A LCX-Based Intrusion-Detection Sensor Using a Broadband Noise Signal

Hang Xu, Ruixin Xie, Hong Han, Zhaoxia Zhang, Jianguo Zhang, Li Liu, Bingjie Wang, Liqiang LiIEEE Access 2019 · DOI

Introduction

Securing critical infrastructure such as airports, military bases, and power plants requires sensors that are both reliable and difficult to evade. Traditional perimeter security technologies like laser scanners, infrared sensors, and video cameras often fail in adverse weather conditions or require a direct line of sight. This paper presents a significant advancement in perimeter security: an invisible, high-precision system based on Leaky Coaxial Cables (LCXs) and broadband noise signals.

System Architecture

The proposed system utilizes a pair of LCXs to create an electromagnetic monitoring field. Unlike traditional sensors that use single-frequency waves or radio pulses, this system employs broadband white noise to achieve superior spatial resolution and interference immunity.

Schematic of the LCX-based intrusion detection system using broadband noise and leaky coaxial cables.

As shown in the system diagram, a custom-designed generator produces a wide-spectrum noise signal. This signal is split into a reference path and a transmit path. The transmit signal is sent through a "transmit" LCX, while a parallel "receive" LCX captures the reflected signals. When an intruder enters the "Protected Area" between the cables, they reflect the electromagnetic waves, creating a detectable change in the received signal.

Methodology: The Power of Broadband Noise

The core innovation of this sensor lies in its use of broadband white noise. By leveraging the unique autocorrelation properties of noise, the system can pinpoint the exact location of an intruder with high precision.

Characteristics of the broadband noise signal: (a) time-domain waveform, (b) power spectrum showing 1.4 GHz bandwidth, and (c) thumbtack-like autocorrelation function.

White noise possesses a "thumbtack-like" autocorrelation function, meaning it only matches itself perfectly at zero time delay. With a bandwidth of 1.4 GHz, the system achieves a range resolution of approximately 30 cm. The correlation processing compares the received echo to the original reference signal. To eliminate static reflections from the environment, a differential analysis is performed by subtracting the baseline "pre-intrusion" correlation curve from the active monitoring signal.

Experimental Results and Key Insights

The system was evaluated through several real-world scenarios, demonstrating its robustness and accuracy:

  • High Precision: The sensor achieved a location error of only 10 cm and a spatial resolution of 30 cm, significantly outperforming traditional LCX sensors which typically offer 3–6 meter resolution.
  • Multi-Target Tracking: The system successfully detected and tracked up to four intruders simultaneously without signal degradation.
  • Target Discrimination: The height of the correlation peak is proportional to the size of the intruder, allowing the system to distinguish between human threats and smaller animals, thereby reducing false alarms.
  • RFI Immunity: Because the system only correlates with its own unique random reference signal, it is inherently immune to external radio frequency interference (RFI) from other electronic devices.

Conclusion

By treating noise as a high-precision ranging tool rather than an unwanted interference, this research provides a robust solution for invisible perimeter monitoring. While current implementations rely on high-speed laboratory equipment, future work will focus on developing cost-effective, specialized hardware to scale the system for long-distance perimeter security.

Summary updated

BibTeX
@article{xu2019lcx,
  author = {Hang Xu and Ruixin Xie and Hong Han and Zhaoxia Zhang and Jianguo Zhang and Li Liu and Bingjie Wang and Liqiang Li},
  title = {{A LCX-Based Intrusion-Detection Sensor Using a Broadband Noise Signal}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2951576},
}
Figure