Nonlinear compensation method and device based on stimulated Brillouin scattering technology and machine learning algorithm

Abstract

The present invention relates to the field of optical communication technology, and specifically discloses a nonlinear compensation method and device based on stimulated Brillouin scattering technology and machine learning algorithm. The method comprises: S1, using stimulated Brillouin scattering (SBS) technology to regenerate and amplify narrowband spectral lines in an optical frequency comb, thereby improving the carrier-to-noise power ratio (CNPR) of the optical frequency comb; S2, performing phase-insensitive coherent detection on the SBS-amplified carrier in a self-coherent receiver, and eliminating the need for a local oscillator; S3, using a low-complexity machine learning algorithm based on StyleGAN to perform real-time compensation for the nonlinear distortion of the optical fiber transmission link and the transmitting end. The nonlinear compensation device comprises an optical frequency comb module, a self-coherent receiving module, and an MLA module. The present invention can solve the transmission performance limitation problem caused by nonlinear distortion, random noise, and frequency drift in existing optical communication systems, and optimize the transmission performance and energy efficiency of optical communication systems.

Description

Technical Field

[0001] The present invention relates to the field of optical communication technology, and in particular to a nonlinear compensation method and device based on stimulated Brillouin scattering technology and a machine learning algorithm. Background Art

[0002] With the explosive growth of the internet and bandwidth-hungry mobile devices, existing telecommunications infrastructure is struggling to meet future capacity demands, and optical networks face an imminent capacity crisis. Replacing existing single-mode fiber (SMF) with multimode fiber, which offers higher capacity potential, is a potential solution, but this approach is limited by nonlinear distortion and signal degradation caused by the optical Kerr effect.

[0003] Currently, nonlinear compensation techniques (such as spectrum inversion, phase conjugation double wave, and digital backpropagation) have, to a certain extent, addressed deterministic nonlinearity issues. However, they still have significant limitations in addressing random nonlinear distortion, optical noise, and frequency drift. Furthermore, existing technologies generally rely on complex digital signal processing (DSP) circuits, which are energy-intensive and difficult to implement in real-time.

[0004] As a novel multi-wavelength carrier source, optical frequency combs have attracted considerable attention for their potential advantages in wavelength division multiplexing (WDM) communications. However, their inherent parametric noise significantly limits their performance. A more efficient nonlinear compensation method is urgently needed to simultaneously address both deterministic and random nonlinearities and optimize the transmission performance and energy efficiency of optical communication systems. Summary of the Invention

[0005] In response to the above-mentioned problems in the prior art, the present invention provides a nonlinear compensation method and device based on stimulated Brillouin scattering technology and machine learning algorithm, which can solve deterministic and random nonlinear problems, optimize the transmission performance and energy efficiency of optical communication systems, significantly improve the performance of optical communication systems, and reduce energy consumption and complexity.

[0006] To achieve the above objectives, the present invention proposes a nonlinear compensation method based on stimulated Brillouin scattering technology and a machine learning algorithm, the method comprising:

[0007] S1. Use stimulated Brillouin scattering (SBS) technology to regenerate and amplify the narrowband spectral lines in the optical frequency comb to improve the carrier-to-noise power ratio (CNPR) of the optical frequency comb.

[0008] S2. Phase-insensitive coherent detection of the SBS-amplified carrier in a self-coherent receiver, eliminating the need for a local oscillator.

[0009] S3. Real-time compensation of nonlinear distortion of optical fiber transmission link and transmitter is performed through low-complexity machine learning algorithm MLA based on StyleGAN.

[0010] Preferably, the real-time compensation includes compensation for nonlinear effects of self-phase modulation SPM, cross-phase modulation XPM and four-wave mixing FWM.

[0011] Preferably, in S1, the SBS technology for regenerating and amplifying narrowband spectral lines in an optical frequency comb includes:

[0012] S11, extracting part of the optical frequency comb spectrum and performing phase modulation to generate frequency-shifted pump light;

[0013] S12. Use an optical bandpass filter (OBPF) in the pump light path to select the target frequency and amplify the narrowband.

[0014] Preferably, the SBS technology is used to achieve narrowband spectrum line regeneration and amplification in the optical frequency comb through a chalcogenide-based nonlinear chip or a highly nonlinear optical fiber HNLF.

[0015] Preferably, in S2, the self-coherent receiver is implemented by the following steps:

[0016] S21, using the SBS-amplified carrier as an ideal local oscillator to recover frequency and phase information from the signal light;

[0017] S22, locking the signal light frequency within the Brillouin amplification bandwidth;

[0018] S23 is applied to polarization-multiplexed optical orthogonal frequency division multiplexing (O-OFDM) systems to improve system spectrum efficiency and reduce the impact of phase noise.

[0019] Preferably, in S3, the machine learning algorithm of the StyleGAN includes the following steps:

[0020] S31. Data preparation: Collect and preprocess the I in-phase and Q quadrature component data of the optical fiber transmission system, normalize them, and format them into a constellation diagram suitable for neural network training;

[0021] S32, Model Training: Generate signal data through the mapping network, synthesis network and style modulation network, and optimize the model based on the Wasserstein loss function with gradient penalty;

[0022] S33, Data Recovery: Generate a clear version of the distorted signal through training, and compensate for the nonlinear effects of the optical fiber and the transmitter in real time.

[0023] Preferably, the StyleGAN model gradually improves the resolution of the training data through progressive growth.

[0024] A nonlinear compensation device based on the method comprises: an optical frequency comb module, a self-coherent receiving module and an MLA module.

[0025] Preferably, the optical frequency comb module is used to generate multi-wavelength carriers and realize narrowband amplification through SBS technology; the autocoherence module is used to extract frequency and phase information using the carriers amplified by SBS; and the MLA module includes a StyleGAN machine learning algorithm for real-time nonlinear compensation.

[0026] Preferably, the device is suitable for high bit rate wavelength division multiplexing (WDM) communication systems and other optical communication systems based on high-order modulation formats.

[0027] Therefore, the present invention proposes a nonlinear compensation method and device based on stimulated Brillouin scattering technology and machine learning algorithm, which has the following beneficial effects:

[0028] (1) This invention utilizes stimulated Brillouin scattering (SBS) technology to regenerate and amplify narrowband spectral lines in an optical frequency comb, significantly improving the carrier-to-noise power ratio of the frequency comb and effectively reducing parameterized noise. SBS is also used in self-coherent receivers to achieve phase-insensitive coherent detection, simplifying receiver design and enhancing immunity to nonlinear phase noise.

[0029] (2) The present invention provides a phase-insensitive coherent detection solution, reducing the complexity of optoelectronic components.

[0030] (3) The present invention uses a low-complexity machine learning algorithm based on StyleGAN to compensate for the nonlinear distortion of the optical fiber and the transmitting end in real time and optimize the transmission performance.

[0031] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the overall architecture of the optical communication system nonlinear compensation method of the present invention;

[0033] Figure 2 Schematic diagram of the optical frequency comb regeneration and amplification process based on SBS of the present invention; (a) is a schematic diagram of comb regeneration (upward) and carrier extraction (downward) based on SBS; (b) is a schematic diagram of the comparison between the machine learning algorithm (MLA) and the most advanced deterministic nonlinear compensator, where SI: spectral inversion, PCTW: phase conjugate dual wave, DBP: digital back propagation;

[0034] Figure 3Schematic diagram of the constellation diagram of the application of StyleGAN in nonlinear compensation of the present invention, wherein the inset constellation diagram: before and after machine learning of 16QAM signal, CD: dispersion. DETAILED DESCRIPTION

[0035] To make the technical solutions, advantages, and purposes of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below. The described embodiments are part of the embodiments of the present invention, not all of them. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0036] Unless otherwise defined, technical or scientific terms used in the present invention shall have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs.

[0037] like Figure 1 As shown, the overall architecture diagram of the nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm of the present invention includes a block diagram of a WDM transmitter and an autocoherent receiver combined with a DSP based on machine learning, and an autocoherent detection process using SBS for comb regeneration or carrier extraction.

[0038] like Figure 1 As shown, the present invention provides a nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm, the method comprising:

[0039] S1. Use stimulated Brillouin scattering (SBS) technology to regenerate and amplify the narrowband spectral lines in the optical frequency comb to improve the carrier-to-noise power ratio (CNPR) of the optical frequency comb.

[0040] S2. In a self-coherent receiver, the SBS-amplified carrier is coherently detected without phase sensitivity, eliminating the need for a local oscillator and reducing the optoelectronics and digital signal processing (DSP) complexity at the receiving end.

[0041] The process of self-coherence detection by the SBS-based self-coherence receiver is as follows:

[0042] At the receiver, a narrowband carrier amplified by SBS technology enables phase-insensitive coherent detection. By locking the signal optical frequency within the Brillouin amplification bandwidth, the need for frequency offset compensation is eliminated. Applied to polarization-multiplexed optical orthogonal frequency division multiplexing (O-OFDM) systems, this approach significantly improves system spectral efficiency and enhances immunity to nonlinear phase noise.

[0043] S3. Real-time compensation of nonlinear distortion of optical fiber transmission link and transmitter is performed through low-complexity machine learning algorithm MLA based on StyleGAN.

[0044] like Figure 2 As shown in (a), the present invention provides an optical frequency comb regeneration and amplification process based on SBS.

[0045] The optical frequency comb regeneration amplification process based on SBS is:

[0046] At the transmitter, a portion of the optical frequency comb spectrum is extracted and phase-modulated to generate frequency-shifted pump light. This is then selected by an optical bandpass filter (OBPF) to achieve narrowband amplification. Regenerative amplification is performed using a chalcogenide-based nonlinear chip or highly nonlinear fiber (HNLF) to increase SBS gain and reduce the effects of phase noise. This technology significantly improves the CNPR of the optical frequency comb, providing higher-quality carrier signals for multi-wavelength modulation in high-bit-rate wavelength division multiplexing (WDM) systems.

[0047] The present invention collects I in-phase and Q quadrature component data of the optical fiber transmission system, constructs and trains the StyleGAN model to achieve nonlinear compensation. The specific steps include:

[0048] Data preparation: normalize and format the signal data into a constellation diagram suitable for training;

[0049] Model training: Generate signal data through mapping networks and synthesis networks, and optimize the model based on the Wasserstein loss function;

[0050] Data recovery: Using trained models to compensate for nonlinear distortions in signals in real time.

[0051] like Figure 3 As shown, the present invention proposes a StyleGAN nonlinear equalizer for 16QAM coherent optical transmission system. In the previous step for linear equalization, time division multiplexing (TDE) is applied only when considering a dual-polarization QAM system.

[0052] The present invention also provides a nonlinear compensation device based on the above method, including: an optical frequency comb module, used to generate multi-wavelength carriers and realize narrowband amplification through SBS technology; a self-coherent receiving module, used to extract frequency and phase information using the SBS-amplified carriers; an MLA module, including the StyleGAN machine learning algorithm, for real-time nonlinear compensation.

[0053] like Figure 2 As shown in (b), the machine learning algorithm MLA of the present invention is superior to the most advanced deterministic nonlinear compensator.

[0054] Experimental results show that the method of the present invention significantly reduces the bit error rate (BER), improves the signal-to-noise ratio (SNR), and achieves low-complexity and high-efficiency nonlinear compensation.

[0055] Therefore, the present invention provides a nonlinear compensation method and device based on stimulated Brillouin scattering technology and a machine learning algorithm, offering a novel solution for nonlinear compensation in optical communication systems. This method not only solves the problem of nonlinear distortion in optical fibers and transmitters, but also simplifies receiver design, improving system transmission performance and energy efficiency. It has broad application prospects and significant economic value.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm, characterized in that: The method includes: S1. Use stimulated Brillouin scattering (SBS) technology to regenerate and amplify the narrowband spectral lines in the optical frequency comb to improve the carrier-to-noise power ratio (CNPR) of the optical frequency comb. S2. Phase-insensitive coherent detection of the SBS-amplified carrier in a self-coherent receiver, eliminating the need for a local oscillator. S3, using the low-complexity machine learning algorithm (MLA) based on StyleGAN to compensate for the nonlinear distortion of the optical fiber transmission link and the transmitter in real time; In S2, the self-coherent receiver is implemented by the following steps: S21, using the SBS-amplified carrier as an ideal local oscillator to recover frequency and phase information from the signal light; S22, locking the signal light frequency within the Brillouin amplification bandwidth; S23, applied to polarization-multiplexed optical orthogonal frequency division multiplexing (O-OFDM) systems, improving system spectrum efficiency and reducing the impact of phase noise; In S3, the machine learning algorithm of StyleGAN includes the following steps: S31. Data preparation: Collect and preprocess the I in-phase and Q quadrature component data of the optical fiber transmission system, normalize them, and format them into a constellation diagram suitable for neural network training; S32, Model Training: Generate signal data through the mapping network, synthesis network and style modulation network, and optimize the model based on the Wasserstein loss function with gradient penalty; S33, Data Recovery: Generate a clear version of the distorted signal through training, and compensate for the nonlinear effects of the optical fiber and the transmitter in real time.

2. The nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 1, characterized in that: The real-time compensation includes compensation for nonlinear effects of self-phase modulation SPM, cross-phase modulation XPM and four-wave mixing FWM.

3. The nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 1, characterized in that: In S1, the SBS technology for regenerating and amplifying narrowband spectral lines in an optical frequency comb includes: S11, extracting part of the optical frequency comb spectrum and performing phase modulation to generate frequency-shifted pump light; S12. Use an optical bandpass filter (OBPF) in the pump light path to select the target frequency and amplify the narrowband.

4. The nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 3, characterized in that: The SBS technology is used to regenerate and amplify narrowband spectral lines in an optical frequency comb through a chalcogenide-based nonlinear chip or a highly nonlinear fiber HNLF.

5. The nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 1, characterized in that: The StyleGAN model gradually improves the resolution of training data through progressive growing.

6. A nonlinear compensation device based on the method according to any one of claims 1 to 5, characterized in that: The device includes an optical frequency comb module, a self-coherent receiver and an MLA module.

7. The nonlinear compensation device based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 6, characterized in that: The optical frequency comb module is used to generate multi-wavelength carriers and achieve narrowband amplification through SBS technology; the self-coherent receiver uses the SBS-amplified carriers to extract frequency and phase information; the MLA module includes the StyleGAN machine learning algorithm for real-time nonlinear compensation.

8. The nonlinear compensation device based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 6, characterized in that: The device is used in high bit rate wavelength division multiplexing (WDM) communication systems and other optical communication systems based on high-order modulation formats.

Images