In-Swath and Out-of-Swath Radio Frequency Interference Mitigation for Elevation Multichannel SAR Data

In-Swath and Out-of-Swath Radio Frequency Interference Mitigation for Elevation Multichannel SAR Data

Z. Lv, Z. Zhang, H. Fan, Z. Chen, J. Bi and W. Wang, "In-Swath and Out-of-Swath Radio Frequency Interference Mitigation for Elevation Multichannel SAR Data," in IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1-19, 2024, Art no. 5217119, doi: 10.1109/TGRS.2024.3428418.

Abstract: The electromagnetic environment is becoming complex as the usable spectrum will be allocated for more services. As a result of this situation synthetic aperture radar (SAR) missions are frequently perturbed by radio frequency interference (RFI) that jeopardizes their scientific observations all over the world. 

The state-of-the-art multichannel SAR has anti-RFI capability since it’s capable of digitally modulating the antenna pattern (AP) in postprocessing, thereby steering the null toward the angle of arrival (AOA) of the RFI in the spatial domain. However, the AOA of RFI is space-variant, meaning that the mitigation performance of beamformers sensitive to AOA will greatly deteriorate. In addition, the AOA of in-swath RFI and the target echo arrive simultaneously, thus the traditional beamformer will generate a distorted AP, deteriorating the SAR imagery. 

In light of these considerations, this article studies the in-swath and out-of-swath RFIs in elevation multichannel SAR and develops their countermeasures. Thereinto, a least  $\ell _{1}$ -norm model is developed to estimate the AOA of the RFI, followed by two schemes developed to separate the RFI. The former develops a beamformer that joint sidelobe control and null expanding to mitigate the space-variant out-of-swaths RFI, whereas the latter develops a blind source separation (BSS)-based technology to mitigate in-swath RFI, avoiding AP distortion and restoring SAR imagery. The effectiveness of the proposed approaches is supported by experiments based on the measured X-band airborne DBF-SAR data as well as the simulated SAR data.

keywords: {Prevention and mitigation;Synthetic aperture radar;Interference;Signal to noise ratio;Remote sensing;Radar antennas;Azimuth;Digital beamforming (DBF);multichannel synthetic aperture radar (SAR);radio frequency interference (RFI)},
URL: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10598185&isnumber=10354519

Summary

This article discusses radio frequency interference (RFI) mitigation techniques for multichannel synthetic aperture radar (SAR) systems. The key points are:

1. RFI is a growing problem for SAR systems due to increasing use of the radio spectrum. It can severely degrade SAR image quality.

2. Multichannel SAR systems with digital beamforming (DBF) capabilities provide new opportunities for RFI mitigation compared to single-channel systems.

3. The paper distinguishes between in-swath RFI (coming from within the imaging swath) and out-of-swath RFI (from outside the swath). These require different mitigation approaches.

4. For out-of-swath RFI, the paper proposes an optimized beamformer that can steer nulls toward the RFI direction while maintaining desired antenna pattern characteristics.

5. For in-swath RFI, a blind source separation (BSS) technique is proposed to separate the RFI from the desired signal.

6. The paper develops a complete processing flow including RFI angle-of-arrival estimation and the mitigation techniques.

7. Experiments are conducted using both simulated data and real airborne X-band SAR data to demonstrate the effectiveness of the proposed methods.

8. Results show the proposed techniques outperform traditional methods like notch filtering and conventional digital beamforming in terms of RFI suppression and preservation of SAR image quality.

9. The methods can handle space-variant RFI and are not limited by signal-to-interference ratio, which are limitations of some existing techniques.

In summary, the paper presents novel RFI mitigation approaches tailored for multichannel SAR systems that show promising results for both in-swath and out-of-swath interference scenarios. 

Authors, Associations, and Prior Work

Based on the information provided in the paper, here are details about the authors, their associated institutions, and some context about related work:

Authors and Institutions:

1. Zongsen Lv: Graduate Student Member of IEEE, affiliated with:
   - Department of Space Microwave Remote Sensing System, Aerospace Information Research Institute, Chinese Academy of Sciences (CAS), Beijing, China
   - University of CAS, Beijing

2. Zhimin Zhang: Affiliated with Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing

3. Huaitao Fan: Member of IEEE, affiliated with Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing

4. Zhen Chen: Member of IEEE, affiliated with Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing

5. Jianzhong Bi: Affiliated with Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing

6. Wei Wang: Member of IEEE, affiliated with Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing

All authors are associated with the Chinese Academy of Sciences, specifically the Aerospace Information Research Institute in Beijing. This suggests a strong focus on aerospace and remote sensing research at this institution.

Prior Related Work:

The paper cites several prior works related to RFI mitigation in SAR systems. Some key related works mentioned include:

1. Early work on notch filtering for RFI suppression in E-SAR and ALOS PALSAR systems (Buckreuss and Horn, 1998; Rosen et al., 2008).

2. Time-domain and time-frequency domain notch filtering improvements (Li et al., 2022; Han and Zhou, 2022).

3. Eigensubspace-based filtering for narrow-band interference suppression (Zhou et al., 2007).

4. Independent component analysis for RFI suppression (Zhou et al., 2013).

5. Subband spectral cancellation for narrow-band interference suppression (Feng et al., 2012).

6. Sparse recovery methods for RFI extraction (Nguyen et al., 2014, 2016).

7. Digital beamforming-based RFI mitigation (Bollian et al., 2018, 2022).

8. Work on multichannel cancellation for barrage jamming suppression (Cheng et al., 2022, 2023).

The authors' own prior work is not explicitly highlighted in the paper, but given their affiliations, it's likely they have been involved in SAR and RFI mitigation research at the Chinese Academy of Sciences.

The paper builds upon this prior work by addressing limitations in existing methods, particularly for multichannel SAR systems, and proposing novel approaches for both in-swath and out-of-swath RFI mitigation. This processing may be of interest to anyone intending to jam or interfere with such Chinese Satellite SAR.

Artifacts and Data

Based on the information provided in the paper, the following databases and artifacts were used for experimental verification:

1. X-band DBF-SAR Data:
   - Acquired by an X-band DBF-SAR system developed by the Institute of Electronics, Chinese Academy of Sciences (IECAS)
   - Collected during an outfield experiment in Guangdong Province, China, in November 2019
   - The system has 16 receiving channels in elevation
   - This data was used to verify in-swath RFI mitigation performance

2. Sentinel-1 SAR Data:
   - Acquired from the European Space Agency (ESA) Sentinel-1A satellite
   - Operating in strip map mode
   - Data acquisition area: Utrecht, The Netherlands, September 2016
   - This data was used to simulate 16-channel elevation multichannel SAR data for verifying out-of-swath RFI mitigation performance

3. Simulated Data:
   - Point target simulations and distributed target simulations were conducted
   - Parameters for these simulations are provided in Tables II and III of the paper

Regarding code artifacts or radar design data for independent validation:

The paper does not explicitly mention any publicly available code repositories or detailed radar design data. The authors do not state that they have made their implementation code or detailed system specifications publicly accessible.

However, the paper provides:

1. Detailed descriptions of the algorithms and processing steps
2. System parameters for the X-band DBF-SAR system (Table I)
3. Simulation parameters (Tables II and III)
4. Flowcharts and block diagrams of the proposed methods

While these details allow for a general understanding and potential reimplementation of the concepts, they may not be sufficient for exact reproduction of the results without additional information.

For independent validation, researchers would likely need to:

1. Implement the described algorithms based on the provided information
2. Use their own SAR systems or simulations with similar parameters
3. Apply the methods to publicly available SAR datasets (e.g., Sentinel-1 data)

It's worth noting that the lack of publicly available code or exact system specifications is not uncommon in radar research due to potential proprietary or sensitive nature of some technologies. Researchers interested in validating or building upon this work might need to contact the authors directly for more detailed information or collaborate with institutions having similar multichannel SAR capabilities.

Figures and Tables

Here's a list of the figures and tables in the article, along with explanations of what each is intended to show:

Figures:

1. Figure 1: Schematic of the working principle and geometric relationship for an elevation multichannel SAR system and two interferers.
   - Shows the basic geometry of the SAR system, swath, and interferers.

2. Figure 2: Schematic of the receive chain for elevation multichannel SAR in SCORE mode.
   - Illustrates how signals are processed in the SAR system with Scan-On-Receive (SCORE) mode.

3. Figure 3: RFI mitigation performance for elevation multichannel SAR in SCORE mode.
   - Demonstrates how SCORE mode performance degrades with increasing signal-to-interference-plus-noise ratio (SINR).

4. Figure 4: Schematic for the anti-RFI capability of SCORE mode.
   - Shows how SCORE beam scans the swath and amplifies both SAR signal and RFI.

5. Figure 5: Space-variation of AOA of RFI.
   - Illustrates how the angle of arrival (AOA) of RFI changes over time due to platform motion.

6. Figure 6: Beam patterns for spatial filtering RFI.
   - Compares beam patterns of SCORE and Minimum Variance Distortionless Response (MVDR) beamformers.

7. Figure 7: Flowchart of the RFI mitigation scheme for elevation multichannel SAR data -illustrates the flowchart of the proposed RFI mitigation scheme for elevation multichannel SAR data. Let me break down the processing steps in detail:

1. Input: The process starts with raw multichannel SAR data contaminated by RFI.

2. RFI AOA Estimation:
   - A least ℓ1-norm model is used to estimate the angle of arrival (AOA) of the RFI.
   - This step helps determine whether the RFI is in-swath or out-of-swath.

3. Decision: Is the RFI in-swath or out-of-swath?
   - Based on the estimated AOA, the process branches into two paths.

4. Out-of-Swath RFI Mitigation:
   - If the RFI is out-of-swath, the following steps are taken:
     a. Range Pulse Compression: This step compresses the signal in the range direction to reduce pulse extension loss.
     b. Optimized Beamformer: A beamformer is designed that jointly controls sidelobes and expands nulls.
     c. Spatial Filtering: The optimized beamformer is applied to filter out the RFI spatially.

5. In-Swath RFI Mitigation:
   - If the RFI is in-swath, a Blind Source Separation (BSS) approach is used:
     a. Data Preprocessing: This includes data rearrangement, centralization, and whitening.
     b. BSS Processing: The JADE (Joint Approximation Diagonalization of Eigen-matrices) algorithm is applied to separate the RFI from the SAR signal.
     c. Data Postprocessing: This step includes signal screening, data rearrangement, and amplitude compensation to address ambiguities in the BSS results.

6. Output: The final output is the RFI-mitigated SAR data.

How the processing works:

1. The AOA estimation helps identify whether the RFI is coming from within the imaging swath or outside it. This is crucial because different mitigation strategies are needed for each case.

2. For out-of-swath RFI:
   - Range pulse compression is performed first to mitigate pulse extension loss.
   - The optimized beamformer is designed to create deep, wide nulls in the direction of the RFI while maintaining desired characteristics of the antenna pattern.
   - Spatial filtering then effectively suppresses the out-of-swath RFI.

3. For in-swath RFI:
   - BSS is used because spatial filtering alone would distort the desired signal (as the RFI comes from the same direction as the SAR echo).
   - The preprocessing steps prepare the data for BSS by removing correlations and centering the data.
   - The JADE algorithm separates the mixed signals into independent components.
   - Postprocessing then identifies which component is the desired SAR signal and performs necessary corrections.

This approach allows for effective mitigation of both in-swath and out-of-swath RFI, addressing the limitations of traditional methods and taking advantage of the multichannel SAR system's capabilities.

8. Figure 8: Beam pattern of SCORE, LCMV with null expanding, and proposed beamformer.
   - Compares the beam patterns of different beamforming methods.

9. Figure 9: Statistical histograms of the I channel, Q channel, amplitude, and phase of SAR and RFI signals.
   - Shows the different statistical characteristics of SAR and RFI signals.
Figures 10 through 17 demonstrate the results and performance analysis of the proposed RFI mitigation techniques compared to existing methods. I'll describe the key differences shown in these figures:

Figure 10:
- Shows experimental results for X-band DBF-SAR data.
- Compares single channel imaging results (b) with mitigation results from FNF method (c), SCORE beam (d), and the proposed approach (e).
- The proposed approach (e) shows the cleanest image with the least visible RFI artifacts.

Figure 11:
- Provides enlarged areas from Figure 10 for detailed comparison.
- The proposed approach shows better preservation of strong scatterers and better suppression of RFI in water areas compared to FNF and SCORE methods.

Figure 12:
- Analyzes signal power after RFI mitigation in nearshore waters.
- The proposed method shows the lowest overall signal power, indicating better RFI suppression.

Figure 13:
- Illustrates the geometric relationship and target distribution in the simulation setup for point target analysis.

Figure 14:
- Compares mitigation results for simulated point targets.
- The proposed approach shows the highest output SINR and best preservation of target impulse response function (IRF).

Figure 15:
- Shows mitigation results for distributed target simulation using Sentinel-1 data.
- The proposed approach (d) demonstrates the best RFI suppression while preserving image details, compared to FNF (a), SCORE (b), and LCMV (c) methods.

Figure 16:
- Presents the zebra map and RFI angle variation in the simulation.
- Illustrates how the RFI angle changes over the scene, which impacts mitigation performance.

Figure 17:
- Displays coherence coefficients between RFI mitigation results and ground truth.
- The proposed approach (d) shows the highest overall coherence, indicating best preservation of the original signal while suppressing RFI.

Key differences:
1. RFI Suppression: The proposed method consistently shows better RFI suppression across different scenarios (airborne, simulated point targets, and distributed targets).

2. Image Quality: The proposed approach better preserves image details and target responses compared to other methods.

3. Coherence: Higher coherence with ground truth for the proposed method indicates better overall performance in maintaining signal integrity while removing RFI.

4. Robustness: The proposed method handles both in-swath and out-of-swath RFI effectively, showing consistent performance across different RFI scenarios.

5. Spatial Variability: The proposed approach demonstrates better handling of spatially variant RFI, as shown in the distributed target simulations.

Overall, these figures collectively demonstrate the superior performance of the proposed RFI mitigation technique compared to traditional methods across various performance metrics and scenarios.

Tables:

1. Table I: Main parameters of X-band DBF-SAR system
   - Lists the key specifications of the airborne SAR system used in experiments.

2. Table II: Main parameters of point target simulation
   - Provides details of the simulation setup for point target analysis.

3. Table III: Main parameters of distributed target simulation
   - Gives parameters for the distributed target simulation using Sentinel-1 data.

These figures and tables collectively illustrate the problem of RFI in SAR systems, explain the proposed mitigation techniques, and demonstrate their effectiveness through various experiments and simulations.