- The paper demonstrates a novel MIMO-DFS system that reduces laser frequency noise by over twofold through active drift stabilization and optimized probing codes.
- The experimental setup uses polarization-diversity multiplexing and M-PSK probing codes to achieve 1-meter spatial resolution and detect weak mechanical events.
- The work provides practical guidelines for integrating high-sensitivity DFS with existing telecom infrastructure without the need for inline amplification.
High-Sensitivity MIMO-DFS over Long-Haul SSMF Using Active Laser Drift Stabilization and Probing Code Optimization
Introduction
This paper presents an experimental demonstration of Distributed Fiber Sensing (DFS) over 100 km of unamplified Standard Single-Mode Fiber (SSMF) based on a polarization-multiplexed coherent transmission scheme—Multi-Input Multi-Output Distributed Fiber Sensing (MIMO-DFS)—augmented with active laser drift stabilization and optimized digital probing codes. The primary objective is to extend the reach and sensitivity of DFS using existing telecom infrastructure, critical for real-world distributed acoustic and environmental sensing applications, without resorting to inline optical amplification. The paper identifies laser frequency noise, particularly its low-frequency components, as the limiting factor for DFS sensitivity in long-haul interrogation regimes and systematically quantifies its impact while validating the mitigation potential of feedback-based laser stabilization.
Impact of Laser Frequency Noise on DFS Sensitivity
DFS using coherent techniques relies on Rayleigh backscatter interrogation, with the primary sensitivity bottleneck arising from both the additive noise of the detection electronics and the coherence properties of the light source. The latter is shown, especially in long-span SSMF applications and time-spread code-based interrogation, to severely constrain the noise floor. Simulations performed in this work reveal that, for realistic system parameters, 50% of the DFS noise floor results from laser frequency noise in the 800 Hz–6 kHz bandwidth (centered around 2 kHz), determined by the convolved timescales of the probing code and fiber round-trip time. When substituting the Lorentzian model for empirical PSD measurements of a commercial high-coherence laser, this critical spectral window shifts to 100 Hz–2.5 kHz, reflecting the prevalence of technical noise at lower frequencies in practical laser sources. These results provide actionable guidelines for engineering laser stabilization: frequency noise must be suppressed predominantly in this band to achieve meaningful sensitivity improvements.
Active Laser Drift Stabilization Architecture
The paper implements an external feedback control loop using an Optical Frequency Discriminator (OFD) for real-time compensation of slow laser drifts. The OFD architecture is specifically tailored to deliver high-bandwidth (≫100 kHz) response and is integrated with a precision low-noise fiber laser. Experimental PSD measurements show a reduction of the laser frequency noise by three orders of magnitude below 10 kHz after stabilization. Notably, the design avoids excessive complexity, integrating compact components that are suitable for practical sensing deployments in telecom bands. The servo control parameters are optimized for strong locking at the relevant resonance, ensuring minimal noise within the desired spectral band.
Experimental validation is carried out on a 123 km SSMF testbed using polarization-diversity multiplexing and optimized M-PSK probing codes with 1-meter spatial resolution. The code parameters are judiciously selected to compress code period, thus shifting the most noise-sensitive region to higher frequencies where laser stabilization is more effective, further relaxing the temporal requirements on the stabilization loop. The noise floor measurements along the first 100 km show a clear and dramatic reduction—a greater than twofold decrease—when active stabilization is engaged. This directly enables the detection of weak mechanical events far beyond 100 km without the use of Raman amplification or distributed gain flattening measures. The system detects a 120 Hz vibration at 101 km, with a spectral SNR that is more than 10 dB above the suppressed noise floor, confirming high fidelity for event discrimination in a long-haul, real-world applicable scenario.
Theoretical and Practical Implications
The study provides a new methodology for quantifying and engineering sensitivity bottlenecks in coherent DFS tied to laser frequency noise PSD and fiber-probing parameters. The demonstration that MIMO-DFS sensitivity is predominantly determined by slow laser fluctuations enables targeted system-level improvements: it is now feasible to design laser control loops and select digital codes that maximize reach and sensitivity without the penalties associated with high optical power or amplification.
From a telecom operator’s perspective, the compatibility of MIMO-DFS with live Dense Wavelength Division Multiplexing (DWDM) transmission is critical. The polarization diversity architecture and digital probing codes do not impair co-propagating data channels, ensuring DFS deployment does not impose penalties on network OPEX/CAPEX or require dark fiber. On the theoretical front, the match between modeled and measured performance solidifies the Rayleigh backscatter–prioritized models for noise floor estimation in long-haul, high-resolution distributed sensing.
Prospects for Future Development
Further work should focus on integration of active stabilization into field-deployed telecom lasers, advanced servo algorithms to address technical noise at sub-100 Hz frequencies, and code design that pushes spatial and temporal resolution to new limits. The insights on frequency noise spectral occupation can be extended to new classes of sensing lasers and may guide the integration of quantum-limited sources for ultimate noise reduction. The demonstrated reach and sensitivity open further study into real-time, city-scale (100 km+) distributed sensing for environmental monitoring, intrusion detection, and structural health diagnostics.
Conclusion
This work delivers a thorough analysis and experimental demonstration showing that targeted laser frequency noise stabilization, informed by quantification of the probe code and fiber timescales, enables long-haul (100+ km) MIMO-DFS with high sensitivity. The reduction of the practical DFS noise floor by more than twofold without amplification, using active feedback and code optimization, confirms the viability of deploying distributed sensing over existing optical telecom infrastructure. The approach sets a new operational standard for scalable, non-intrusive, and highly sensitive environmental telemetry leveraging global fiber networks.