A method for signal amplification and a signal amplifier for submarine cable transmission signals

By using a two-stage cascaded signal amplifier and a signal amplification method combining forward and backward pumping modules with Raman gain media, the problem of broadband signal amplification in submarine optical cable systems was solved. This achieved flat and efficient gain for S-band, C-band, and L-band signals, thereby improving the transmission capacity of submarine optical cable systems.

CN123122671APending Publication Date: 2026-11-17BEIJING INST OF TECH +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611594513.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-15
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing single-band optical amplification schemes are insufficient to meet the requirements of submarine optical cable systems for broadband signal amplification, thus limiting the transmission capacity of submarine optical cable systems.

Method used

A two-stage cascaded signal amplifier is used, with the forward pump module and the backward pump module providing gains in the S-band, C-band, and L-band, respectively. The first-stage and second-stage amplification components are connected through forward and backward couplers. The signal is amplified using Raman gain medium, and the pump light reverse transmission is suppressed through isolation components to avoid interference.

Benefits of technology

It achieves flat and efficient gain for S-band, C-band and L-band signals, improving the transmission capacity of submarine optical cable systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122671A_ABST
    Figure CN123122671A_ABST
Patent Text Reader

Abstract

This application provides a signal amplification method and signal amplifier for submarine cable transmission signals, relating to the field of optical fiber communication technology. The signal amplifier includes a forward pump module, a backward pump module, a forward coupler, a backward coupler, an isolation component, a first-stage amplification component, and a second-stage amplification component. The input end of the forward coupler is connected to the forward pump module and the submarine cable transmission optical fiber, and its output end is connected to the first-stage amplification component. The other end of the first-stage amplification component is connected to the second-stage amplification component through the isolation component, and the other end of the second-stage amplification component is connected to the bidirectional port of the backward coupler. The input end of the backward coupler is connected to the backward pump module. The forward pump module generates pump light that provides gain for S-band signals; the backward pump module generates pump light that provides gain for C-band and L-band signals; the first-stage and second-stage amplification components amplify the submarine cable transmission signal. This solution can improve the transmission capability of submarine optical cable systems.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of optical fiber communication technology, and in particular to a signal amplification method and signal amplifier for submarine cable transmission signals. Background Technology

[0002] In recent years, with the rapid development of global internet services, cross-regional data interaction, and emerging applications such as artificial intelligence, the demand for data transmission worldwide has continued to grow, driving submarine optical cable systems to develop towards ultra-long-distance, high-capacity, and multi-band collaborative transmission.

[0003] Currently, submarine optical cable systems typically employ rare-earth-doped fiber amplifiers to compensate for power attenuation of optical signals during long-distance transmission, enabling transoceanic information transmission. However, rare-earth-doped fiber amplifiers are limited by the energy level structure of the doping materials; different energy level structures of doping materials can only be used for signal amplification in specific wavelength bands. For example, erbium-doped fiber amplifiers primarily amplify in the C-band (Conventional Band), while thulium-doped fiber amplifiers primarily amplify in the S-band (Short Band).

[0004] Therefore, existing single-band optical amplification schemes are insufficient to meet the requirements of submarine optical cable systems for broadband submarine cable signal amplification, thus limiting the transmission capacity of submarine optical cable systems. Summary of the Invention

[0005] The purpose of this application is to provide a signal amplification method and a signal amplifier for submarine cable transmission signals, so as to improve the transmission capability of submarine optical cable systems. The specific technical solution is as follows:

[0006] In a first aspect, embodiments of this application provide a signal amplifier for submarine cable transmission signals, the signal amplifier comprising: a forward pump module, a forward coupler, a first-stage amplification component, an isolation component, a second-stage amplification component, a backward coupler, and a backward pump module;

[0007] The two input terminals of the forward coupler are respectively connected to the forward pump module and the submarine cable transmission optical fiber; the output terminal of the forward coupler is connected to one end of the first-stage amplification component, and the other end of the first-stage amplification component is connected to the input terminal of the isolation component; the output terminal of the isolation component is connected to one end of the second-stage amplification component, and the other end of the second-stage amplification component is connected to the bidirectional port of the backward coupler; the input terminal of the backward coupler is connected to the backward pump module.

[0008] The isolation component suppresses the transmission of pump light from the secondary amplification component to the primary amplification component;

[0009] The forward pump module generates a first pump light that provides gain for the S-band signal; the backward pump module generates a second pump light that provides gain for the C-band and L-band signals.

[0010] The first-stage amplification component uses a first Raman gain medium to amplify the submarine cable transmission signal transmitted through the submarine cable transmission optical fiber under the action of the first pump light to obtain a first signal; the submarine cable transmission signal is an optical signal including S-band, C-band and L-band.

[0011] The secondary amplification component uses a second Raman gain medium to amplify the first signal under the action of the second pump light to obtain the second signal.

[0012] Secondly, embodiments of this application provide a signal amplification method for submarine cable transmission signals. The method employs a signal amplifier for signal amplification. The signal amplifier includes: a forward pump module, a backward pump module, a forward coupler, a backward coupler, an isolation component, a first-stage amplification component, and a second-stage amplification component. The two input terminals of the forward coupler are respectively connected to the forward pump module and the submarine cable transmission optical fiber. The output terminal of the forward coupler is connected to one end of the first-stage amplification component, and the other end of the first-stage amplification component is connected to the input terminal of the isolation component. The output terminal of the isolation component is connected to one end of the second-stage amplification component, and the other end of the second-stage amplification component is connected to the bidirectional port of the backward coupler. The input terminal of the backward coupler is connected to the backward pump module. The method includes:

[0013] The forward pump module outputs a first pump light that provides gain for an S-band signal;

[0014] The forward coupler couples and outputs the first pump light and the submarine cable transmission signal transmitted by the submarine cable transmission fiber; the submarine cable transmission signal is an optical signal including S-band, C-band and L-band.

[0015] The first-stage amplification component amplifies the submarine cable transmission signal under the action of the first pump light to obtain the first signal;

[0016] The backward pump module outputs a second pump light that provides gain for C-band and L-band signals. The isolation component suppresses the transmission of the pump light from the secondary amplification component to the primary amplification component. Under the action of the second pump light, the secondary amplification component amplifies the first signal to obtain the second signal.

[0017] Beneficial effects of the embodiments in this application:

[0018] The signal amplifier provided in this application embodiment couples the submarine cable transmission signal transmitted via the submarine cable optical fiber with the first pump light generated by the forward pump module through a forward coupler, and outputs it to a first-stage amplification component. The first-stage amplification component uses a first Raman gain medium. The first pump light, which provides gain for the S-band signal, undergoes stimulated Raman scattering in this medium, transferring energy to the S-band signal in the submarine cable transmission signal, thereby amplifying the S-band signal. The amplified submarine cable transmission signal is transmitted to a second-stage amplification component through an isolation component. Furthermore, the second pump light, which provides gain for the C-band and L-band signals generated by the backward pump module, is coupled through a backward coupler and injected into the second-stage amplification component from the other end of the second-stage amplification component. The second-stage amplification component uses a second Raman gain medium. The second pump light undergoes stimulated Raman scattering in this medium, transferring energy to the C-band and L-band signals in the submarine cable transmission signal, thereby amplifying the C-band and L-band signals.

[0019] The aforementioned signal amplifier enables signal amplification for broadband submarine cable transmission signals. Furthermore, this scheme injects a first pump light before the first-stage amplification component and a second pump light after the second-stage amplification component. Because the isolation component between the two amplification components effectively suppresses the reverse propagation of the second pump light, the pump light used for L-band gain is almost unable to enter the first-stage amplification component. This effectively prevents interference from the L-band pump light on the S-band signal gain process within the first-stage amplification component. Thus, compared to existing Raman amplifiers where the energy of the pump light used for S-band gain is transferred to the pump light used for L-band gain, resulting in over-amplification of the L-band signal and insufficient gain in the S-band, this scheme, by preventing the L-band pump light from entering the first-stage amplification component, allows for focused amplification of the S-band signal within the first-stage amplification component. Therefore, the two-stage cascaded amplification components achieve flat and efficient gain for S-band, C-band, and L-band signals. Using the signal amplifier provided in this solution in a submarine optical cable system can effectively amplify broadband submarine cable transmission signals, thereby improving the transmission capacity of the submarine optical cable system.

[0020] Of course, implementing any product or method of this application does not necessarily require achieving all of the advantages described above at the same time. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other embodiments can be obtained based on these drawings.

[0022] Figure 1 This is a schematic diagram of the structure of a first type of signal amplifier provided in an embodiment of this application;

[0023] Figure 2 This is a schematic diagram of the structure of a second type of signal amplifier provided in an embodiment of this application;

[0024] Figure 3 A flowchart illustrating the adjustment of configuration parameters for a pump light source, provided as an embodiment of this application;

[0025] Figure 4 A flowchart illustrating how to obtain initial configuration parameters of a pump light source, as provided in an embodiment of this application;

[0026] Figure 5 A flowchart illustrating a training method for an inverse network provided in this application embodiment;

[0027] Figure 6 This is a schematic diagram of the structure of a third type of signal amplifier provided in an embodiment of this application;

[0028] Figure 7 This is a schematic diagram of the structure of the fourth signal amplifier provided in the embodiments of this application;

[0029] Figure 8 A flowchart illustrating a signal amplification method provided in an embodiment of this application;

[0030] Figure 9 A schematic diagram illustrating a specific example of a signal amplifier provided in an embodiment of this application;

[0031] Figure 10 A schematic diagram illustrating the processing steps of a forward network and a reverse network provided in an embodiment of this application;

[0032] Figure 11 This is a schematic diagram of the gain-noise figure curve of a signal amplifier provided in an embodiment of this application;

[0033] Figure 12 A schematic diagram comparing the gain-noise figure curves of the signal amplifier provided in this application embodiment with those of a conventional solution;

[0034] Figure 13 This is a schematic diagram of the gain curves of the signal amplifier provided in the embodiments of this application under different target gains;

[0035] Figure 14 A schematic diagram illustrating the bit error rate of a C-band 30Gbaud 16QAM wavelength division multiplexed signal amplified by the signal amplifier provided in the application embodiment;

[0036] Figure 15A schematic diagram illustrating the bit error rate of the signal amplifier provided in the application embodiment after amplifying an L-band 30Gbaud 16QAM wavelength division multiplexed signal;

[0037] Figure 16 A schematic diagram illustrating the bit error rate of a C-band 50Gbaud 16QAM wavelength division multiplexed signal amplified by the signal amplifier provided in the application embodiment;

[0038] Figure 17 This is a schematic diagram showing the bit error rate of a 50Gbaud 16QAM wavelength division multiplexed signal amplified by the signal amplifier provided in the application embodiment. Detailed Implementation

[0039] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art based on this application are within the scope of protection of this application.

[0040] like Figure 1 As shown, this application embodiment provides a signal amplifier for submarine cable transmission signals, including a forward pump module 110, a forward coupler 120, a first-stage amplification component 130, an isolation component 140, a second-stage amplification component 150, a backward coupler 160, and a backward pump module 170. The two input terminals of the forward coupler 120 are connected to the forward pump module 110 and the submarine cable transmission optical fiber; the output terminal of the forward coupler 120 is connected to one end of the first-stage amplification component 130, and the other end of the first-stage amplification component 130 is connected to the input terminal of the isolation component 140; the output terminal of the isolation component 140 is connected to one end of the second-stage amplification component 150, and the other end of the second-stage amplification component 150 is connected to the bidirectional port of the backward coupler 160; The input terminal of the coupler 160 is connected to the backward pump module 170; the isolation component 140 suppresses the transmission of pump light from the secondary amplification component 150 to the primary amplification component; the forward pump module 110 generates a first pump light to provide gain for the S-band signal; the backward pump module 170 generates a second pump light to provide gain for the C-band and L-band signals; the primary amplification component 130 uses a first Raman gain medium to amplify the submarine cable transmission signal under the action of the first pump light to obtain a first signal; the submarine cable transmission signal includes optical signals in the S-band, C-band and L-band; the secondary amplification component 150 uses a second Raman gain medium to amplify the first signal under the action of the second pump light to obtain a second signal.

[0041] The signal amplifier provided in this application embodiment couples the submarine cable transmission signal transmitted via the submarine cable optical fiber with the first pump light generated by the forward pump module through a forward coupler, and then outputs it to a first-stage amplification component. The first-stage amplification component uses a first Raman gain medium. The first pump light, used to provide gain for the S-band signal, undergoes stimulated Raman scattering in this medium, transferring energy to the S-band signal in the submarine cable transmission signal, thereby amplifying the S-band signal. The amplified submarine cable transmission signal is transmitted to a second-stage amplification component through an isolation component. Furthermore, the second pump light, generated by the backward pump module, which provides gain for the C-band and L-band signals, is coupled through a backward coupler and injected into the second-stage amplification component from the other end of the second-stage amplification component. The second-stage amplification component uses a second Raman gain medium. The second pump light undergoes stimulated Raman scattering in this medium, transferring energy to the C-band and L-band signals in the submarine cable transmission signal, thereby amplifying the C-band and L-band signals.

[0042] The aforementioned signal amplifier enables signal amplification for broadband submarine cable transmission signals. Furthermore, this scheme injects a first pump light before the first-stage amplification component and a second pump light after the second-stage amplification component. Because the isolation component between the two amplification components effectively suppresses the reverse propagation of the second pump light, the pump light used for L-band gain is almost unable to enter the first-stage amplification component. This effectively prevents interference from the L-band pump light on the S-band signal gain process within the first-stage amplification component. Thus, compared to existing Raman amplifiers where the energy of the pump light used for S-band gain is transferred to the pump light used for L-band gain, resulting in over-amplification of the L-band signal and insufficient gain in the S-band, this scheme, by preventing the L-band pump light from entering the first-stage amplification component, allows for focused amplification of the S-band signal within the first-stage amplification component. Therefore, the two-stage cascaded amplification components achieve flat and efficient gain for S-band, C-band, and L-band signals. Using the signal amplifier provided in this solution in a submarine optical cable system can effectively amplify broadband submarine cable transmission signals, thereby improving the transmission capacity of the submarine optical cable system.

[0043] The forward pump module 110 may include at least one pump light source. It is understood that, since the pump light is used for a signal with a gain of approximately 100 nm above its center wavelength, in practical applications, the center wavelength of the pump light source in the forward pump module can be determined based on the wavelength of the S-band used for the submarine cable transmission signal.

[0044] For example, if the submarine cable transmission signal uses a wavelength of 1485nm, then the forward pumping module may include at least one pump light source with a center wavelength of 1385nm. For example, if the submarine cable transmission signal uses wavelengths of 1465nm and 1485nm, then the forward pumping module may include at least one pump light source with a center wavelength of 1365nm and at least one pump light source with a center wavelength of 1385nm.

[0045] The forward coupler 120 can be a broadband wavelength division multiplexer (B-WDM) or a pump / signal combiner, etc., to combine two optical signals into a single optical fiber for transmission.

[0046] The forward coupler includes two input terminals and one output terminal. Specifically, the forward coupler may include a pump port, a signal port, and an output port. The pump port is connected to the output terminal of the forward pump module and is used to receive the first pump light output by the forward pump module. The signal port is connected to the submarine cable transmission fiber and is used to receive the submarine cable transmission signal transmitted by the submarine cable transmission fiber. The output port is connected to one end of the first-stage amplification component and is used to output the combined light obtained by coupling the first pump light and the submarine cable transmission signal to the first-stage amplification component.

[0047] The connections between the devices involved in this embodiment are optical path connections achieved through single-mode optical fiber.

[0048] The first-stage amplification component 130 employs a first Raman gain medium to induce stimulated Raman scattering of the first pump light within this medium, thereby amplifying the submarine cable transmission signal to obtain a first signal. In this embodiment, the submarine cable transmission signal comprises optical signals in the S-band (1460 nm-1530 nm), C-band (1530 nm-1565 nm), and L-band (1565 nm-1625 nm).

[0049] For example, Raman gain fiber can be used as a first-stage amplification component. This Raman gain fiber can be a photonic crystal fiber (PCF) with high nonlinearity, or another type of highly nonlinear fiber (HNLF).

[0050] For example, in practical applications, the first-stage amplification component can employ a pump coupling structure that matches the first Raman gain medium. This pump coupling structure is responsible for coupling the pump light into the Raman gain fiber, allowing the pump light to be injected into the fiber core with low loss and maintaining a high power density within the fiber core to efficiently excite the stimulated Raman scattering effect, thereby providing gain to the signal light.

[0051] Isolation component 140 has its input end connected to the primary amplification component and its output end connected to the secondary amplification component. The isolation component can be composed of an optical isolator. When connected, the positive direction marked on the optical isolator points towards the secondary amplification component to suppress the transmission of various types of light, including pump light, from the secondary amplification component to the primary amplification component.

[0052] Understandably, since the pump light is used to gain signals approximately 100 nm above their center wavelength, the wavelength of the pump light used to gain the L-band (1565 nm-1625 nm) signal falls precisely in the S-band (1460 nm-1530 nm). If the pump light gaining the L-band is not isolated, the pump light gaining the S-band will amplify both the S-band and L-band signals simultaneously, resulting in over-amplification of the L-band signal and insufficient gain in the S-band, thus limiting the transmission distance of the S-band signal. This solution injects a first pump light before the first-stage amplification component and a second pump light after the second-stage amplification component. Because the isolation component between the two amplification components effectively suppresses the reverse propagation of the second pump light, the pump light used to gain the L-band can hardly enter the first-stage amplification component. This effectively prevents the pump light used to gain the L-band from interfering with the S-band signal gain process within the first-stage amplification component, ultimately achieving effective amplification of signals in all bands.

[0053] The secondary amplification component 150 employs a second Raman gain medium to induce stimulated Raman scattering of the second pump light within this medium, thereby amplifying the first signal to obtain the second signal. The second Raman gain medium may be the same as or different from the first Raman gain medium. For example, a Raman gain fiber can be used as the secondary amplification component; this Raman gain fiber may be a photonic crystal fiber with high nonlinearity, or other types of highly nonlinear fibers.

[0054] For example, in practical applications, the secondary amplification component can employ a pump coupling structure that matches the second Raman gain medium. This pump coupling structure is responsible for coupling the pump light into the Raman gain fiber, allowing the pump light to be injected into the fiber core with low loss and maintaining a high power density within the fiber core to efficiently excite the stimulated Raman scattering effect, thereby providing gain to the signal light.

[0055] The backward coupler 160 can be a reflective coupler, including an input terminal, an output terminal, and a bidirectional port. The input terminal is connected to the output terminal of the backward pump module to receive the second pump light output by the backward pump module; the bidirectional port is connected to the end of the secondary amplification component away from the isolation component to inject the second pump light in reverse into the secondary signal amplification unit and receive the second signal amplified by the secondary amplification unit.

[0056] For example, the backward pumping module may include multiple pump light sources. In this case, the backward coupler may include multiple input ports, each of which is used to receive pump light generated by a pump light source. The backward coupler couples each backward pump light into a second pump light and then injects it into the secondary amplification component.

[0057] For example, the backcoupler may include at least one output port, one of which is connected to another segment of submarine cable transmission fiber, allowing the amplified submarine cable transmission signal to continue long-distance transmission through the submarine cable transmission fiber. To further reduce transmission loss, hollow-core fiber can be used for the submarine cable transmission fiber. In practical applications, one or more repeater stations can be set up along the submarine cable transmission link, and signal amplifiers can be used in the repeater stations to amplify the submarine cable transmission signal with low noise to compensate for transmission loss.

[0058] The backward pumping module 170 may include at least one pump source for gaining a C-band signal and at least one pump source for gaining an L-band signal. It is understood that since the pump light is used to gain a signal approximately 100 nm more than its center wavelength, in practical applications, the center wavelength of each pump source in the backward pumping module can be determined based on the wavelengths of the C-band and L-band signals used in the submarine cable transmission.

[0059] For example, if the submarine cable transmission signal uses wavelengths of 1550nm (C-band) and 1610nm (L-band), then the backward pumping module includes at least one pump light source with a center wavelength of 1450nm and at least one pump light source with a center wavelength of 1510nm.

[0060] In the embodiments of this application, "forward" and "backward" refer to the transmission direction of the signal light. That is, transmission in the same direction as the signal light is "forward" and transmission in the opposite direction to the signal light is "backward".

[0061] Alternatively, in another embodiment of this application, in Figure 1 Based on the illustrated embodiments, as Figure 2As shown, the signal amplifier further includes a processor 210, a first converter 220, and a second converter 230. The input terminal of the first converter 220 is connected to the output terminal of the forward coupler 120, and the input terminal of the second converter 230 is connected to the output terminal of the backward coupler 160. The processor is electrically connected to the first converter 220, the second converter 230, the forward pump module 110, and the backward pump module 170, respectively. The first converter 220 converts the output of the forward coupler 120 into a first electrical signal. The second converter 230 converts the output of the backward coupler 160 into a second electrical signal. The processor 210 generates an adjustment command for adjusting the configuration parameters of the target pump light source based on the first and second electrical signals, and sends it to the target pump light source so that the target pump light source adjusts its own configuration parameters according to the received adjustment command. The target pump light source is the pump light source in the forward pump module 110 and / or the backward pump module 170.

[0062] Understandably, since the optical signal output from the forward coupler includes the initial submarine cable transmission signal to be amplified, the first converter can convert the unamplified submarine cable transmission signal into a first electrical signal. Similarly, since the optical signal output from the backward coupler includes the amplified submarine cable transmission signal, the second converter can convert the amplified submarine cable transmission signal into a second electrical signal. Because the first and second electrical signals correspond to the submarine cable transmission signals before and after amplification, respectively, the processor can determine whether the current amplification performance of the signal amplifier meets the requirements based on the first and second electrical signals, and accordingly generate an adjustment command to adjust the configuration parameters of the target pump light source. Upon receiving the adjustment command, the target pump light source adjusts its own configuration parameters. This enables real-time closed-loop control based on signal feedback, thereby further improving the amplification performance and long-term stability of the signal amplifier.

[0063] The processor 210 can be a processor of type MCU (Microcontroller Unit) or ASIC (Application-Specific Integrated Circuit). The processor is electrically connected to the first converter, the second converter, the forward pump module, and the backward pump module, respectively. It is used to receive the first electrical signal output by the first converter and the second electrical signal output by the second converter, and generate adjustment instructions based on the first and second electrical signals to adjust the configuration parameters of the pump light source in the first pump module and / or the second pump module.

[0064] The above electrical connections can be achieved by wire soldering or PCB (Printed Circuit Board) traces.

[0065] The first converter 220 can be a spectrometer or other conversion device capable of converting optical signals of different wavelengths into electrical signals. It is understood that, because the first converter can convert optical signals of different wavelengths into electrical signals, even if the output of the forward coupler contains the first pump light, the first converter can separate the spectra corresponding to the first pump light and the signal light based on their different wavelengths, thereby obtaining independent electrical signals corresponding to the pump light and the signal light respectively. That is, the first electrical signal contains independent electrical signals corresponding to the pump light and the signal light respectively.

[0066] The second converter 230 can be a spectrometer or other conversion device capable of converting optical signals of different wavelengths into electrical signals. For example, in one implementation, the first and second converters use the same type of converter to reduce measurement deviations introduced by differences in photoelectric responsivity between different converters, thereby improving the accuracy of the processor's gain calculation for the signal amplifier.

[0067] It is understandable that, since the second converter can convert light signals of different wavelengths into electrical signals respectively, even if the output of the back coupler contains the second pump light, the second converter can still separate the spectra corresponding to the second pump light and the signal light according to the different wavelengths of the second pump light and the signal light, thereby obtaining independent electrical signals corresponding to the pump light and the signal light respectively. That is, the second electrical signal also contains independent electrical signals corresponding to the pump light and the signal light respectively.

[0068] For example, the current gain value of the signal amplifier can be determined based on the power difference between the electrical signal corresponding to the signal light in the first electrical signal and the electrical signal corresponding to the signal light in the second electrical signal, and the current gain value can be compared with a preset target gain value. If the current gain value is lower than the target gain value, an adjustment command is generated to increase the pump power of the target pump light source; if the current gain value is higher than the target gain value, an adjustment command is generated to decrease the pump power driven by the target pump light source. The target pump light source can be a pump light source in a forward pump module and / or a backward pump module.

[0069] For example, the configuration parameters can be the driving current or driving voltage of the pump light source, and the pump power of the pump light source can be adjusted by adjusting the driving current or driving voltage.

[0070] In one implementation, the configuration parameter is the drive current. It is understood that, since pump power is linearly related to current, controlling the pump power of the pump source by adjusting the drive current offers advantages such as fast adjustment response, precise power control, and ease of achieving automatic power stabilization through feedback control.

[0071] This application does not limit the specific implementation of the processor adjusting the configuration parameters of the target pump light source in the embodiments. For the sake of clarity of the scheme layout, a specific implementation of the processor adjusting the configuration parameters of the target pump light source is introduced in the following embodiments.

[0072] Alternatively, in another embodiment of this application, such as Figure 3 As shown, the processor 210 is specifically used to execute the following steps: S301, determining the current amplification performance index of the signal amplifier based on the first electrical signal and the second electrical signal; S302, in each iteration, determining whether there is a target pump light source among the pump light sources that needs to be adjusted in terms of configuration parameters. If so, sending an adjustment instruction carrying the adjustment step size and adjustment direction to the target pump light source so that the target pump light source adjusts its own configuration parameters according to the adjustment step size and adjustment direction indicated by the received adjustment instruction; wherein, the adjustment step size decreases as the iteration increases, and the adjustment direction in each iteration is to make the predicted amplification performance index corresponding to the adjusted combination of configuration parameters of each pump light source closer to the target amplification performance index than the current amplification performance index; S303, if the preset iteration stop condition is not met, returning to the step of determining the current amplification performance index of the signal amplifier based on the first electrical signal and the second electrical signal until the iteration stop condition is met and the adjustment is stopped.

[0073] Understandably, in each iteration, if a target pump source is identified that requires parameter adjustment, the configuration parameters of the target pump source are adjusted by a step size in the direction that makes the adjusted predicted amplification performance index closer to the target amplification performance index. This adjustment step size decreases as the iteration number increases. In this way, through iterative adjustment with decreasing step sizes, the gap with the target amplification performance index can be quickly narrowed in the early stages of the iteration, and then finely adjusted with small step sizes in the later stages to smoothly approach the target amplification performance index. This ensures both adjustment speed and rapid and stable convergence of the amplification performance index.

[0074] Regarding step S301, the current amplification performance index of the signal amplifier can be determined based on the power difference between the electrical signal corresponding to the signal light in the first electrical signal and the electrical signal corresponding to the signal light in the second electrical signal. For example, the amplification performance index may include gain spectrum, noise figure, gain flatness, and link transmission performance parameters (such as optical signal-to-noise ratio and wavelength drift).

[0075] In one implementation, gain spectrum, noise figure, and gain flatness are used as amplification performance indicators. It is understood that the gain spectrum reflects the amplification capability of the signal amplifier at different wavelengths; the effective operating bandwidth of the signal amplifier can be determined by the continuous wavelength range in the gain spectrum that meets the gain requirements. The noise figure characterizes the degree of signal-to-noise ratio reduction after the signal passes through the signal amplifier; the lower the value, the less additional noise introduced by the signal amplifier, and the better the signal is preserved. Gain flatness reflects the consistency of gain for signals of different wavelengths within the operating band; the smaller the value, the smaller the gain difference between wavelengths. Using these three amplification performance indicators, the overall performance of the signal amplifier can be comprehensively reflected from three dimensions: gain capability, signal quality, and gain flatness.

[0076] In this implementation, the power difference between the electrical signals corresponding to the same wavelength in the first and second electrical signals can be calculated to obtain the gain value corresponding to each wavelength, thereby obtaining the current gain spectrum of the signal amplifier (i.e., the distribution curve of the gain value at each wavelength position of the signal light). Based on the gain spectrum, the maximum deviation of the gain value at each wavelength (i.e., the difference between the maximum and minimum gain value) can be calculated to obtain the gain flatness. By calculating the ratio of the peak power of the signal at the same wavelength in the first and second electrical signals to the power of the adjacent noise floor, the input signal-to-noise ratio (SNR) and output SNR are obtained. The ratio of the input SNR to the output SNR is then calculated to obtain the noise figure at that wavelength.

[0077] For step S302, in each iteration, it can be determined whether the configuration parameters of each pump light source need to be adjusted, and the pump light source that needs to be adjusted is determined as the target pump light source.

[0078] For example, each pump light source can be traversed, and for each pump light source traversed, the adjustment step size corresponding to the current iteration round for increasing or decreasing the configuration parameters of that pump light source can be predicted, while the pump parameters of other pump light sources remain unchanged, and the predictive amplification performance index of the signal amplifier can be obtained.

[0079] If the predicted amplification performance index obtained after adjustment is closer to the target amplification performance index than the current amplification performance index, then the adjustment direction is determined to be adjustment. At this time, the pump light source is the target pump light source, and the processor sends an adjustment instruction to the pump light source carrying the adjustment direction of adjustment and the adjustment step size corresponding to the current iteration round.

[0080] If the predicted amplification performance index obtained after adjustment is closer to the target amplification performance index than the current amplification performance index, then the adjustment direction is determined to be adjustment. At this time, the pump light source is the target pump light source, and the processor sends an adjustment instruction to the pump light source with the adjustment direction being adjustment and the adjustment step size being the adjustment step size corresponding to the current iteration round.

[0081] If the predicted amplification performance index obtained after increasing or decreasing the amplification performance index is not as close to the target amplification performance index as the current amplification performance index, then the pump source does not need to be adjusted, and the next pump source is traversed.

[0082] After receiving the adjustment command, the target pump light source adjusts its configuration parameters according to the adjustment step size and adjustment direction indicated by the received adjustment command.

[0083] For example, an initial adjustment step size and a step size decay coefficient can be preset. The first iteration uses the initial adjustment step size, and the adjustment step size gradually decreases as the iteration increases. The adjustment step size corresponding to the subsequent iteration satisfies the following relationship with the adjustment step size corresponding to the previous iteration:

[0084] ;

[0085] in, This is the adjustment step size corresponding to the (m+1)th iteration. This represents the adjustment step size corresponding to the m-th iteration. This is the step size decay coefficient.

[0086] For example, the initial adjustment step size can be set to allow the pump power to change by 0.2dB or 0.3dB, with a step size decay coefficient of 0.95 or 0.9. For example, if the initial adjustment step size is set to allow the pump power to change by 0.2dB, and the configuration parameter uses the drive current, then the change in drive current corresponding to the 0.2dB change in pump power can be calculated as the initial adjustment step size.

[0087] For example, in practical applications, the target amplification performance metrics can be set according to requirements. When comparing whether the predicted amplification performance metrics are closer to the target amplification performance metrics than the current amplification performance metrics, the metrics can be compared separately for each dimension. Then, the results of the comparison of each dimension are weighted and summed to obtain the final comparison result. The metrics for each dimension include gain spectrum, noise figure, and gain flatness, etc., and the weights of each dimension can be set by relevant technical personnel based on experience and requirements.

[0088] The predicted amplification performance indicators corresponding to the configuration parameter combinations of each pump light source can be determined by simulating the actual amplification process of each pump light source under the given configuration parameter combinations using a pre-built Raman amplification performance prediction model. The predicted amplification performance indicators are then determined based on the amplified signal obtained from the simulation. For example, this pre-built Raman amplification performance prediction model can be a simulation model based on the signal amplifier and submarine cable transmission characteristics, used to quickly calculate the predicted gain spectrum and noise figure of the signal after Raman amplification under a given configuration parameter combination.

[0089] For step S303, the preset iteration stopping condition can be reaching a preset number of iterations or reaching a preset minimum adjustment step size. For example, the preset number of iterations can be set to 100 or 200 times, and the preset minimum adjustment step size can be set to 1e-4 dB or 1e-3 dB, etc.

[0090] If the preset iteration stop condition is not met, return to step S301; otherwise, stop the iteration.

[0091] Alternatively, in another embodiment of this application, such as Figure 4 As shown, the initial configuration parameters of each pump light source in the forward pumping module and the backward pumping module are obtained according to the following steps: S401, the processor inputs the target amplification performance index into the pre-trained inverse network to obtain the configuration parameter combination of each pump light source output by the inverse network; wherein, the inverse network is trained based on the training dataset under the architecture of the forward-inverse cascaded network, and the training dataset includes each sample configuration parameter combination and its corresponding actual amplification performance index. The actual amplification performance index corresponding to each sample configuration parameter combination represents the amplification performance that the signal amplifier can achieve when configuring each pump light source using the sample configuration parameter combination; S402, the initial configuration parameters of each pump light source are determined according to the obtained configuration parameter combination.

[0092] Understandably, by using a training dataset that includes each sample's configuration parameter combination and its corresponding actual amplification performance index, the inverse network under the forward-inverse cascade architecture can be trained. This allows the inverse network to learn the forward physical mapping from configuration parameter combinations to amplification performance indexes and then establish a reverse mapping relationship from amplification performance indexes to configuration parameter combinations. When multiple configuration parameter combinations correspond to the same or similar amplification performance, the forward network in the cascade architecture can perform forward verification on the configuration parameter combinations output by the inverse network. This enables the inverse network to converge to a single solution when facing a many-to-one mapping problem, without having to exhaustively enumerate all possible configurations. This effectively simplifies the process of determining the initial configuration parameters and provides a feasible starting point for subsequent fine-tuning.

[0093] Regarding step S401, configuration parameters for each pump light source can be randomly generated within the range of their respective values ​​to obtain sample configuration parameter combinations. The actual amplification performance index corresponding to each sample configuration parameter combination can be obtained through testing with the signal amplifier provided in this embodiment, or by simulating the actual amplification process of each pump light source under each sample configuration parameter combination using the aforementioned Raman amplification performance prediction model. The amplification performance index determined based on the amplified signal obtained from the simulation can then be used as the actual amplification performance index corresponding to each sample configuration parameter combination.

[0094] The inverse network and the forward network are cascaded to form a forward-inverse cascaded network. The inverse network is trained within this cascaded network architecture. That is, the training of the inverse network is not independent, but is based on the forward physical mapping established by the forward network, and is trained in conjunction with the forward network within the forward-inverse cascaded architecture.

[0095] Since the inverse network is trained on a training dataset containing sample configuration parameter combinations and their corresponding actual amplification performance metrics within a forward-inverse cascade architecture, the training data already covers the correspondence between configuration and performance. Therefore, the inverse network can learn the reverse mapping of this correspondence. Simultaneously, the forward network in the cascade structure provides physical constraints to the inverse network during training, ensuring that the output of the inverse network remains consistent with the physical laws reflected in the training data, rather than being a mathematical pseudo-solution deviating from reality. Based on this training mechanism, the trained inverse network can directly map the matching configuration parameter combinations to the input target amplification performance metrics.

[0096] To ensure clarity of the solution, the specific training process of the inverse network will be described in the following embodiments, and will not be repeated here.

[0097] For step S402, each configuration parameter in the configuration parameter combination obtained in step S401 can be determined as the initial configuration parameter of the pump light source corresponding to that parameter, based on the correspondence between each configuration parameter in the configuration parameter combination and each pump light source.

[0098] Alternatively, in another embodiment of this application, such as Figure 5As shown, the inverse network is trained through the following steps: S501, based on the configuration parameter combinations of each sample and their corresponding actual amplification performance indicators, the forward network in the forward-inverse cascaded network is trained so that the forward network fits the first mapping relationship between the configuration parameter combinations and the amplification performance indicators; S502, each actual amplification performance indicator is input into the inverse network in the forward-inverse cascaded network so that the inverse network outputs the predicted configuration parameter combinations corresponding to each actual amplification performance indicator based on the inverse mapping relationship of the first mapping relationship; S503, the Raman amplification performance corresponding to each predicted configuration parameter combination is predicted to obtain each predicted amplification performance indicator; S504, based on the first difference between each predicted amplification performance indicator and each actual amplification performance indicator, and the second difference between each predicted configuration parameter combination and each sample configuration parameter combination, the loss value is determined; S505, based on the loss value, the network parameters of the inverse network are adjusted to obtain the trained inverse network.

[0099] It is understandable that by inputting the combination of configuration parameters of each sample and its corresponding actual amplification performance index into the forward network to establish the first mapping relationship, and on this basis driving the inverse network to learn the inverse mapping of the first mapping relationship, the first difference and the second difference are introduced as dual constraints during the training process to ensure that the predicted performance is close to the actual performance and the predicted configuration does not deviate from the true configuration distribution. Thus, the trained inverse network can directly output a reasonable combination of configuration parameters according to the target amplification performance, thereby providing a reliable initial configuration for subsequent iterative adjustments.

[0100] For step S501, each sample configuration parameter combination can be used as input data, and the actual amplification performance index corresponding to each sample configuration parameter combination can be used as supervision label to supervise the training of the forward network, so that the forward network fits the mapping relationship between the configuration parameter combination and the amplification performance index, and obtains the first mapping relationship.

[0101] For example, the initial forward network can be an MLP (Multilayer Perceptron) or a residual network, and the network parameters can be initialized randomly.

[0102] For example, the configuration parameters of each sample can be used as input data into the initial forward network. The loss is calculated based on the difference between the network output and the corresponding supervision label, and the network parameters (including weights and biases) are updated by backpropagation to minimize the loss until the forward network converges, resulting in a trained forward network. Through training, the weights and biases within the forward network are adjusted to a certain numerical combination, enabling the network to achieve a deterministic mapping from input to output, resulting in a forward mapping function between input and output, which is the first mapping relationship.

[0103] For step S502, the inverse network can be of type MLP or residual network, and the initial network parameters of the inverse network can be obtained based on the inverse mapping relationship of the first mapping relationship. Specifically, the inverse network can be pre-initialized, that is, the actual amplification performance index corresponding to each sample configuration parameter combination is input into the inverse network, and the parameters of the inverse network are adjusted according to the difference between the output result of the inverse network and the corresponding sample configuration parameter combination. When a predetermined number of iterations is reached or the difference between the output result of the inverse network and the corresponding sample configuration parameter combination is less than a preset threshold, the current parameters of the inverse network are used as the initial network parameters to obtain the pre-initialized inverse network. The inverse network obtained after pre-initialization can fit the inverse mapping relationship of the first mapping relationship, so that after the actual amplification performance index is input into the inverse network, the inverse network can output the predicted configuration parameter combination based on the inverse mapping relationship of the first mapping relationship.

[0104] For example, the input dimension of the inverse network can be the dimension of the amplification performance index, and the output dimension can be the number of pump light sources. Each actual amplification performance index is input into the inverse network to obtain a set of parameters output by the inverse network. This set of parameters is the combination of prediction configuration parameters corresponding to the actual amplification performance index.

[0105] For step S503, the above-mentioned Raman amplification performance prediction model can be used to simulate the Raman amplification process, thereby predicting the Raman amplification performance corresponding to each combination of prediction configuration parameters and obtaining each predicted amplification performance index.

[0106] For example, the amplification performance index corresponding to each combination of prediction configuration parameters can be determined as the predicted amplification performance index based on the power difference between the amplified signal and the signal before amplification obtained from the simulation. The specific implementation method for determining the amplification performance index based on the power difference between the amplified signal and the signal before amplification can be referred to the relevant description in the above embodiments, and will not be repeated here.

[0107] For step S504, at least one amplification performance metric can be used, including gain spectrum, noise figure, and gain flatness. If multiple amplification performance metrics are used, the difference between the predicted and actual amplification performance metrics for each sample configuration parameter combination in the same dimension can be calculated separately. Then, the differences in each dimension are weighted and summed to obtain the first difference. For example, the weights corresponding to the differences in each dimension can be set by relevant technical personnel based on experience and needs. For instance, the weights of each dimension can be the same, or the weights of dimensions that are more important in the current scenario can be set higher.

[0108] For example, the difference between each sample configuration parameter combination and its corresponding predicted configuration parameter combination can be calculated separately, and the sum of the obtained differences is determined as the second difference. The above differences can be calculated using the L1 distance function or the L2 distance function, and the specific function used to calculate the differences is not limited in the embodiments of this application.

[0109] For example, after obtaining the first difference and the second difference, the first difference and the second difference can be weighted and summed to obtain the loss value. The weights corresponding to the first difference and the second difference can be set by relevant technical personnel based on experience; for example, the weights corresponding to the first difference and the second difference can be the same.

[0110] For example, in another implementation, if the amplification performance metrics include at least gain spectrum, noise figure, and gain flatness, the combined difference between the gain spectrum and noise figure can be used as a third difference. The first, second, and third differences are then weighted and summed to obtain the loss value. It is understood that since gain spectrum and noise figure are two more important dimensions of the signal amplification process, obtaining the third difference based solely on these two dimensions, and using the loss value including the third difference to train the inverse network, allows the combined difference between the gain spectrum and noise figure to be included as an independent constraint term in the loss. This ensures that the accuracy of these two dimensions is prioritized during the inverse network training process, enabling the trained inverse network to output at least a combination of configuration parameters that meets the gain and noise performance requirements.

[0111] Regarding step S505, the network parameters of the inverse network can be adjusted through backpropagation that minimizes the loss value or other optimization algorithms until a preset stopping condition is met, resulting in a trained inverse network. For example, the stopping condition could be reaching a preset number of iterations, or the loss value being less than a preset threshold.

[0112] The training process of the aforementioned reverse network can be carried out in the processor within the signal amplifier, or it can be carried out in an electronic device with data processing capabilities outside the signal amplifier; both are reasonable.

[0113] Alternatively, in another embodiment of this application, in Figure 1 Based on the illustrated embodiments, as Figure 6As shown, the forward pumping module 110 includes a pump combiner 1103, multiple forward optical isolators 1102, and multiple forward pumping light sources 1101. The number of forward optical isolators 1102 is the same as the number of forward pumping light sources 1101. Each forward pumping light source 1101 corresponds to a forward optical isolator 1102. The output terminal of each forward pumping light source 1101 is connected to the input terminal of its corresponding forward optical isolator 1102. The output terminal of each forward optical isolator 1102 is connected to the pump combiner. The input terminal of the pump combiner 1103 is connected to the input terminal of the forward coupler 120; each forward pump light source 1101 generates pump light to provide gain for the S-band signal; each forward optical isolator 1102 suppresses the reverse transmission light of the pump light generated by the corresponding forward pump light source 1101; the pump combiner 1103 couples the multi-channel isolated pump light obtained after the pump light generated by each forward pump light source 1101 passes through the forward optical isolator 1102 into the first pump light.

[0114] Understandably, by employing multiple forward pump sources, multiple wavelengths of pump light can be provided, thereby effectively broadening the gain bandwidth. Furthermore, each pump source is connected to its corresponding optical isolator, which can suppress the pump light from reflecting back to the source, thereby avoiding the instability of the pump source's operating state and potential damage risk caused by reflected light. After passing through the optical isolator, each pump light is combined by a beam combiner, which can efficiently couple multiple pump lights to obtain broadband pump light.

[0115] The forward pump module can include multiple pump light sources of different wavelengths, each wavelength being used to gain S-band signals, and the same wavelength can also correspond to multiple pump light sources.

[0116] For example, the forward pumping module may include at least one pump light source with a wavelength of 1365 nm and at least one pump light source with a wavelength of 1385 nm to generate pump light that amplifies S-band signals of different wavelengths.

[0117] The output of each forward pump light source is connected to the input of its corresponding forward optical isolator. The output of the optical isolator is connected to the input of the pump combiner. The forward optical isolator can suppress the reverse transmission of pump light generated by the forward pump light source to ensure unidirectional transmission of pump light. This protects the pump light source and allows the pump light to enter the subsequent optical path with the most efficient efficiency, i.e., minimizes pump power loss.

[0118] For example, the specific type of pump combiner can be selected according to the number of forward pump light sources. For instance, if there are 10 forward pump light sources, a pump combiner for combining 10 beams into one beam can be used. This pump combiner has 10 input ports and 1 output port.

[0119] The pump light generated by each forward pump light source is passed through a forward optical isolator to obtain multiple isolated pump lights, which are then input to the respective input ports of the pump combiner. After coupling within the pump combiner, the first pump light is output through the output port of the pump combiner.

[0120] Alternatively, in another embodiment of this application, in Figure 1 Based on the illustrated embodiments, as Figure 7 As shown, the backward pumping module 170 includes multiple backward optical isolators 1702 and multiple backward pumping light sources 1701. The number of backward optical isolators 1702 is the same as the number of backward pumping light sources 1701. Each backward pumping light source 1701 corresponds to a backward optical isolator 1702. The output terminal of each backward pumping light source 1701 is connected to the input terminal of its corresponding backward optical isolator 1702. The output terminal of each backward optical isolator 1702 is connected to the input terminal of the backward coupler 160. Each backward pumping light source 1701 generates pump light that provides gain for C-band or L-band signals. Each backward optical isolator 1702 suppresses the reverse transmission of the pump light generated by its corresponding backward pumping light source 1701. The backward coupler 160 couples the multi-channel isolated pump light obtained after the pump light generated by each backward pumping light source 1701 passes through the backward optical isolator into a second pump light.

[0121] Understandably, by employing multiple back-pumped light sources, multiple wavelengths of pump light can be provided, thereby effectively broadening the gain bandwidth. Furthermore, each pump light source is connected to its corresponding optical isolator, which can suppress the reflection of pump light back to the light source, thereby avoiding the instability of the pump light source's operating state and potential damage risk caused by reflected light. After passing through the optical isolator, each back-pumped light is coupled through a back-coupler, which can efficiently couple multiple pump lights to obtain broadband pump light.

[0122] The backward pumping module can include multiple pump sources of different wavelengths. Each wavelength includes the wavelength used to gain the signals in the C-band and L-band. Furthermore, the same wavelength can also correspond to multiple pump sources.

[0123] For example, the backward pumping module may include at least one pump light source with a wavelength of 1450 nm and at least one pump light source with a wavelength of 1510 nm to generate pump light that amplifies signals in the C-band and L-band.

[0124] The output of each backward pump light source is connected to the input of its corresponding backward optical isolator. The output of the backward optical isolator is connected to the input of the backward coupler. The backward optical isolator can suppress the reverse transmission of pump light generated by the backward pump light source to ensure unidirectional transmission of pump light. This protects the pump light source and allows the pump light to enter the subsequent optical path with the most efficient power, i.e., minimizes pump power loss.

[0125] For example, the specific type of back coupler can be selected according to the number of back pump light sources. For instance, if there are 10 back pump light sources, a back coupler for combining 10 beams into one beam can be used. In addition to a bidirectional port and at least one output port, the back coupler may also include 10 input ports.

[0126] The pump light generated by each backward pump light source is passed through a backward optical isolator to obtain multiple isolated pump lights, which are then input to the respective input ports of the backward coupler. After coupling within the backward coupler, they are injected into the secondary amplification component through a bidirectional port.

[0127] Corresponding to the aforementioned signal amplifier, this application embodiment also provides a signal amplification method. The method uses a signal amplifier to amplify the signal. The signal amplifier includes: a forward pump module, a backward pump module, a forward coupler, a backward coupler, an isolation component, a first-stage amplification component, and a second-stage amplification component. The two input terminals of the forward coupler are respectively connected to the forward pump module and the submarine cable transmission optical fiber; the output terminal of the forward coupler is connected to one end of the first-stage amplification component, and the other end of the first-stage amplification component is connected to the input terminal of the isolation component; the output terminal of the isolation component is connected to one end of the second-stage amplification component, and the other end of the second-stage amplification component is connected to the bidirectional port of the backward coupler; the input terminal of the backward coupler is connected to the backward pump module. Figure 8 As shown, the method includes: S801, the forward pump module outputs a first pump light that provides gain for an S-band signal; S802, a forward coupler couples the first pump light and the submarine cable transmission signal transmitted by the submarine cable transmission optical fiber and outputs the signal; the submarine cable transmission signal includes optical signals in S-band, C-band and L-band; S803, a first-stage amplification component amplifies the submarine cable transmission signal under the action of the first pump light to obtain a first signal; S804, the backward pump module outputs a second pump light that provides gain for C-band and L-band signals, an isolation component suppresses the transmission of pump light from a second-stage amplification component to a first-stage amplification component, and the second-stage amplification component amplifies the first signal under the action of the second pump light to obtain a second signal.

[0128] The signal amplification method provided in this application involves coupling a submarine cable transmission signal transmitted via an optical fiber to a first pump light generated by a forward pump module via a forward coupler, and then outputting the signal to a first-stage amplification component. The first-stage amplification component amplifies the S-band signal in the submarine cable transmission signal under the action of the first pump light. The amplified submarine cable transmission signal is then transmitted to a second-stage amplification component via an isolation component. Furthermore, a second pump light generated by a backward pump module, which provides gain for the C-band and L-band signals, is coupled through a backward coupler and injected into the second-stage amplification component from the other end. The second-stage amplification component amplifies the C-band and L-band signals in the submarine cable transmission signal under the action of the second pump light.

[0129] The aforementioned signal amplification method enables signal amplification for broadband submarine cable transmission signals. Furthermore, this scheme injects a first pump light before the first-stage amplification component and a second pump light after the second-stage amplification component. Because the isolation component between the two amplification components effectively suppresses the reverse propagation of the second pump light, the pump light used for L-band gain is almost unable to enter the first-stage amplification component. This effectively prevents interference from the L-band pump light on the S-band signal gain process within the first-stage amplification component. Thus, compared to existing Raman amplifiers where the energy of the pump light used for S-band gain is transferred to the pump light used for L-band gain, resulting in over-amplification of the L-band signal and insufficient gain in the S-band, this scheme, by preventing the L-band pump light from entering the first-stage amplification component, allows for focused amplification of the S-band signal within the first-stage amplification component. Therefore, the two-stage cascaded amplification components achieve flat and efficient gain for S-band, C-band, and L-band signals. The signal amplification method provided in this solution can effectively amplify broadband submarine cable transmission signals in submarine optical cable systems, thereby improving the transmission capacity of submarine optical cable systems.

[0130] Optionally, the signal amplifier further includes a processor, a first converter, and a second converter; the input terminal of the first converter is connected to the output terminal of the forward coupler, and the input terminal of the second converter is connected to the output terminal of the backward coupler; the processor is electrically connected to the first converter, the second converter, the forward pumping module, and the backward pumping module, respectively; the method further includes: the first converter converting the output of the forward coupler into a first electrical signal; the second converter converting the output of the backward coupler into a second electrical signal; the processor generating an adjustment command for adjusting the configuration parameters of the target pumping light source based on the first electrical signal and the second electrical signal, and sending it to the target pumping light source so that the target pumping light source adjusts its own configuration parameters according to the received adjustment command; the target pumping light source is a pumping light source in the forward pumping module and / or the backward pumping module.

[0131] Optionally, the method further includes: the processor determining the current amplification performance index of the signal amplifier based on the first electrical signal and the second electrical signal; in each iteration, determining whether there is a target pump light source among the pump light sources that needs configuration parameter adjustment; if so, sending an adjustment instruction carrying an adjustment step size and adjustment direction to the target pump light source, so that the target pump light source adjusts its own configuration parameters according to the adjustment step size and adjustment direction indicated by the received adjustment instruction; wherein, the adjustment step size decreases as the iteration increases, and the adjustment direction in each iteration is such that the predicted amplification performance index corresponding to the adjusted combination of configuration parameters of each pump light source is closer to the target amplification performance index than the current amplification performance index; if a preset iteration stop condition is not met, returning to the step of determining the current amplification performance index of the signal amplifier based on the first electrical signal and the second electrical signal, until the iteration stop condition is met and the adjustment stops.

[0132] The implementation of each step in the signal amplification method provided in this application embodiment can be referred to the relevant description of the above-described signal amplifier embodiment, and will not be repeated here.

[0133] In another embodiment of this application, if the submarine cable transmission signal uses C-band + L-band, i.e., includes optical signals of both C-band and L-band, then the forward pump module in the aforementioned signal amplifier can generate pump light to provide gain for the C-band signal, and the backward pump module can generate pump light to provide gain for the L-band signal. For example, if the submarine cable transmission signal uses wavelengths of 1550nm (C-band) and 1610nm (L-band), then the forward pump module includes at least one pump light source with a center wavelength of 1450nm, and the backward pump module includes at least one pump light source with a center wavelength of 1510nm.

[0134] Understandably, traditional Raman amplifiers suffer from complex power transfer issues, most notably in S-band + C-band + L-band amplification. However, power transfer also occurs in C-band + L-band amplification, leading to gain unevenness. The signal amplifier provided in this embodiment effectively suppresses the back propagation of backward-injected pump light by using an isolation component between the two amplification components. This prevents pump light used for L-band gain from entering the first-stage amplification component, thus effectively avoiding interference from L-band gain pump light on the C-band signal gain process. This allows the first-stage amplification component to focus on C-band signal amplification, while the second-stage amplification component focuses on L-band signal amplification. Therefore, the two cascaded amplification stages provide flat and efficient gain for both C-band and L-band signals.

[0135] To better understand the signal amplifier provided in the embodiments of this application, a specific example will be used for description below.

[0136] like Figure 9 As shown, the signal amplifier provided in this example adopts a bidirectional pumped two-stage cascaded discrete Raman amplification structure, including a forward pump module (corresponding to the forward pump module and forward coupler mentioned above), a backward pump module (corresponding to the backward pump module and backward coupler mentioned above), a first-stage signal amplification unit (corresponding to the first-stage amplification component mentioned above), a second-stage signal amplification unit (corresponding to the second-stage amplification component mentioned above), and a performance dynamic monitoring feedback unit (corresponding to the processor mentioned above).

[0137] The forward pump module includes multiple forward pump sources with different center wavelengths, optical isolators connected to each forward pump source, a multi-wavelength pump combiner, and a pump-signal coupler (corresponding to the forward coupler mentioned above). For example... Figure 9 As shown, the forward pumping module is equipped with forward pumping light sources with center wavelengths of 1405nm and 1425nm, respectively.

[0138] The pump light output from each forward pump source first passes through a corresponding optical isolator to suppress the reverse optical feedback at the pump source end face and improve the pump output stability. Subsequently, pump light of different wavelengths is combined by a multi-wavelength pump combiner to form broadband Raman pump light, which is then coupled with the broadband submarine cable transmission signal to be amplified through a pump-signal coupler before being input into a first-stage signal amplification unit.

[0139] The primary signal amplification unit includes a first Raman gain medium and a matching pump coupling structure. The first Raman gain medium uses photonic crystal fiber with high nonlinearity as the Raman gain medium, used for primary gain compensation of the input broadband submarine cable signal. By rationally configuring the pump wavelength and power ratio in the primary signal amplification unit, it primarily achieves gain establishment in the short-wavelength band, while reducing energy competition and gain imbalance between different pumps during broadband amplification. The output of the primary signal amplification unit is connected to a selective band isolation component (corresponding to the isolation component mentioned above). The selective band isolation component is composed of optical isolation devices with selective optical transmission characteristics, used to manage pump light between the primary and secondary amplification units and suppress backscattered signals. The selective band isolation component allows the target communication signal after primary amplification to pass through with low loss, while suppressing light transmitted from the secondary signal amplification unit to the primary amplification unit, including suppressing residual pump light, backscattered light, and non-target light signals generated by the secondary signal amplification unit from returning to the primary signal amplification unit, thereby reducing the impact of interstage pump leakage and optical feedback on system gain flatness and noise performance.

[0140] The secondary signal amplification unit includes a second Raman gain medium and a matching pump coupling structure, used for secondary gain compensation of the broadband submarine cable signal after primary amplification. The signal light after passing through the selective band isolation component enters the second Raman gain medium, while the multi-wavelength pump light output from the backward pump module is injected back into the second Raman gain medium through the backward pump coupler, realizing broadband high-gain submarine cable signal amplification.

[0141] By coordinating the primary and secondary signal amplification units, signals of different communication bands can obtain matched gain compensation, thereby improving the wideband output gain flatness.

[0142] The back-pumping module includes multiple back-pumping light sources, corresponding optical isolators, and a back-pumping coupler (corresponding to the back-coupler mentioned above). For example... Figure 9 As shown, the back-pumping module is equipped with back-pumping light sources with center wavelengths of 1470nm, 1490nm and 1510nm respectively.

[0143] Pump light from multiple backward-pumped light sources is directionally controlled by an optical isolator and then injected into a secondary signal amplification unit through the bidirectional ends of a backward-pump coupler, providing reverse pump energy for subsequent signal amplification. Through the synergistic effect of the forward and backward pump modules, broadband Raman gain coverage for signals in different frequency bands is achieved.

[0144] The performance dynamic monitoring feedback unit monitors the input of the first-stage signal amplification unit and the output of the second-stage signal amplification unit. Specifically, the performance dynamic monitoring feedback unit is connected to the output of the backward pump coupler via a spectrometer to acquire the output of the second-stage signal amplification unit. Similarly, it is connected to the output of the pump-signal coupler via a spectrometer to acquire the input of the first-stage signal amplification unit (i.e., monitor the input broadband submarine cable signal). This allows for real-time acquisition of the operating status of the two-stage amplification components and the characteristic parameters of the submarine cable transmission link, including performance indicators such as input and output optical power, gain spectrum of different bands, noise figure, and gain flatness.

[0145] The performance dynamic monitoring and feedback unit integrates a physical information reverse design optimization network and a genetic algorithm fine-tuning model. The physical information reverse design optimization network extracts the nonlinear correlation between submarine cable link parameters and amplification performance indicators through a feature extraction module. It also introduces latent variables to constrain the complex mapping relationship between pump configuration parameters and target gain and noise performance, transforming the pump parameter optimization problem with multiple possible solutions into a deterministic parameter mapping problem and outputting an initial pump configuration scheme. Subsequently, a genetic algorithm iteratively optimizes the power ratio of different pump light sources, obtaining the optimal pump configuration based on comprehensive evaluation indicators such as gain, noise figure, and gain flatness.

[0146] Specifically, different submarine cable transmission states, pump power configurations, and corresponding broadband amplification performance parameters are collected to construct a training dataset containing link loss characteristics, signal band distribution, pump configuration parameters, and gain-noise performance indicators. This dataset is used to train the physical information in the performance dynamic monitoring feedback unit to design an optimization network.

[0147] The physical information reverse design optimization network adopts a reverse mapping structure, mapping amplification performance indicators to the corresponding pump parameter configuration space. Traditional Raman amplification systems exhibit a many-to-one mapping relationship, i.e.:

[0148] ;

[0149] in, The pump configuration parameters (including power parameters) for the i-th pump light source, where n is the number of pump light sources. The amplification performance metrics include parameters such as gain spectrum, noise figure, and gain flatness. To address the many-to-one mapping problem between the aforementioned pump configuration parameters and amplification performance metrics, the physical information inverse design optimization network introduces latent variables to transform the many-to-one mapping into a one-to-one mapping relationship.

[0150] like Figure 10 As shown, the physical information reverse design optimization network consists of two parts: a forward network and a reverse network. Figure 10Different types of lines correspond to different processing stages, including three stages: forward network, reverse network, and feedback loop. The forward network is used to establish a positive mapping relationship between pump configuration parameters and amplification performance indicators, that is:

[0151] ;

[0152] in, Indicates the first Group pump configuration parameters, including the output power of each pump light source; Indicates the relationship with the first The actual amplification performance indicators corresponding to the pump configuration parameters include output gain spectrum, noise figure, and gain flatness; This represents the nonlinear mapping function of the forward network, used to learn the complex coupling relationship between pump configuration parameters and amplification performance. The latent variables introduced (corresponding to the first mapping relationship in the above text).

[0153] The inverse network is used to solve for the corresponding pump configuration parameters in reverse based on the actual amplification performance indicators, that is:

[0154] ;

[0155] in This represents the inverse mapping function corresponding to the forward network; Indicates the relationship with the first The actual amplification performance indicators corresponding to the pump configuration parameters include performance constraint parameters such as target gain spectrum, noise figure, and gain flatness; Indicates the relationship with the first The pump configuration parameters output by the reverse network corresponding to the group pump configuration parameters (corresponding to the predicted configuration parameter combination mentioned above).

[0156] The pump configuration parameters output from the inverse network are input into the broadband Raman amplification performance prediction model, i.e.:

[0157] ;

[0158] in, This represents a Raman amplification performance prediction model based on the transmission characteristics of submarine cables, used to simulate the actual amplification process corresponding to the pump configuration parameters. This represents the predicted amplification performance index calculated by the Raman amplification performance prediction model based on the output pump configuration parameters of the inverse network. Next, the feedback loop compares the predicted amplification performance index with the actual amplification performance index and adjusts the network parameters of the inverse network. For example, by comparing the predicted amplification performance index... Performance indicators of target amplification The deviation between the parameters is used to optimize the network parameters of the inverse network, so that the pump configuration parameters generated by the trained inverse network can meet the target amplification performance requirements.

[0159] In the training process of the physical information reverse design optimization network, a physical information loss function incorporating submarine cable transmission characteristics is used for optimization. The loss function includes forward prediction error loss, reverse reconstruction loss, and physical constraint loss, and its expression is:

[0160] ;

[0161] ;

[0162] ;

[0163] ;

[0164] in, To predict amplification performance indicators Compared with actual amplification performance indicators The error between them (corresponding to the first difference mentioned above), This is the weight corresponding to the error; This indicates the pump configuration parameters output by the reverse network. Compared with actual pump configuration parameters The reconstruction error between them (corresponding to the second difference mentioned above). This is the weight corresponding to the error; This is the physical constraint loss (corresponding to the third difference mentioned above), used to constrain the network output to meet the physical laws governing the Raman amplification process. This represents the weight corresponding to the loss. and These represent the gain and noise figure calculated based on the pump configuration parameters output by the network, respectively. and These represent the gain and noise figure in the actual amplification performance indicators, respectively.

[0165] When using the signal amplifier provided in this example to amplify a signal, the initial pump configuration parameters output by the reverse design optimization network based on physical information can be iteratively fine-tuned using a greedy optimization algorithm to obtain the optimal pump configuration parameters that meet the target amplification performance requirements.

[0166] Specifically, the initial iteration step size (corresponding to the adjustment step size mentioned above), minimum iteration step size, and maximum number of iterations are set based on the initial pump configuration parameters. During each iteration, the power parameters of each pump source are adjusted sequentially, and the amplification performance index corresponding to different pump configuration parameters is calculated using a Raman amplification performance prediction model. Each candidate pump configuration parameter is evaluated according to a preset comprehensive performance evaluation function. The pump power adjustment direction that can reduce the performance evaluation function value is selected as the current iteration direction, and the power parameters of the corresponding pump source are updated. When all pump configuration parameter adjustments within the current iteration cycle fail to further reduce the performance evaluation function, the iteration step size is reduced, and the next round of search continues until the preset termination condition is met, obtaining the optimal pump configuration parameters.

[0167] Among them, the greedy optimization algorithm uses the pump configuration parameters output by the optimization network designed in reverse with physical information as the initial solution:

[0168] ;

[0169] in, This represents a set of initial pump configuration parameters. Let represent the pump configuration parameters of the i-th pump source, and n represent the number of pump sources. In each iteration, the i-th pump source is independently optimized while keeping the pump configuration parameters of other pump sources unchanged. Candidate configuration parameters are calculated after increasing and decreasing the iteration step size corresponding to the current iteration round, respectively.

[0170] ;

[0171] ;

[0172] in, Indicates the current iteration step size. This represents the candidate configuration parameters obtained after increasing the current iteration step size. This represents the candidate configuration parameters obtained after reducing the current iteration step size.

[0173] Based on the broadband Raman amplification performance under different pump configuration parameters, a comprehensive performance evaluation function is established. :

[0174] ;

[0175] in, This represents the average gain evaluation index. This represents the gain flatness evaluation index. Indicates noise performance evaluation indicators. , and These represent the weight coefficients of the corresponding evaluation indicators. By comparing the comprehensive performance evaluation function values ​​under different power adjustment directions, the optimal adjustment direction for the i-th pump light source is determined:

[0176] ;

[0177] in, This indicates the power update direction of the i-th pump source. This indicates finding the value of the independent variable that minimizes the value within the parentheses.

[0178] Optimize the parameters of each pump source sequentially using the above method to obtain the pump configuration corresponding to the current iteration cycle:

[0179] ;

[0180] in, This represents a set of pump configuration parameters at the start of the (m+1)th iteration. This represents a set of pump configuration parameters at the start of the m-th iteration. This represents the amount of change adjusted during the m-th iteration.

[0181] When adjusting the direction of all pump light sources fails to further reduce the overall performance evaluation function value, the iteration step size is reduced. The next iteration step size is obtained using the following formula:

[0182] ;

[0183] in, Indicates the step size decay coefficient. Indicates the first The iteration step size corresponding to each iteration round. Indicates the first The iteration step size corresponds to each iteration. The above optimization process continues until the current iteration step size is less than the preset minimum step size or the maximum number of iterations is reached. The pump configuration parameters obtained at this point are the final optimization results.

[0184] like Figure 11 The figure shows the gain-noise figure curves obtained by testing the signal amplifier using the above method. The solid line represents the gain curve, and the dashed line represents the noise figure curve. Figure 11 The horizontal axis represents wavelength (in nm), the vertical axis on the left represents gain (in dB), and the vertical axis on the right represents noise figure (in dB).

[0185] from Figure 11As can be seen, with the optimized pump configuration, this solution can achieve high-gain amplification of wideband signals (covering at least 1540-1600nm), while maintaining a low noise figure (maximum noise figure of about 6dB) and good gain flatness (gain around 35dB), meeting the requirements of ultra-long-distance submarine cable transmission systems for high-quality amplification of optical signals.

[0186] To verify the performance advantage of the proposed signal amplifier compared to the traditional Raman amplification scheme, a comparative analysis was conducted between the proposed scheme and the traditional scheme. The results are as follows: Figure 12 As shown. Figure 12 The horizontal axis represents wavelength (in nm), the vertical axis on the left represents gain (in dB), and the vertical axis on the right represents noise figure (in dB).

[0187] In traditional Raman amplification structures, due to the complex energy coupling relationships between multiple pump beams and between the pump beam and the signal beam, gain imbalances can easily occur during the amplification process for signals of different bands, affecting noise performance. This example, however, effectively reduces pump back crosstalk and unwanted power transfer during the two-stage amplification process through a two-stage synergistic cascaded amplification structure and selective band isolation components. This allows signals of different bands to obtain stable gains, thereby improving broadband amplification performance.

[0188] Furthermore, to verify the adaptability of the signal amplifier provided in this example under different target gain requirements, multiple target gain conditions were set, and the corresponding pump configuration parameters were optimized through the performance dynamic monitoring feedback unit. Figure 13 The horizontal axis represents wavelength (in nm), and the vertical axis represents gain (in dB). Different types of curves represent different target gains, including target gains of 15dB, 20dB, 25dB, 30dB, and 35dB.

[0189] like Figure 13 As shown, under different target gain conditions, the signal amplifier provided in this example can maintain continuous and stable wideband gain output. As the target gain increases, the power ratio of different pump sources is dynamically adjusted to achieve a reasonable distribution of pump energy among different communication bands, thereby ensuring that the output gain curve has good flatness.

[0190] In addition, to further verify the application capability of the signal amplifier provided in this example in a practical communication system, a test system based on wavelength division multiplexing coherent transmission was built to test the bit error rate performance of the broadband submarine cable signal amplified by the signal amplifier provided in this example.

[0191] like Figure 14 and Figure 15As shown, the bit error rate changes of C-band and L-band 30 Gbaud 16QAM (i.e., 30 billion symbols per second, using 16-state quadrature amplitude modulation) wavelength division multiplexing signals after transmission and amplification are presented respectively. Figure 14 and 15 The horizontal axis represents wavelength (in nm), and the vertical axis represents bit error rate. Hollow circles represent traditional schemes, and solid circles represent this scheme. The bit error rate at the position of the dashed line in the figure is 3.8E-3.

[0192] like Figure 16 and Figure 17 As shown, the bit error rate test results for C-band and L-band 50 Gbaud 16QAM (i.e., 50 billion symbols per second, using 16-state quadrature amplitude modulation) wavelength division multiplexing signals are presented respectively. Figure 16 and 17 The horizontal axis represents wavelength (in nm), and the vertical axis represents bit error rate. Hollow circles represent traditional schemes, and solid circles represent this scheme. The bit error rate at the position of the dashed line in the figure is 3.8E-3.

[0193] from Figures 14-17 As can be seen, this scheme outperforms the traditional scheme in both the C-band and L-band bit error rate.

[0194] The test results above show that the signals of different bands amplified by the signal amplifier provided in this example can maintain good transmission performance and meet the signal quality requirements of high-speed, long-distance submarine cable communication systems.

[0195] This example demonstrates a high-gain amplification device (i.e., signal amplifier) ​​for ultra-wideband signal transmission in submarine cables, constructed based on a bidirectional cascaded discrete Raman amplification architecture. It achieves high-gain amplification of wideband signals through a two-stage collaborative cascaded discrete Raman amplification structure, reduces interstage pump crosstalk through selective band isolation components, and achieves adaptive adjustment of pump parameters through a physical information reverse design optimization network combined with intelligent optimization algorithms. This results in low-noise, high-flatness, and high-stability wideband submarine cable signal amplification capabilities.

[0196] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0197] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. In particular, the method embodiments are basically similar to the signal amplifier embodiments, so the description is relatively simple; relevant parts can be referred to the description of the method embodiments.

[0198] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of protection of this application.

Claims

1. A signal amplifier for submarine cable signal transmission, characterized in that, The signal amplifier includes: a forward pump module, a forward coupler, a first-stage amplification component, an isolation component, a second-stage amplification component, a backward coupler, and a backward pump module; The two input terminals of the forward coupler are respectively connected to the forward pump module and the submarine cable transmission optical fiber; the output terminal of the forward coupler is connected to one end of the first-stage amplification component, and the other end of the first-stage amplification component is connected to the input terminal of the isolation component; the output terminal of the isolation component is connected to one end of the second-stage amplification component, and the other end of the second-stage amplification component is connected to the bidirectional port of the backward coupler; the input terminal of the backward coupler is connected to the backward pump module. The isolation component suppresses the transmission of pump light from the secondary amplification component to the primary amplification component; The forward pump module generates a first pump light that provides gain for the S-band signal; the backward pump module generates a second pump light that provides gain for the C-band and L-band signals. The first-stage amplification component uses a first Raman gain medium to amplify the submarine cable transmission signal transmitted through the submarine cable transmission optical fiber under the action of the first pump light to obtain a first signal; the submarine cable transmission signal is an optical signal including S-band, C-band and L-band. The secondary amplification component uses a second Raman gain medium to amplify the first signal under the action of the second pump light to obtain the second signal; The signal amplifier also includes a processor, a first converter, and a second converter; The input terminal of the first converter is connected to the output terminal of the forward coupler, and the input terminal of the second converter is connected to the output terminal of the backward coupler; the processor is electrically connected to the first converter, the second converter, the forward pump module, and the backward pump module, respectively. The first converter converts the output of the forward coupler into a first electrical signal; the second converter converts the output of the backward coupler into a second electrical signal. The processor generates an adjustment instruction for adjusting the configuration parameters of the target pump light source based on the first electrical signal and the second electrical signal, and sends it to the target pump light source so that the target pump light source adjusts its own configuration parameters according to the received adjustment instruction; the target pump light source is the pump light source in the forward pump module and / or the backward pump module.

2. The signal amplifier according to claim 1, characterized in that, The processor adjusts the configuration parameters of the target pump light source in the following manner: Based on the first electrical signal and the second electrical signal, determine the current amplification performance index of the signal amplifier; In each iteration, it is determined whether there is a target pump source that needs to be adjusted in terms of configuration parameters. If so, an adjustment command carrying the adjustment step size and adjustment direction is sent to the target pump source so that the target pump source adjusts its own configuration parameters according to the adjustment step size and adjustment direction indicated by the received adjustment command. The adjustment step size decreases as the iteration increases, and the adjustment direction in each iteration is such that the predicted amplification performance index corresponding to the adjusted combination of configuration parameters of each pump source is closer to the target amplification performance index than the current amplification performance index. If the preset iteration stop condition is not met, return to the step of determining the current amplification performance index of the signal amplifier based on the first electrical signal and the second electrical signal, until the iteration stop condition is met and the adjustment stops.

3. The signal amplifier according to claim 1, characterized in that, The initial configuration parameters of each pump light source in the forward pumping module and the backward pumping module were obtained in the following manner: The processor inputs the target amplification performance index into a pre-trained inverse network to obtain the configuration parameter combination of each pump light source output by the inverse network. The inverse network is trained on a training dataset under a forward-inverse cascaded network architecture. The training dataset includes each sample configuration parameter combination and its corresponding actual amplification performance index. The actual amplification performance index corresponding to each sample configuration parameter combination represents the amplification performance that the signal amplifier can achieve when configuring each pump light source using the sample configuration parameter combination. Based on the obtained combination of configuration parameters, determine the initial configuration parameters for each pump light source.

4. The signal amplifier according to claim 3, characterized in that, The inverse network is trained in the following way: Based on the configuration parameter combinations of each sample and their corresponding actual amplification performance indicators, the forward network in the forward-reverse cascaded network is trained so that the forward network fits the first mapping relationship between the configuration parameter combinations and the amplification performance indicators. Each actual amplification performance index is input into the inverse network of the forward-inverse cascaded network, so that the inverse network outputs the predicted configuration parameter combination corresponding to each actual amplification performance index based on the inverse mapping relationship of the first mapping relationship. The Raman amplification performance corresponding to each combination of predicted configuration parameters is predicted to obtain each predicted amplification performance index. The loss value is determined based on the first difference between each predicted amplification performance index and each actual amplification performance index, and the second difference between each predicted configuration parameter combination and each sample configuration parameter combination. The network parameters of the inverse network are adjusted based on the loss value to obtain the trained inverse network.

5. The signal amplifier according to claim 1, characterized in that, The forward pumping module includes a pump combiner, multiple forward optical isolators, and multiple forward pumping light sources, wherein the number of forward optical isolators is the same as the number of forward pumping light sources. Each forward pump light source corresponds to a forward optical isolator. The output end of each forward pump light source is connected to the input end of its corresponding forward optical isolator. The output end of each forward optical isolator is connected to the input end of the pump combiner. The output end of the pump combiner is connected to the input end of the forward coupler. Each forward pump source generates pump light that provides gain for the S-band signal. Each forward optical isolator suppresses the reverse transmission of the pump light generated by the corresponding forward pump source. The pump combiner couples the multi-channel isolated pump light obtained after the pump light generated by each forward pump source passes through the forward optical isolator into a first pump light.

6. The signal amplifier according to claim 1, characterized in that, The backward pumping module includes multiple backward optical isolators and multiple backward pumping light sources, wherein the number of backward optical isolators is the same as the number of backward pumping light sources; Each backward-pumped light source corresponds to a backward optical isolator, the output terminal of each backward-pumped light source is connected to the input terminal of its corresponding backward optical isolator, and the output terminal of each backward optical isolator is connected to the input terminal of the backward coupler. Each backward pump source generates pump light that provides gain for C-band or L-band signals. Each backward optical isolator suppresses the reverse transmission of the pump light generated by the corresponding backward pump source. The backward coupler couples the multi-channel isolated pump light obtained after the pump light generated by each backward pump source passes through the backward optical isolator into a second pump light.

7. A method for signal amplification in submarine cable transmission, characterized in that, The method employs a signal amplifier for signal amplification. The signal amplifier includes: a forward pump module, a backward pump module, a forward coupler, a backward coupler, an isolation component, a first-stage amplification component, and a second-stage amplification component. The two input terminals of the forward coupler are respectively connected to the forward pump module and the submarine cable transmission optical fiber. The output terminal of the forward coupler is connected to one end of the first-stage amplification component, and the other end of the first-stage amplification component is connected to the input terminal of the isolation component. The output terminal of the isolation component is connected to one end of the second-stage amplification component, and the other end of the second-stage amplification component is connected to the bidirectional port of the backward coupler. The input terminal of the backward coupler is connected to the backward pump module. The method includes: The forward pump module outputs a first pump light that provides gain for an S-band signal; The forward coupler couples and outputs the first pump light and the submarine cable transmission signal transmitted by the submarine cable transmission fiber; the submarine cable transmission signal is an optical signal including S-band, C-band and L-band. The first-stage amplification component amplifies the submarine cable transmission signal under the action of the first pump light to obtain the first signal; The backward pump module outputs a second pump light that provides gain for C-band and L-band signals. The isolation component suppresses the transmission of pump light from the secondary amplification component to the primary amplification component. Under the action of the second pump light, the secondary amplification component amplifies the first signal to obtain the second signal. The signal amplifier further includes a processor, a first converter, and a second converter; the input terminal of the first converter is connected to the output terminal of the forward coupler, and the input terminal of the second converter is connected to the output terminal of the backward coupler; the processor is electrically connected to the first converter, the second converter, the forward pump module, and the backward pump module, respectively; the method further includes: The first converter converts the output of the forward coupler into a first electrical signal; The second converter converts the output of the backward coupler into a second electrical signal; The processor generates an adjustment command to adjust the configuration parameters of the target pump light source based on the first electrical signal and the second electrical signal, and sends it to the target pump light source so that the target pump light source adjusts its own configuration parameters according to the received adjustment command; the target pump light source is the pump light source in the forward pump module and / or the backward pump module.

8. The method according to claim 7, characterized in that, The method further includes: The processor determines the current amplification performance index of the signal amplifier based on the first electrical signal and the second electrical signal; In each iteration, it is determined whether there is a target pump source that needs to be adjusted in terms of configuration parameters. If so, an adjustment command carrying the adjustment step size and adjustment direction is sent to the target pump source so that the target pump source adjusts its own configuration parameters according to the adjustment step size and adjustment direction indicated by the received adjustment command. The adjustment step size decreases as the iteration increases, and the adjustment direction in each iteration is such that the predicted amplification performance index corresponding to the adjusted combination of configuration parameters of each pump source is closer to the target amplification performance index than the current amplification performance index. If the preset iteration stop condition is not met, return to the step of determining the current amplification performance index of the signal amplifier based on the first electrical signal and the second electrical signal, until the iteration stop condition is met and the adjustment stops.