[Paper Review] Airborne Radar STAP using Sparse Recovery of Clutter Spectrum

Ke Sun, Hao Zhang+3 more|arXiv (Cornell University)|Aug 25, 2010

Radar Systems and Signal Processing25 references18 citations

TL;DR

This paper proposes SR-STAP, a novel airborne radar STAP algorithm that leverages sparse recovery to estimate the clutter space-time spectrum in angle-Doppler domain using limited training data. By exploiting the intrinsic sparsity of clutter and employing joint sparse recovery across multiple training samples, SR-STAP achieves fast convergence and reduced dependence on prior knowledge, outperforming traditional knowledge-based STAP methods in non-stationary clutter environments with improved robustness to prior mismatch.

ABSTRACT

Space-time adaptive processing (STAP) is an effective tool for detecting a moving target in spaceborne or airborne radar systems. Statistical-based STAP methods generally need sufficient statistically independent and identically distributed (IID) training data to estimate the clutter characteristics. However, most actual clutter scenarios appear only locally stationary and lack sufficient IID training data. In this paper, by exploiting the intrinsic sparsity of the clutter distribution in the angle-Doppler domain, a new STAP algorithm called SR-STAP is proposed, which uses the technique of sparse recovery to estimate the clutter space-time spectrum. Joint sparse recovery with several training samples is also used to improve the estimation performance. Finally, an effective clutter covariance matrix (CCM) estimate and the corresponding STAP filter are designed based on the estimated clutter spectrum. Both the Mountaintop data and simulated experiments have illustrated the fast convergence rate of this approach. Moreover, SR-STAP is less dependent on prior knowledge, so it is more robust to the mismatch in the prior knowledge than knowledge-based STAP methods. Due to these advantages, SR-STAP has great potential for application in actual clutter scenarios.

Motivation & Objective

• To address the limitation of conventional STAP methods that require large amounts of statistically independent and identically distributed (IID) training data for clutter covariance matrix estimation.
• To overcome the poor performance of existing STAP techniques in real-world clutter environments, which are often locally stationary and lack sufficient IID training samples.
• To develop a robust STAP algorithm that minimizes reliance on prior knowledge while maintaining high detection performance in practical airborne radar scenarios.
• To exploit the inherent sparsity of clutter distribution in the angle-Doppler domain to enable accurate and efficient clutter spectrum estimation with limited training data.

Proposed method

• The method models the clutter space-time spectrum as a sparse signal in the angle-Doppler domain, leveraging the fact that clutter energy is concentrated in a few dominant directions and Doppler shifts.
• It applies sparse recovery techniques—specifically compressive sensing principles—to reconstruct the clutter spectrum from a small number of training samples.
• Joint sparse recovery is employed across multiple training snapshots to enhance estimation accuracy and stability.
• A novel clutter covariance matrix (CCM) estimator is derived based on the recovered sparse spectrum, ensuring optimal STAP filter design.
• The STAP filter is then computed using the estimated CCM, enabling effective moving target detection in the presence of strong clutter.
• The approach is validated using both real Mountaintop radar data and simulated scenarios to demonstrate robustness and convergence speed.

Experimental results

Research questions

• RQ1 Can sparse recovery techniques effectively estimate the clutter spectrum in airborne radar STAP when training data is limited and non-IID?
• RQ2 How does joint sparse recovery across multiple training samples improve the accuracy and robustness of clutter spectrum estimation compared to single-sample recovery?
• RQ3 To what extent does SR-STAP reduce dependency on prior knowledge compared to conventional knowledge-based STAP methods?
• RQ4 What is the convergence performance of SR-STAP in real-world and simulated clutter environments with limited training data?
• RQ5 How does the proposed method maintain detection performance under prior mismatch conditions common in practical radar systems?

Key findings

• SR-STAP demonstrates a significantly faster convergence rate compared to traditional statistical STAP methods, especially in low-sample regimes.
• The algorithm achieves improved detection performance in both simulated and real Mountaintop radar data, even with limited training samples.
• Joint sparse recovery across multiple training samples enhances the stability and accuracy of clutter spectrum estimation.
• SR-STAP reduces sensitivity to prior knowledge mismatch, making it more robust than knowledge-based STAP approaches in practical, non-ideal clutter environments.
• The estimated clutter covariance matrix via sparse recovery enables effective STAP filtering, leading to high mainlobe suppression and improved target detection.