Separation and localization of multiple distributed wideband chirps using the fractional Fourier transform

Jiexiao Yu, Liang Zhang, Kaihua Liu, Deliang Liu

Abstract

In complex environments like shallow water sonar or radar in rugged terrain, signals often arrive as distributed sources with significant angular spread rather than single point sources. For wideband chirps, this problem is compounded by time-varying spatial signatures that render traditional localization algorithms ineffective. This paper presents a novel localization framework based on the Fractional Fourier Transform (FrFT). By rotating signals into a specialized "Dechirping domain," the time-varying nature of wideband chirps is effectively neutralized. We derive a Generalized Array Manifold (GAM) using a first-order Taylor series expansion to model the distributed nature of the sources. Our method allows for the separation of overlapping sources in the FrFT domain, enabling the localization of more sources than physical sensors. Numerical simulations confirm that the proposed approach offers superior accuracy and robustness to noise and angular spread compared to existing benchmarks.

Introduction

Traditional array signal processing typically relies on the "point source" model, assuming signals arrive from precise, singular angles. However, real-world multipath propagation and scattering often result in "distributed sources," where signals appear as clusters with an associated angular spread. While narrowband distributed sources have been extensively studied, wideband chirps—signals with time-varying frequencies—pose a unique challenge. Their spatial signatures change over time, causing standard subspace-based estimation algorithms like MUSIC (Multiple Signal Classification) to fail. Existing wideband solutions often suffer from high computational complexity or an inability to handle signals that overlap in both time and frequency.

Methodology

The proposed framework leverages the Fractional Fourier Transform (FrFT), a mathematical tool that generalizes the standard Fourier Transform by rotating the signal in the time-frequency plane.

Domain Transformation and Source Separation

The authors identify two critical FrFT domains:

  1. Energy-Concentrated Domain: In this domain, the chirp's energy is compressed into a sharp peak, allowing for the isolation of individual sources. By applying a band-pass filter here, the algorithm can separate overlapping chirps before localization.
  2. Dechirping Domain: This domain is specifically chosen to "freeze" the time-varying nature of the wideband signal. Here, the signal's spatial location vector becomes related primarily to its central angle.
The Fractional Fourier Transform (FrFT) can be visualized as a rotation of the signal in the time-frequency plane, facilitating signal analysis in specialized domains.

Mathematical Modeling

To handle the distributed nature of the sources, the authors employ a first-order Taylor series expansion to derive a Generalized Array Manifold (GAM). This approximation simplifies the complex spatial characteristics of distributed sources into a model compatible with subspace algorithms. Finally, an adapted version of the MUSIC algorithm is applied within the Dechirping domain using the calibrated location vectors from the GAM.

Experimental Results

The algorithm was validated through extensive numerical simulations using a Uniform Linear Array (ULA) with 8 sensors and two uncorrelated wideband chirps with distinct initial frequencies and angular distributions (uniform and Gaussian).

Accuracy and Robustness

The proposed method consistently demonstrated a lower Root Mean Square Error (RMSE) for Direction of Arrival (DOA) estimation compared to traditional FrFT-MUSIC and other recent benchmarks. Notably, the method remained effective at signal-to-noise ratios (SNR) as low as -5 dB.

Handling Angular Spread and Source Density

A key advantage of this approach is its tolerance to angular spread. Performance gains over traditional methods became more pronounced as the spread increased (tested up to 12 degrees). Furthermore, because the algorithm separates sources in the FrFT domain prior to localization, it can successfully resolve more signal sources ($Q$) than there are physical sensors ($P$) in the array ($Q > P$), bypassing a fundamental limitation of standard DOA estimation.

Conclusion

This paper introduces a robust method for the localization of distributed wideband chirp sources by utilizing the FrFT to create a static representation in the Dechirping domain. The integration of a Taylor-series-based GAM and pre-localization source separation allows for high precision and the ability to handle dense signal environments. Future research could look into automating the chirp rate estimation to speed up the domain transformation or adapting the model to handle coherent (correlated) sources.

Summary updated

Abstract

In this paper, we consider the problem of localizing the multiple distributed wideband chirp sources using the fractional Fourier transform. The model in the time domain and that in the fractional Fourier domain derived by the Taylor series expansion are presented respectively. The representation of location vector in the Dechirping domain is illustrated which is only related to the central angle. A novel direction of arrival (DOA) estimation algorithm in the Dechirping domain is proposed, which is extended from the conventional multiple signal classification (MUSIC) algorithm in the time domain. Using this algorithm, except for estimating the DOA, the incidence source number can be determined as well which is allowed to exceed the sensor number in the array. To demonstrate the performance of proposed algorithm, numerical results are conducted. Compared to the previous FrFT-MUSIC algorithm based on the assumption of point source model, the proposed algorithm performs a better estimation performance, especially for large angular spread and low signal-to-noise ratio.
BibTeX
@article{yu2015separation,
  author = {Jiexiao Yu and Liang Zhang and Kaihua Liu and Deliang Liu},
  title = {{Separation and localization of multiple distributed wideband chirps using the fractional Fourier transform}},
  journal = {EURASIP J. Wirel. Commun. Netw.},
  year = {2015},
  doi = {10.1186/s13638-015-0497-9},
}
Figure