusing is proposed. The proposed approach includes three important steps: DBF with null steering onboard, multichannel azimuth data reconstruction, and single-channel burst mode SAR imaging. The key point of the proposed approach is DBF onboard and azimuth multichannel reconstruction for multichannel raw data preprocessing. The DBF step with null steering operated onboard separates echoes corresponding to different subpulses to extract more spatial diversities. Compared with the conventional DBF on receive approach named as scan-on-receive (SCORE) [
This paper is organized as follows. Section
For future spaceborne microwave remote sensing missions, SAR sensors will require a complete and frequent coverage of the Earth with a reasonably high geometric resolution, for example, an imaged swath width of no less than 400 km with an azimuth resolution of below 5 m [
Beamspace MIMO-SAR imaging scheme. (a) Four subpulses are transmitted by two subapertures in a single PRI, which results in more spatial diversities (SDs) in azimuth. (b) Two subpulses transmitted by the same Tx are used to illuminate different subswaths, while echoes corresponding to subpulses with different time delays transmitted by different subapertures are separated by narrow beam scanning on receive with null steering.
In azimuth
In elevation
As shown in Figure
Echoes from different subswaths could be easily separated by DBF in elevation, since the difference between their arriving directions is much larger than the beamwidth of the narrow scanning receive beam. Unfortunately, echoes of different subpulses transmitted into the same subswath are difficult to be separated by the conventional DBF receive approach, since the small angular interval between their echoes arriving directions is caused by the short time delay between two transmitted subpulses such as point targets P and Q as shown in Figure
Two subpulses transmitted in a single PRI to illuminate the same swath and their corresponding echoes superimpose in the receiving window. (a) Imaging geometry of two subpulses in a single PRI to illuminate the same swath. (b) Echoes of different subpulses superimpose in the receiving window.
To implement the beamspace MIMO-SAR imaging scheme, two subpulses should be transmitted into different subswaths by the same Tx, and echoes of multiple subpulses are, respectively, received by different narrow scanning receive beams. Furthermore, echoes are individually received by multiple subapertures arranged in azimuth. Therefore, a large receive antenna, which is divided into multiple subapertures in both azimuth and elevation, is required in a beamspace MIMO-SAR system. Two subapertures are used to transmit radar pulses and receive backscatter echoes, while others are only used to receive radar echoes. Furthermore, each subaperture contains lots of element antennas as shown in Figure
A large antenna adopted in beamspace MIMO-SAR. (a) The large receive antenna is divided into multiple subapertures in both azimuth and elevation directions, and each subaperture contains lots of element antenna. (b) Intrapulse beamsteering to transmit multiple subpulses via the ABF net. (c) Sharp pencil beam to receive echoes by a real DBF processor.
In beamspace MIMO-SAR systems, subpulses with different time delays and different beam pointing directions are transmitted, and their corresponding echoes could be separated by DBF on receive in elevation. The conventional DBF on receive onboard technique is detailed and investigated in [
The block diagram of the real time conventional DBF processor in elevation onboard.
The above-mentioned beamspace MIMO-SAR system based on the burst mode imaging scheme utilizes multiple subpulses in a single PRI with the same carrier frequency and phase coding but with different time delays and transmitting beam pointing direction to obtain more spatial samples in azimuth. Therefore, the proposed imaging approach includes three important parts: DBF with null steering in elevation to separate echoes corresponding to different subpulses, azimuth multichannel reconstruction to resolve the azimuth nonuniform sampling, and a single-channel burst mode SAR imaging processor as shown in Figure
Block diagram of the proposed imaging approach.
According to above-mentioned analysis results, as shown in Figure
As multiple subpulses are transmitted in a single PRI, their corresponding echoes will overlap at each receiver. For example, there are two point targets in the designed scene, point target P with the far slant range and point target Q with the near slant range as shown in Figure
Echoes received by multiple subapertures and stored by individual channels in elevation could be expressed as a vector
The block diagram of the real time DBF processor with null steering in elevation onboard is shown in Figure
The block diagram of the real time DBF processor with null steering in elevation onboard.
After separating echoes corresponding to different subpulses in the same swath, echoes received by all azimuth subapertures should be combined together. Compared with the SIMO SAR system, more effective phase centers are obtained, since two transmitters are adopted as shown in Figure
As the mentioned beamspace MIMO-SAR is mainly based on the burst imaging mode. The resulting raw data after DBF in elevation and azimuth multichannel reconstruction could be handled by burst mode (ScanSAR or TOPS) SAR processors.
To validate the proposed imaging approach, experiments on simulated raw data are carried out. Simulation parameters are listed in Table
Simulation parameter.
Parameter | Value |
---|---|
Satellite height | 570 km |
Satellite velocity | 7585 m/s |
Subpulse duration | 35 |
Bandwidth | 150 MHz |
Sampling frequency | 180 MHz |
Time delay between two subpulses | 40 |
Operated PRF | 1180 Hz |
Number of azimuth subapertures | 4 |
Number of elevation subapertures | 10 |
Subaperture height | 0.2 m |
Subaperture length | 2.4 m |
Look angle | 28.6° |
Slant range of the scene center | 634.68 km |
Burst duration | 0.32 s |
Figure
Echoes separation experiment. (a) Echo received by one of elevation receive channel (top) and range compressing results. (b) Echo separation results handled by the convectional DBF processor. (c) Echo separation results handled by the proposed two-step DBF processor.
To validate the whole processor imaging processor, an imaging scene consisting of six point targets is designed as shown in Figure
Imaging parameters of simulated point targets.
Targets | Azimuth | Range | ||||
---|---|---|---|---|---|---|
Resolution (m) | PSLR (dB) | ISLR (dB) | Resolution (m) | PSLR (dB) | ISLR (dB) | |
P1 | 3.26 | −13.26 | −10.08 | 0.89 | −13.26 | −9.92 |
P2 | 3.27 | −13.26 | −10.02 | 0.89 | −13.26 | −9.98 |
P3 | 3.27 | −13.24 | −9.96 | 0.89 | −13.26 | −9.93 |
The designed imaged scene in the slant range plane with six point targets.
Simulation results of the designed scene. (a) Echoes received by one of the subapertures. (b) Echoes of the imaged scene corresponding to the first subpulse after DBF with null steering. (c) Echoes of the imaged scene corresponding to the second subpulse after DBF with null steering. (d) Two-dimension spectrum of the imaged scene.
Contour plots of three point targets. (a) P1. (b) P2. (c) P3.
Simulation results of the designed distributed scene. (a) The designed extended scene. (b) Echoes of the imaged scene corresponding to the first subpulse after DBF with null steering. (c) Echoes of the imaged scene corresponding to the second subpulse after DBF with null steering. (d) Two-dimension spectrum of the imaged scene.
In this paper, an imaging processor to handle the beamspace MIMO-SAR raw data is proposed. The key point of the proposed imaging processor is its echoes separation corresponding to different subpulses and azimuth multichannel raw data construction. Based on the relationship between the transmitted time delay and the SAR side-looking imaging geometry, echoes corresponding to different subpulses are separated by a DBF processor with null steering onboard. In the proposed DBF processor, echoes corresponding to different subpulses are received by different sharp scanning pencil beams, which are implemented by the steering matrix and a series of time delayers. Afterwards, the interference signal corresponding to the undesired subpulse received by the sidelobe of the sharp scanning pencil beam is further suppressed via the matrix of the second DBF step for null steering. Then, the azimuth multichannel raw data is reconstructed by a modified multichannel reconstruction filter according to the subaperture arrangement in azimuth of beamspace MIMO-SAR. Imaging results on simulated raw data validate the proposed imaging approach.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This work is supported by NSF of China (no. 61271177).