Noncircularity exploitation in signal processing overview and application to radar

Frédéric Barbaresco, Pascal Chevalier2008 · DOI

TL;DR

This paper explores the concept of "noncircularity" in signal processing, a property of signals that has been largely ignored in traditional radar systems. By exploiting this noncircularity, particularly in phase-coded radar waveforms, significant performance gains can be achieved in areas like detection, anti-jamming, and direction finding, opening new avenues for advanced radar capabilities.

What Problem Does This Paper Address?

Traditionally, signal processing in radar systems has assumed signals to be "second-order circular" (SO circular). This means that the second-order correlation function of the signal is zero, simplifying analysis and processing. However, many real-world signals, especially those from radio communications, are "SO noncircular" (or improper). These signals contain valuable information in their second-order correlation function that is lost when treated as SO circular. This paper addresses the gap by highlighting the potential of exploiting this noncircularity to improve radar performance, moving beyond the limitations of classical signal processing assumptions. The authors aim to provide an overview of this concept and its specific applications in radar, particularly for phase-coded waveforms.

Methods

The core of the paper revolves around the concept of "Widely Linear" (WL) filtering. Unlike standard linear filters, WL filters can process signals that are SO noncircular by incorporating the complex conjugate of the signal. This allows them to extract information from the non-zero pseudo-correlation function. The paper discusses how WL filters are optimal for processing nonstationary signals with noncircular complex envelopes. It then details various applications of WL filtering in signal processing, including:

  • Cochannel Interference Rejection: For cellular and radio communication systems, WL filters can effectively cancel noncircular interferences, leading to significant capacity gains.
  • CDMA Radiocommunications: WL filters are used in Multi-user Detection (MUD) schemes to mitigate Multiple Access Interference (MAI).
  • Detection Systems: New detectors based on WL filtering are proposed for improved signal detection in the presence of noncircular interferences.
  • MVDR Beamforming: WL MVDR beamformers are designed to enhance passive listening applications by handling SO noncircular interferences.
  • Equalization: WL equalizers are presented for various communication links (single carrier, DMT/OFDM, UWB) and MIMO systems to combat Inter-Symbol Interference (ISI) and interferences.
  • Parameter Estimation: The paper also touches upon exploiting noncircularity for direction finding and blind channel identification without necessarily using WL filters.

For radar applications, the paper proposes exploring phase-coded waveforms (like BPSK and PN sequences) and exploiting their noncircularity. This involves analyzing their ambiguity functions and considering new processing chains that utilize the complex symmetric unconjugated spatial covariance matrix (pseudocovariance matrix).

Datasets

Not discussed in the paper. The paper focuses on theoretical concepts and signal processing techniques rather than empirical evaluation on specific datasets.

Key Insights and Ideas

  • Noncircularity is a Rich Source of Information: Signals that are not SO circular contain information in their second-order correlation function that is missed by traditional processing. Exploiting this can lead to dramatic performance improvements.
  • Widely Linear Filtering is Key: WL filters are the natural extension of linear filters for noncircular signals, enabling the extraction of this hidden information.
  • Broad Applicability Beyond Communications: While noncircularity has been explored in radio communications, its potential in radar systems, particularly for phase-coded waveforms, is significant and largely untapped.
  • Radar Waveform Design Can Be More Holistic: The paper advocates for designing radar waveforms with multiple objectives beyond just the classical ambiguity function, including detection, anti-jamming, and communication capabilities.
  • Potential for Enhanced Radar Capabilities: Exploiting noncircularity can lead to improved radar detection, better anti-jamming performance (handling more interference sources), and more precise direction finding.

Key Results

The paper presents a comprehensive overview of the benefits of noncircularity exploitation across various signal processing domains. While specific quantitative results are detailed within the cited references, the paper highlights:

  • Significant performance gains in interference cancellation, equalization, and detection when using WL filters compared to conventional linear methods.
  • Doubled system capacity in some communication scenarios (e.g., Direct Sequence Spread Spectrum).
  • Improved detection probabilities in the presence of noncircular interferences, as illustrated by the example in Figure 2, where conventional receivers fail to detect a signal while optimal receivers succeed.
  • The potential to process more interferences than conventional methods, for example, up to 2(N-1) rectilinear interferences from an array of N sensors using WL MVDR beamformers.

Limitations and Future Work

The authors acknowledge that the concept of noncircularity and its exploitation, particularly in radar, is an evolving area.

  • Radar Waveform Exploitation is Nascent: The paper emphasizes that the noncircularity properties of phase-coded radar waveforms have not yet been exploited, suggesting this as a major area for future research.
  • Need for New Processing Chains: Developing new processing chains that leverage the pseudocovariance matrix is crucial for radar applications.
  • Multi-objective Waveform Design: Future work should focus on designing waveforms that serve multiple purposes, such as detection, anti-jamming, and communication, simultaneously.
  • Extension to Advanced Radar Systems: The authors suggest that the exploitation of noncircularity can be extended to multi-static, MIMO, and Ultra-Wideband radar systems.
  • The "What to Transmit" Question Remains: Echoing Woodward's sentiment, the paper concludes that the fundamental question of what to transmit for optimal radar performance, considering noncircularity, remains a significant challenge.

Summary updated

Abstract

With new generation of Active Digital Radar Antenna, there is a renewal of waveform generation and processing approaches, and new strategies can be explored to optimize waveform design and waveform analysis and to benefit of all potential waveform diversity. Among these strategies, building and exploitation of the Noncircularity of waveforms is a promising issue. Up to the middle of the nineties, most of the signals encountered in practice are assumed to be second order (SO) circular (or proper), with a zero second correlation function. However, in numerous operational contexts such as in radio communications, the observed signals are either SO noncircular (or improper) or jointly SO noncircular with a particular signal to estimate, to detect or to demodulate, with some information contained in the second correlation function of the signals. Exploitation of this information in the processing of SO noncircular signals may generate dramatic gain in performance with respect to conventional processing and opens new perspective in signal processing. The purpose of this paper is to present a short overview of the interest of taking into account the potential SO noncircularity of the signals in signal processing and to describe the potential interest of SO noncircular waveforms for radar applications. (50 pages)
BibTeX
@misc{barbaresco2008noncircularity,
  author = {Frédéric Barbaresco and Pascal Chevalier},
  title = {{Noncircularity exploitation in signal processing overview and application to radar}},
  year = {2008},
  doi = {10.1049/ic.2008.0743},
}
Figure