Abstract
The wideband linear frequency modulation (LFM) signals are widely used in information systems. The conventional direction-of arrival (DOA) estimation algorithms of LFM signals rely on a large number of snapshots, some of which are not reliable in numerous practical applications such as underwater array processing. To solve the above problem, we present a modified sparse iterative covariance (MSPICE)-based estimation method in fractional Fourier transform (FrFT) domain to estimate the DOA of wideband LFM signals. First, we extend the original SPICE algorithm in FrFT domain with a specific transform order for wideband LFM signals. Then, we utilize the energy centrobaric modification method to make the original SPICE more accurate without adding more computational complexity. The simulation results demonstrate the effectiveness of the proposed method.
Introduction
In sonar and radar systems, determining the Direction of Arrival (DOA) of incoming signals is a fundamental task. However, wideband Linear Frequency Modulation (LFM) signals, or "chirps," present a significant challenge because their frequency changes over time. This time-variance makes the steering vector—the mathematical model of how signals interact with a sensor array—difficult to process using standard narrowband techniques. Furthermore, high-resolution algorithms like MUSIC typically require a large number of data snapshots. In dynamic environments, such as underwater acoustics, the available window for reliable data collection is often too short to meet these requirements.
This paper introduces FrFT-MSPICE, a novel framework designed to achieve high-resolution DOA estimation for wideband LFM signals using as few as two snapshots. By combining the Fractional Fourier Transform (FrFT) to "dechirp" signals with a modified sparsity-based iterative algorithm, the method achieves high-resolution tracking in scenarios where traditional methods fail.
Methodology
The proposed approach leverages the Fractional Fourier Transform (FrFT) to simplify the signal processing pipeline. The FrFT can be interpreted as a rotation in the time-frequency plane. By selecting an optimal rotation angle related to the chirp rate of the LFM signal, the wideband signal is "dechirped" into a single-frequency sine wave in the FrFT domain.
Once the signal is transformed into a narrowband-like representation, the steering vector becomes time-invariant. The authors then apply a modified version of the Sparse Iterative Covariance (SPICE) algorithm. SPICE is a parameter-free, sparsity-based method that is inherently robust in low-snapshot scenarios. To overcome the "grid effect"—where accuracy is limited by the resolution of the search grid—an energy centrobaric modification is employed. This technique interpolates between grid points based on spectral energy leakage, providing sub-grid precision without increasing the computational burden of a denser search grid.
Results and Discussion
The effectiveness of FrFT-MSPICE was evaluated using simulated datasets featuring a 32-sensor linear array and LFM signals with a 12 MHz center frequency. The simulation tested a challenging scenario with three targets (at 70°, 72°, and 110°) and only two snapshots at a 20 dB Signal-to-Noise Ratio (SNR).
As shown in the angular spectrum comparison, the conventional FrFT-MUSIC method fails to resolve the closely spaced targets at 70° and 72°, merging them into a single peak. While the Spatial Resampling with Iterative Adaptive Approach (SR-IAA) separates the targets, it introduces significant sidelobes and "ghost" signals. In contrast, the proposed FrFT-MSPICE (labeled as FrFT-CSPICE in simulation results) clearly resolves all three targets with high precision and minimal sidelobe interference. Quantitative analysis further confirmed that the modified SPICE approach significantly reduces Root Mean Squared Error (RMSE) compared to the standard SPICE algorithm, particularly at higher SNR levels.
Conclusion
FrFT-MSPICE provides a robust solution for DOA estimation of wideband LFM signals in data-starved environments. By combining the dechirping capabilities of the FrFT with the efficiency of sparsity-based recovery and centrobaric interpolation, the method achieves superior resolution over traditional benchmarks. Future work will focus on reducing the iterative computational complexity for real-time applications and validating the model against real-world underwater acoustic data where non-Gaussian noise and multipath effects are prevalent.
Summary updated