Multi-bit Distributed Detection of Sparse Stochastic Signals over Error-Prone Reporting Channels

Linlin Mao, Shefeng Yan, Zeping Sui, Hongbin Li

Abstract

We consider a distributed detection problem within a wireless sensor network (WSN), where a substantial number of sensors cooperate to detect the existence of sparse stochastic signals. To achieve a trade-off between detection performance and system constraints, multi-bit quantizers are employed at local sensors. Then, two quantization strategies, namely raw quantization (RQ) and likelihood ratio quantization (LQ), are examined. The multi-bit quantized signals undergo encoding into binary codewords and are subsequently transmitted to the fusion center via error-prone reporting channels. Upon exploiting the locally most powerful test (LMPT) strategy, we devise two multi-bit LMPT detectors in which quantized raw observations and local likelihood ratios are fused respectively. Moreover, the asymptotic detection performance of the proposed quantized detectors is analyzed, and closed-form expressions for the detection and false alarm probabilities are derived. Furthermore, the multi-bit quantizer design criterion, considering both RQ and LQ, is then proposed to achieve near-optimal asymptotic performance for our proposed detectors. The normalized Fisher information and asymptotic relative efficiency are derived, serving as tools to analyze and compensate for the loss of information introduced by the quantization. Simulation results validate the effectiveness of the proposed detectors, especially in scenarios with low signal-to-noise ratios and poor channel conditions.

Introduction

In Wireless Sensor Networks (WSNs), detecting "sparse" signals—events that are geographically or temporally rare, such as gas leaks or battlefield intrusions—is a critical task. However, these sensors are often constrained by limited battery life and narrow communication bandwidth. To manage these constraints, sensors must compress their high-resolution readings into a few bits (quantization) before transmission. This process introduces two primary challenges: significant information loss from quantization and potential bit-flip errors during transmission over noisy reporting channels.

While previous research has often focused on one-bit quantization or idealized perfect channels, this work addresses the realistic scenario of multi-bit quantization over error-prone reporting channels.

System Model and Quantization Strategies

The proposed framework utilizes a distributed architecture where local sensors process observations and transmit quantized data to a central Fusion Center (FC).

System model of the distributed detection network featuring local multi-bit quantization and error-prone reporting channels.

Two primary quantization strategies are examined:

  1. Raw Quantization (RQ): Sensors quantize the actual observed signal values directly.
  2. Likelihood Ratio Quantization (LQ): Sensors first calculate the likelihood ratio—a statistical measure of signal presence—and then quantize this value.

At the FC, a Locally Most Powerful Test (LMPT) is employed. This statistical framework is optimized for detecting weak signals by fusing the received multi-bit codewords. To maximize performance, the authors utilize Particle Swarm Optimization (PSOA) to determine optimal quantization thresholds that maximize the Fisher Information (FI), accounting for both quantization loss and channel noise.

Performance Analysis and Key Results

The study reveals that the choice of quantization strategy and bit depth significantly impacts detection reliability. A key finding is that LQ consistently outperforms RQ. This superiority stems from LQ's inherent ability to prioritize information most relevant to the detection task. Furthermore, LQ maps high-probability signal regions to binary codewords with large Hamming distances, providing natural resilience against bit-flip errors in the reporting channel.

Normalized Fisher Information across different bit depths (1b, 2b, 3b) and quantization strategies (RQ, LQ) as a function of channel error probability $P_e$.

As shown in the performance analysis, increasing the bit depth from 1-bit to 3-bit provides a substantial performance boost, nearing the theoretical limit of an uncompressed "Clairvoyant" detector. The Asymptotic Relative Efficiency (ARE) analysis indicates that RQ may require 2 to 4 times more sensors than LQ to achieve equivalent performance in noisy environments. Additionally, the optimal quantization thresholds are found to be adaptive; as channel noise increases, thresholds must decrease to ensure sufficient information reaches the FC.

Conclusion

This research establishes a robust framework for sparse signal detection in WSNs using multi-bit quantization. By optimizing quantization thresholds and leveraging Likelihood Ratio Quantization, the proposed system achieves near-optimal performance even under poor channel conditions and low signal-to-noise ratios. Future work may explore heterogeneous networks with varying sensor SNRs and real-time channel state estimation to further enhance system robustness.

Summary updated

Abstract

We consider a distributed detection problem within a wireless sensor network (WSN), where a substantial number of sensors cooperate to detect the existence of sparse stochastic signals. To achieve a trade-off between detection performance and system constraints, multi-bit quantizers are employed at local sensors. Then, two quantization strategies, namely raw quantization (RQ) and likelihood ratio quantization (LQ), are examined. The multi-bit quantized signals undergo encoding into binary codewords and are subsequently transmitted to the fusion center via error-prone reporting channels. Upon exploiting the locally most powerful test (LMPT) strategy, we devise two multi-bit LMPT detectors in which quantized raw observations and local likelihood ratios are fused respectively. Moreover, the asymptotic detection performance of the proposed quantized detectors is analyzed, and closed-form expressions for the detection and false alarm probabilities are derived. Furthermore, the multi-bit quantizer design criterion, considering both RQ and LQ, is then proposed to achieve near-optimal asymptotic performance for our proposed detectors. The normalized Fisher information and asymptotic relative efficiency are derived, serving as tools to analyze and compensate for the loss of information introduced by the quantization. Simulation results validate the effectiveness of the proposed detectors, especially in scenarios with low signal-to-noise ratios and poor channel conditions.
BibTeX
@article{mao2024multi,
  author = {Linlin Mao and Shefeng Yan and Zeping Sui and Hongbin Li},
  title = {{Multi-bit Distributed Detection of Sparse Stochastic Signals over Error-Prone Reporting Channels}},
  journal = {IEEE Trans. Signal Inf. Process. Over Networks},
  year = {2024},
  doi = {10.1109/tsipn.2024.3496253},
}
Figure