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

Abstract

The invention relates to the technical field of optical communication, and particularly discloses a nonlinear compensation method and device based on a stimulated Brillouin scattering technology and a machine learning algorithm. The method comprises the following steps: S1, regenerating and amplifying a narrow-band spectral line in an optical frequency comb by using a stimulated Brillouin scattering (SBS) technology, and improving the carrier-to-noise power ratio (CNPR) of the optical frequency comb; s2, performing phase-sensitive-free coherent detection on the carrier amplified by the SBS in a self-coherent receiver, and eliminating the requirement on a local oscillator; and S3, performing real-time compensation on the nonlinear distortion of the optical fiber transmission link and the transmitting end through a low-complexity machine learning algorithm based on the StyleGAN. The nonlinear compensation device comprises an optical frequency comb module, a self-coherent receiving module and an MLA module. According to the invention, the problem of limited transmission performance caused by nonlinear distortion, random noise and frequency drift in the existing optical communication system can be solved, and the transmission performance and energy efficiency of the optical communication system are optimized.

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 mobile devices with high bandwidth requirements, the existing telecommunication infrastructure can no longer meet future capacity requirements, and optical networks are facing an imminent capacity crisis. Replacing existing single-mode fiber (SMF) with multimode fiber with higher capacity potential is a possible solution, but this solution is limited by nonlinear distortion and signal degradation caused by the optical Kerr effect.

[0003] At present, nonlinear compensation technologies (such as spectrum inversion, phase conjugated double wave and digital back propagation) have solved the deterministic nonlinear problem to a certain extent, but they still have obvious limitations in dealing with random nonlinear distortion, optical noise and frequency drift. In addition, existing technologies generally rely on complex digital signal processing (DSP) circuits, which not only consumes a lot of energy but also makes real-time applications difficult to achieve.

[0004] As a new type of multi-wavelength carrier source, optical frequency comb has attracted much attention for its potential advantages in wavelength division multiplexing (WDM) communications. However, its inherent parameterized noise problem has a significant impact on performance, and a more efficient nonlinear compensation method is urgently needed to simultaneously solve deterministic and random nonlinear problems and optimize the transmission performance and energy efficiency of optical communication systems. Summary of the invention

[0005] In view of 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 object, 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 and eliminating the need for a local oscillator;

[0009] S3. The nonlinear distortion of the optical fiber transmission link and the transmitter is compensated in real time through the 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 comprises:

[0012] S11, extracting part of the spectrum of the optical frequency comb 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 narrow band.

[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 multiplexing optical orthogonal frequency division multiplexing (O-OFDM) system to improve the system spectrum efficiency and reduce the impact of phase noise.

[0019] Preferably, in S3, the machine learning algorithm of the StyleGAN comprises 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 and format them into a constellation diagram suitable for neural network training;

[0021] S32, model training: Generate signal data through mapping network, synthesis network and style modulation network, and optimize the model based on 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 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 realize the extraction of frequency and phase information using the carrier amplified by SBS; the MLA module includes the 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, and its beneficial effects are as follows:

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

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

[0030] (3) The present invention uses a low-complexity machine learning algorithm based on StyleGAN to perform real-time compensation for the nonlinear distortion of the optical fiber and the transmitting end, thereby optimizing 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 It is a schematic diagram of the overall architecture of the nonlinear compensation method for an optical communication system 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, wherein SI: spectral inversion, PCTW: phase conjugate dual wave, DBP: digital back propagation;

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

[0035] In order to make the technical solutions, advantages and purposes of the present invention clearer, the technical solutions of the embodiments of the present invention are clearly and completely described below. The described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of this application.

[0036] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field 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 subjected to phase-insensitive coherent detection, and the need for a local oscillator is eliminated, and the optoelectronics and digital signal processing (DSP) complexity at the receiving end is reduced;

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

[0042] At the receiving end, the narrowband carrier amplified by SBS technology is used to achieve phase-insensitive coherent detection. By locking the signal optical frequency within the Brillouin amplification bandwidth, the need for frequency offset compensation is avoided. Applied to polarization-multiplexed optical orthogonal frequency division multiplexing (O-OFDM) systems, this method significantly improves the system spectrum efficiency and enhances the ability to resist nonlinear phase noise.

[0043] S3. The nonlinear distortion of the optical fiber transmission link and the transmitter is compensated in real time through the 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 transmitting end, a frequency-shifted pump light is generated by extracting part of the optical frequency comb spectrum and performing phase modulation, and then the target frequency is selected by an optical bandpass filter OBPF to achieve narrowband amplification. Regenerative amplification is performed using a chalcogenide-based nonlinear chip or a highly nonlinear fiber HNLF to increase SBS gain and reduce the impact of phase noise. This technology significantly improves the CNPR of the optical frequency comb and provides a higher quality carrier signal 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, and 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 network and synthesis network, and optimize the model based on Wasserstein loss function;

[0050] Data recovery: Use 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) multiple-input multiple-output (MIMO) is applied only when a dual-polarization QAM system is considered.

[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 carrier amplified by SBS; an MLA module, including a 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 currently available.

[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 realizes 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 machine learning algorithm, which provides a new solution for nonlinear compensation of optical communication systems. This method not only solves the nonlinear distortion problem of optical fiber and transmitting end, but also simplifies the design of receiving end, improves the transmission performance and energy efficiency of the system, and 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 solution of the present invention rather than to limit it. 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 solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution 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 and eliminating the need for a local oscillator; S3. The nonlinear distortion of the optical fiber transmission link and the transmitter is compensated in real time through the low-complexity machine learning algorithm MLA based on StyleGAN.

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 of 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 spectrum of the optical frequency comb 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 narrow band.

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 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.

5. The nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 1, characterized in that: 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 is applied to polarization multiplexing optical orthogonal frequency division multiplexing (O-OFDM) system to improve the system spectrum efficiency and reduce the impact of phase noise.

6. The nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 1, characterized in that: 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 and format them into a constellation diagram suitable for neural network training; S32, model training: Generate signal data through mapping network, synthesis network and style modulation network, and optimize the model based on 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.

7. The nonlinear compensation method based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 6, characterized in that: The StyleGAN model gradually improves the resolution of training data through progressive growing.

8. A nonlinear compensation device based on the method according to any one of claims 1 to 7, characterized in that: The device comprises: an optical frequency comb module, a self-coherent receiving module and an MLA module.

9. The nonlinear compensation device based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 8, characterized in that: The optical frequency comb module is used to generate multi-wavelength carriers and realize narrowband amplification through SBS technology; the self-coherence module is used to extract frequency and phase information using the carrier amplified by SBS; the MLA module includes the StyleGAN machine learning algorithm for real-time nonlinear compensation.

10. The nonlinear compensation device based on stimulated Brillouin scattering technology and machine learning algorithm according to claim 7, characterized in that: The device is used for high bit rate wavelength division multiplexing (WDM) communication systems and other optical communication systems based on high-order modulation formats.

Images