Nonparametric Detector Performance for Active Sonar in Ocean Reverberation Background with Multiple False Targets Jamming

Songwen Li, Yang SongAdvances 2015 · DOI

Abstract

Passive sonar, which used to be the most effective device for submarine detection, cannot satisfy the requirement of anti-submarine operation after the appearance of quiet submarine decades ago, and active sonar is playing a more important role. Active sonar often works in reverberation limited condition in shallow water area because of the strong backscattering of the sea bottom. Unlike the volume reverberation often encountered by sonar working in deep ocean, bottom reverberation is usually not Rayleigh distributed. Experiment results show that the sea bottom reverberation is more likely to be k-distributed, thus the nonparametric detector, such as generalize sign detector, is more suitable for active sonar working in shallow water area. Multiple false targets can decrease the detection possibility of parametric and nonparametric CFAR detector, and usually used as an important method to counter active sonar. From Monte Carlo simulation, it is shown that only several false targets are needed to make the active sonar detection performance decline severely when it is using nonparametric detector in ocean reverberation background.

Introduction

In modern anti-submarine warfare, the advent of "quiet" submarines has significantly reduced the efficacy of traditional passive sonar systems. Consequently, active sonar has become a critical tool. However, active sonar performance is often limited by reverberation, particularly in shallow water where signals bounce off the sea floor. This "bottom reverberation" typically follows a non-Rayleigh distribution, often modeled as a $K$-distribution.

To maintain a Constant False Alarm Rate (CFAR) in these complex, non-stationary environments, nonparametric detectors like the Generalized Sign (GS) test are preferred. This study investigates the vulnerability of these robust detectors to Multiple False Target Jamming (MFTJ), a common countermeasure used to mask real targets.

Methodology

The study utilizes Monte Carlo simulations to evaluate detector performance under various conditions:

  • Background Modeling: Sea bottom reverberation is modeled using a $K$-distribution, while volume reverberation is modeled as Rayleigh for baseline comparison.
  • Target Dynamics: Targets are simulated using both non-fluctuating and fluctuating (Swerling II) models.
  • Detection Algorithm: A Generalized Sign (GS) test detector is employed. It uses rank quantization, comparing a detection unit's amplitude against 16 reference units. Detection is declared based on hits across multiple sweeps ($M=8$).
  • Jamming Scenarios: The simulation introduces 2, 4, 6, and 8 false targets into the reference window to measure the resulting degradation in the Probability of Detection ($P_d$).

Results and Discussion

The simulations reveal that while nonparametric detectors are robust to the specific statistical distribution of the background noise, they are highly sensitive to the nature of jamming signals.

Impact of Target Fluctuation

A critical finding is the difference in impact between non-fluctuating and fluctuating false targets. In scenarios with non-fluctuating targets, increasing the Signal-to-Noise Ratio (SNR) can eventually overcome the jamming effect, allowing the probability of detection to reach 1.0.

Figure 3: Probability of detection ($P_d$) vs. SNR for non-fluctuating targets. High SNR eventually allows for complete detection despite jamming.

However, when false targets fluctuate (Swerling II model), the detector experiences a "saturation" effect. Because the detector relies on the rank of the signal relative to its neighbors, fluctuating false targets frequently exceed the rank threshold by chance. This "blinds" the detector, causing the $P_d$ to plateau well below 1.0, regardless of the real target's SNR.

Figure 1: Probability of detection ($P_d$) vs. SNR for fluctuating targets. Note the saturation effect where $P_d$ plateaus below 1.0 for various jamming intensities.

Jamming Severity

The severity of the performance drop is stark. With 8 non-fluctuating false targets, the SNR requirement for 50% detection increases by approximately 5dB. In the fluctuating case, if 4 or more false targets are present, a 50% detection probability may become mathematically unreachable as the detector's rank test fails to distinguish the real target from the fluctuating noise floor.

Conclusion

Nonparametric detectors offer valuable robustness against varying ocean reverberation backgrounds. However, this study demonstrates a significant vulnerability to multiple false target jamming, especially when those targets fluctuate. The rank-based quantization method is easily overwhelmed, leading to a hard ceiling on detection probability. Future research should focus on adaptive windowing or spatial processing techniques to mitigate these jamming effects in real-world sonar applications.

Summary updated

Abstract

Passive sonar, which used to be the most effective device for submarine detection, cannot satisfy the requirement of anti-submarine operation after the appearance of quiet submarine decades ago, and active sonar is playing a more important role. Active sonar often works in reverberation limited condition in shallow water area because of the strong backscattering of the sea bottom. Unlike the volume reverberation often encountered by sonar working in deep ocean, bottom reverberation is usually not Rayleigh distributed. Experiment results show that the sea bottom reverberation is more likely to be k-distributed, thus the nonparametric detector, such as generalize sign detector, is more suitable for active sonar working in shallow water area. Multiple false targets can decrease the detection possibility of parametric and nonparametric CF AR detector, and usually used as an important method to counter active sonar. F rom Monte Carlo simulation, it is shown that only several false targets are needed to make the active sonar detection performance decline severely when it is using nonparametric detector in ocean reverberation background.
BibTeX
@inproceedings{li2015nonparametric,
  author = {Songwen Li and Yang Song},
  title = {{Nonparametric Detector Performance for Active Sonar in Ocean Reverberation Background with Multiple False Targets Jamming}},
  booktitle = {Advances},
  year = {2015},
  doi = {10.2991/esac-15.2015.55},
}

Loading related content...

Figure