Optical power adaptive method and system for optical fiber transmission

Abstract

The invention discloses an optical power self-adaption method and system for optical fiber transmission, and particularly relates to the field of optical fiber communication and optical transmission, and the method comprises the steps: deploying fiber core combining platforms at two ends of a machine room, and accessing a service bearer; acquiring a signal-to-noise ratio and an optical signal-to-noise ratio of each optical path, and converting a service bearer into a QoT index time sequence; self-adaptive reference power is generated based on the obtained signal-to-noise ratio and optical signal-to-noise ratio of each optical path, and a driving quantity error is obtained by combining the self-adaptive reference transmitting power; a discrete PID controller is adopted, online setting is carried out through a particle swarm algorithm, and the time weighted absolute integral of a driving quantity error is used as a cost function to realize optical power self-adaption; the system comprises an ISAP261 type optical transmission platform, a service access board card, a multiplexer / demultiplexer board card, a gain-adjustable erbium-doped fiber amplification board card, a network management electric adjustable dispersion compensation board card, a network management card and an optical monitoring board card. According to the method, the transmitting power in the multi-optical-path coexistence network is kept in a self-adaptive dynamic interval.

Description

Technical Field

[0001] This invention relates to the field of optical fiber communication and optical transmission technology, and more specifically, to an adaptive optical power method and system for optical fiber transmission. Background Technology

[0002] Traditionally, the physical optical fiber cores between backbone / metropolitan area transmission equipment rooms are often small in number (such as 12 or 24 cores) when they are first built. As business grows and existing cores age, the addition of new optical cables is constrained by investment, time, and construction conditions, making it difficult to meet the expansion needs in a timely manner.

[0003] In wavelength division multiplexing (WDM) backbones with multiple standards such as 10 / 40 / 100Gb / s, the link status changes dynamically due to amplifier gain fluctuations, channel power coupling, dispersion and nonlinear effects, as well as the addition or removal of services. Simply relying on fixed transmit power and conservative system margins during the planning phase is insufficient to balance transmit power and transmission quality.

[0004] Therefore, there is an urgent need for a method and system that can multiplex multiple services on the same fiber to meet capacity expansion requirements, while maintaining transmission quality within an adaptive dynamic range of transmit power in a multi-optical-path coexistence network. Summary of the Invention

[0005] To overcome the aforementioned deficiencies in the prior art, this invention provides an adaptive optical power method and system for optical fiber transmission. This method involves constructing a closed-loop optical performance monitoring system based on a coherent receiver, periodically collecting the signal-to-noise ratio (SNR) of each optical path, and performing SNR and optical SNR consistency conversion according to modulation scheme, symbol rate, and reference bandwidth. It also introduces online self-tuning based on a particle swarm optimization algorithm to continuously maintain the mapping between "drive quantity error → transmit power adjustment" within the adaptive dynamic range, thereby solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: An adaptive optical power method for fiber optic transmission includes the following steps: S1, fiber-to-fiber combining platforms are deployed at both ends of the data center to carry out access services; S2, obtain the signal-to-noise ratio and optical signal-to-noise ratio of each optical path, and convert the service carrying capacity into a QoT indicator time series; S3. Based on the obtained signal-to-noise ratio and optical signal-to-noise ratio of each optical path, an adaptive reference power is generated, and the driving quantity error is obtained by using the adaptive reference power and the adaptive reference transmit power. S4 employs a discrete PID controller, tuned online using a particle swarm optimization algorithm, and uses the time-weighted absolute integral of the driving quantity error as the cost function to achieve adaptive optical power. The particle swarm optimization algorithm adaptively updates the three gains of the PID controller as the link state, coupling strength, and disturbance level change, continuously maintaining the mapping between "drive quantity error → transmit power adjustment" within an adaptive dynamic range. This includes the following steps: S41, finding the most suitable PID gain for the current link is regarded as an optimization problem that rolls over time; S42 uses the time-weighted integral of the absolute value of the driving error as the cost function; S43 maps a set of particles to candidate PID gain triplet sets; S44, use the latest segment of the driving quantity error sequence obtained in S41 to calculate the time-weighted integral value of each particle; S45 updates the velocity of each particle at every moment using a triple superposition of three terms; S46, select the particle with the smallest time-weighted integral as the global optimum for the current round; S47, the control output is manifested as updating the power setting value of the transmitter, so that the drive quantity error is mapped to the transmit power adjustment quantity and kept within the adaptive dynamic range.

[0007] As a further aspect of the present invention, in S47, within each discrete update cycle, the current proportional, integral, and differential gains are first determined online by the particle swarm optimization algorithm; the current error timing is input into the discrete PID controller to generate the transmit power adjustment amount for each optical path and form a new transmit power setting value accordingly, which is implemented at the transmitting end; the control flow explicitly includes the forward delay from the transmitting end to the link output, the return delay from the receiving end to the control / execution side, and the processing and execution delay, and enters the next sampling cycle after completing "measurement-calculation-transmission-execution"; the transmit power setting value is jointly constrained by the target transmission quality, the minimum guaranteed code rate, and the upper and lower limits of the transmit power; through unified network management, network management card, and optical monitoring channel, multi-site parameter transmission and status return are realized on the 1490 / 1510nm management channel, thereby ensuring that the transmit power setting update is visible, controllable, and traceable.

[0008] As a further aspect of this invention, S1, fiber-to-fiber combining platforms are deployed at both ends of the equipment room to access services, including the following specific details: ISAP261 optical transmission platforms are mounted at both ends of the equipment room. The chassis of the optical transmission platform provides 7 general-purpose service slots, 1 network management slot, fans, and dual power supplies, supporting various services from 1.25Gbit / s to 100Gbit / s on CWDM (Coarse Wavelength Division Multiplexing) / DWDM (Dense Wavelength Division Multiplexing). On the service access side, 3R regeneration and wavelength conversion are completed through boards, which provide SFP / SFP+ (LC / UPC) interfaces and monitor the port operating status.

[0009] As a further aspect of the present invention, S2, obtaining the signal-to-noise ratio (SNR) and optical signal-to-noise ratio (OSNR) of each optical path, and converting the service carrying capacity into a QoT indicator time series, includes the following specific content: deploying an optical performance monitoring module at the receiving end of the monitored optical path at the peer site, the optical performance monitoring module being integrated into a coherent receiver based on digital signal processing, the coherent receiver being able to directly provide the signal-to-noise ratio (SNR) of the optical path, and under the condition of known modulation format, bit rate, and reference bandwidth, the OSNR can be converted from the receiving-side SNR; the coherent receiver collecting and reporting the SNR and OSNR of each optical path at a preset fixed period.

[0010] As a further aspect of the present invention, S3, based on the obtained signal-to-noise ratio and optical signal-to-noise ratio of each optical path, an adaptive reference power is generated, and the driving quantity error is obtained using the adaptive reference power and the adaptive reference transmit power, including the following specific contents: the control plane first starts from the optical performance monitoring results, and uses the known modulation format, bit rate and reference bandwidth to convert the signal-to-noise ratio given by the coherent receiver into optical signal-to-noise ratio, and obtains the "measured optical signal-to-noise ratio" of each optical path as the sensing baseline of the current control cycle.

[0011] Optical performance monitoring is enabled at the receiver of each optical path to read the real-time signal-to-noise ratio (SNR), and then the SNR is converted into the optical SNR of that optical path. Subsequently, the target optical SNR is set: In the mixed line-rate network, each optical path corresponds to a nominal bit rate and a minimum guaranteed code rate. The target optical SNR is set as the minimum usable optical SNR under the premise of satisfying the minimum guaranteed code rate and the required bit error rate, making it a hard constraint parameter of the control plane.

[0012] Based on this, in order to associate the target optical signal-to-noise ratio (OSR) and the measured OSR with the executable transmit power setting, the adaptive reference power at the current moment is generated by multiplying the ratio of the target OSR to the measured OSR by the transmit power at the previous moment. Then, the drive quantity error is defined by subtracting the adaptive reference power at the current moment from the adaptive reference transmit power.

[0013] As a further aspect of this invention, S4 employs a discrete PID controller, tuned online by a particle swarm optimization algorithm, using the time-weighted absolute integral of the driving quantity as the cost function to achieve adaptive optical power. This includes the following specific steps: The discrete PID controller performs a "proportional-integral-derivative" combined response on the driving quantity, providing a structured "parameter tuning lever" for the subsequent particle swarm optimization algorithm: In the proportional stage, an adjustment is immediately made based on the amplitude of the current driving quantity error. In the integral stage, the accumulation of the driving quantity error over time is used to eliminate steady-state deviation. In the derivative stage, the trend of driving quantity error change is suppressed to reduce overshoot and oscillation.

[0014] An optical power adaptive system for fiber optic transmission includes an ISAP261 optical transmission platform, a service access card (BD-OTU10G), a multiplexing / demultiplexing card (Mux), a gain-adjustable erbium-doped fiber amplifier card (EDFA), a network-managed electrically adjustable dispersion compensation card (TDCM), a network management card (NUM), and an optical monitoring card (OSC). The optical power adaptive system uses the ISAP261 optical transmission platform as its core. The service access card at the front end performs multi-standard access and wavelength conversion. The optical layer uses the multiplexing / demultiplexing card for multiplexing / demultiplexing. The gain-adjustable erbium-doped fiber amplifier card ensures power and noise margins across distances. The network-managed electrically adjustable dispersion compensation card suppresses dispersion in high-speed links. The network-managed electrically adjustable dispersion compensation card and the optical monitoring card provide unified management and alarms across different sites.

[0015] The technical effects and advantages of the optical power adaptive method and system for optical fiber transmission proposed in this invention are as follows: The proposed optical power adaptive method and system takes "meeting the QoT constraint + minimizing transmit power" as the control objective. Addressing the time-varying and uncertainties of link states caused by amplifier gain fluctuations, channel power coupling, group velocity dispersion, and nonlinearity, a closed-loop optical performance monitoring system based on a coherent receiver is constructed. The SNR of each optical path is periodically collected and converted to a consistent OSNR according to the modulation scheme, symbol rate, and reference bandwidth, forming a QoT time series as a unified metric on the control side. The control layer employs discrete PID control with online self-tuning using a particle swarm optimization algorithm, using the time-weighted absolute error integral as the cost function. This significantly improves the recovery speed after disturbances, suppresses overshoot, and reduces steady-state residuals, maintaining rapid convergence of transmit power adjustment under coupled multi-optical path conditions. To ensure robustness, the system reduces conservative system margins and unnecessary power consumption while guaranteeing the minimum guaranteed bit rate and target bit error rate threshold. In terms of engineering implementation, the system is centered on a fiber core combining platform. The front end uses service access and multiplexing / demultiplexing boards to complete multi-standard access and multiplexing / demultiplexing. The span is provided by a gain-adjustable EDFA to provide power and noise margin (noise figure less than 6dB, flatness less than 1.5dB). High-speed long-distance links achieve electrically adjustable compensation in the C-band at a range of 0-60km and ±800ps / nm through a network-managed electrically adjustable dispersion compensation board. A unified network management card and optical monitoring board complete cross-site measurement, command issuance, and traceable alarms. This enables efficient reuse and energy efficiency improvement of existing fiber resources, reduces the probability of service interruption, and enhances robustness and maintainability under multi-service concurrency. It has practical value for smooth implementation in existing networks. Attached Figure Description

[0016] Figure 1 This is a flowchart of an adaptive optical power method for optical fiber transmission according to the present invention.

[0017] Figure 2 This is a schematic diagram of the application scenario of the fiber core combining platform of the present invention.

[0018] Figure 3 This is a diagram showing the appearance and slot structure of the ISAP261 type optical transmission platform of the present invention.

[0019] Figure 4 This is a rolling curve of ITAE over time according to the present invention.

[0020] In the diagram: SDH2.5G represents equipment with a Synchronous Digital Hierarchy speed of approximately 2.5Gb / s, IPRUN represents an Ethernet-based wireless access bearer network, and SDH10G represents equipment with a Synchronous Digital Hierarchy speed of approximately 10Gb / s. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1

[0023] See Figure 1 The present invention provides an adaptive optical power method for optical fiber transmission, comprising the following steps: S1, fiber optic combining platforms are deployed at both ends of the equipment room, the fiber optic combining platforms as follows: Figure 2 As shown, the access service is carried out; S2, obtain the signal-to-noise ratio and optical signal-to-noise ratio of each optical path, and convert the service carrying capacity into a QoT indicator time series; S3. Based on the obtained signal-to-noise ratio and optical signal-to-noise ratio of each optical path, an adaptive reference power is generated, and the driving quantity error is obtained by using the adaptive reference power and the adaptive reference transmit power. S4 employs a discrete PID controller, tuned online using a particle swarm optimization algorithm, and uses the time-weighted absolute integral of the driving quantity error as the cost function to achieve adaptive optical power.

[0024] Furthermore, as a further embodiment of the present invention, S1, a fiber-to-fiber combining platform is deployed at both ends of the equipment room to access services, including: mounting an ISAP261 type optical transmission platform at each end of the equipment room, wherein the appearance and slot structure of the ISAP261 type optical transmission platform are as follows. Figure 3 As shown, the optical transmission platform chassis provides 7 general-purpose service slots, 1 network management slot, fans, and dual power supplies, supporting various services ranging from 1.25Gbit / s to 100Gbit / s on CWDM / DWDM. The service access side completes 3R regeneration and wavelength conversion through boards, providing SFP / SFP+ (LC / UPC) interfaces and monitoring the port operating status.

[0025] Multi-channel services are fed into the fiber via multiplexing / demultiplexing boards. Channel insertion loss and isolation between adjacent / non-adjacent channels are clearly defined. The interface is LC / PC, passive, and low-power, effectively saving fiber optic resources. Long-distance or multi-segment links can be inserted with EDFAs (power / line / pre-amplifier types) as needed, supporting 40 / 80 / 96-wavelength DWDM with a noise figure <6dB and flatness <1.5dB. The network management system can monitor input / output optical power, gain, and temperature. For high-speed long-distance links, a TDCM can be optionally installed, employing G.652 slope-matched electrically adjustable dispersion compensation, covering the entire C-band channel, with a compensation range of 0–60km and adjustable dispersion of +800 to −800 ps / nm.

[0026] Further, S2, acquiring the signal-to-noise ratio (SNR) and optical signal-to-noise ratio (OSNR) of each optical path, and converting service carrying capacity into a QoT indicator time series, includes: deploying an optical performance monitoring module at the receiving end of the monitored optical path at the peer site; the optical performance monitoring module is integrated in a coherent receiver based on digital signal processing; the coherent receiver can directly provide the SNR of the optical path, and under the condition of known modulation format, bit rate, and reference bandwidth, the OSNR can be converted from the SNR at the receiving side to characterize the link noise margin and service carrying capacity; the coherent receiver collects and reports the SNR and OSNR of each optical path at a preset fixed period.

[0027] In this embodiment, the process of converting the signal-to-noise ratio (SNR) to the optical SNR is as follows: First, based on the modulation format and whether dual polarization is used, the bit rate is mapped to the symbol rate to determine the equivalent noise bandwidth after the receiver's matched filter. Then, the measured electrical noise power is normalized according to the bandwidth, making it equivalent to a 12.5 GHz reference window, thus converting the electrical SNR into an SNR measured by the optical reference bandwidth. Next, a small correction is made according to the actual roll-off and bandwidth shape of the receiver / optical filter to ensure the consistency of bandwidth normalization. Regarding noise composition, the conversion steps maintain the same statistical caliber as the SNR estimation. The calculation takes into account key physical terms such as link amplified spontaneous emission, inter-channel out-of-band crosstalk, XPM (cross-phase modulation) / SPM (self-phase modulation) nonlinearity, and receiver shot and thermal noise. For different modulation schemes, the calculation applies their respective format-dependent mapping constants and calibration coefficients to eliminate systematic deviations caused by different demodulation paths, estimation algorithms, and filter shapes, ensuring that the optical signal-to-noise ratio derived from the signal-to-noise ratio is consistent with the QoT model of that format. Finally, the optical signal-to-noise ratio of each optical path is output and timestamped back to the control plane, which receives the QoT estimation results (signal-to-noise ratio and optical signal-to-noise ratio) reported by the optical performance monitoring.

[0028] Further, S3, based on the obtained signal-to-noise ratio and optical signal-to-noise ratio of each optical path, an adaptive reference power is generated, and the driving quantity error is obtained by using the adaptive reference power and the adaptive reference transmit power. This includes: the control surface first starts from the optical performance monitoring results, and uses the known modulation format, bit rate and reference bandwidth to convert the signal-to-noise ratio given by the coherent receiver into the optical signal-to-noise ratio, and obtains the "measured optical signal-to-noise ratio" of each optical path as the sensing baseline of the current control cycle.

[0029] Optical performance monitoring is enabled at the receiver of each optical path to read the real-time signal-to-noise ratio (SNR), and then the SNR is converted into the optical SNR of that optical path. Subsequently, the target optical SNR is set: In the mixed line-rate network, each optical path corresponds to a nominal bit rate and a minimum guaranteed code rate. The target optical SNR is set as the minimum usable optical SNR under the premise of satisfying the minimum guaranteed code rate and the required bit error rate, making it a hard constraint parameter of the control plane.

[0030] Based on this, in order to associate the target optical signal-to-noise ratio (OSR) and the measured OSR with the executable transmit power setting, the adaptive reference power at the current moment is generated by multiplying the ratio of the target OSR to the measured OSR by the transmit power at the previous moment. Then, the drive quantity error is defined by subtracting the adaptive reference power at the current moment from the adaptive reference power. The adaptive reference transmit power is obtained by proportionally scaling the actual transmit power of the optical path at the previous moment using the ratio of the target OSR to the measured OSR, resulting in a reference power adaptively scaled according to the QoT notch.

[0031] Further, in step S4, a discrete PID controller is employed and tuned online using a particle swarm optimization algorithm. The time-weighted absolute integral of the driving quantity is used as the cost function to achieve adaptive optical power. This includes: using a discrete PID controller to perform a "proportional-integral-derivative" combined response on the driving quantity, providing a structured "parameter tuning lever" for the subsequent particle swarm optimization algorithm. In the proportional stage, an adjustment is immediately given based on the magnitude of the current driving quantity error. That is, the discrete PID controller directly multiplies the driving quantity error by the proportional gain to obtain the instantaneous power correction for the current iteration. Because the control law applies a negative sign to the proportional term in the transfer function, the direction of the correction is opposite to the error. If the current transmit power is higher than the reference (positive error), the proportional stage immediately corrects by lowering the power; if it is lower than the reference (negative error), the power is immediately increased, thus pulling the controlled quantity back to the target trajectory as quickly as possible. Subsequently, before the proportional correction is superimposed with the integral and derivative outputs, it is subject to upper and lower power constraints (including the minimum and maximum allowable transmit power for each optical path) to prevent the executed quantity from exceeding the safe range of the equipment and link. In the integral stage, the accumulation of the driving error over time is used to eliminate steady-state deviation. That is, the integral continuously accumulates the driving error over time (in discrete implementation, the error is multiplied by the sampling period and added to the integral for each sampling step), and this integral is superimposed on the control output, causing the control quantity to slowly drift in the direction of "reducing error." As long as the driving error is not zero, the integral continues to grow or decay until the control output exactly cancels out that constant disturbance / deviation, the controlled quantity coincides with the target, and the driving error returns to zero. After this, when the driving error is zero, the integral no longer changes, effectively forming a "new static bias" to maintain the equilibrium point. In the derivative term, the trend of the driving error is suppressed to reduce overshoot and oscillation. In other words, the derivative term adopts a feedforward suppression of the driving error trend. That is, the difference between the current error and the previous error is amplified by a coefficient and incorporated into the control law. Therefore, when the driving error is increasing rapidly, the derivative term generates a reverse correction in time to offset the excessive driving force that the proportional and integral terms may bring in advance, thereby suppressing overshoot. When the driving error decreases rapidly and approaches the target, the derivative term weakens the change amplitude of the control output to avoid secondary backshoot caused by lag and coupling, thereby reducing the oscillation amplitude.

[0032] In this embodiment, unlike traditional fixed tuning, the present invention does not fix the proportional, integral, and derivative terms to a certain set of constants. Instead, it introduces online self-tuning based on the particle swarm optimization algorithm, allowing the three gains of the PID to be adaptively updated as the link state, coupling strength, and disturbance level change. This continuously keeps the mapping of "driving quantity error → transmit power adjustment" within an adaptive dynamic range, including the following steps: S41, the online particle swarm tuning is regarded as an optimization problem that rolls over time, which is to "find the most suitable PID gain for the current link". The optical performance monitoring at the receiver is integrated into a coherent receiver based on digital signal processing. First, the signal-to-noise ratio of each optical path is obtained. Under the premise of preset modulation format, bit rate and reference bandwidth, the signal-to-noise ratio is converted into optical signal-to-noise ratio and used as QoT estimation input to the control plane. Then, the adaptive reference power at the current moment is generated by multiplying the ratio of the target optical signal-to-noise ratio and the measured optical signal-to-noise ratio with the transmit power at the previous moment. Finally, the drive quantity error is defined by subtracting the adaptive reference power at the current moment from the adaptive reference power.

[0033] S42 uses the time-weighted integral (ITAE) of the absolute value of the driving error as the cost function, with a larger weight for later time steps. This penalizes residual errors and forces a solution that is better in steady state and disturbance rejection. For multiple interacting optical paths on the same route, the definition of the time-weighted integral also explicitly counts the number of coupled optical paths, ensuring that the tuning reflects the amplification effect of coupling on the error. In this embodiment, as... Figure 4 As shown, the "time-weighted absolute error integral" is calculated for the error time series using a 30-second window. This means that the error weight increases as it occurs later in the time series. Therefore, after two disturbance injections at 50s and 120s (where the error suddenly increases and then decays), the rolling ITAE shows two significant increases and gradually decreases under control. When the system enters steady state and both the error magnitude and duration decrease simultaneously, the rolling ITAE remains at a lower level. This curve intuitively reflects the measurement characteristics of ITAE: it not only focuses on the magnitude of the error but also penalizes the duration of the error, thus being more sensitive to "long-tailed errors" and "recovery speed." In a closed-loop system using PID+PSO online self-tuning, if the parameters are properly selected, the curve should show a rapid surge after a disturbance, followed by a monotonic or step-like decrease and stabilization. This can be used as a comprehensive performance indicator to evaluate the performance of different tuning strategies in terms of disturbance rejection, tuning time, and steady-state residuals.

[0034] S43, a set of particles is mapped to a candidate PID gain triplet (Kp, Ki, Kd; where Kp is the proportional gain, i.e. the amplification factor of the error at the current time; Ki is the integral gain, i.e. the coefficient that weights the cumulative amount of error over time; Kd is the differential gain, i.e. the coefficient that weights the rate of change of error). Each particle maintains both the candidate gain and the search step size. In the search subspace, each particle records its "individual historical best position", and all particles jointly maintain the "group global best position".

[0035] S44, at each discrete time, calculate the time-weighted integral value of each particle using the latest segment of the driving quantity error sequence obtained in S41; the smaller the time-weighted integral value, the more suitable the group (Kp, Ki, Kd) is under the current link and coupling conditions.

[0036] In step S45, the particle swarm optimization algorithm updates the velocity of each particle at each time step using a superposition of three terms: an inertial weight term for the velocity at the previous time step, an attraction term toward the individual's historical best position, and an attraction term toward the global best position of the swarm. Both attraction terms are multiplied by a 0–1 uniformly distributed random factor and a corresponding acceleration coefficient to balance exploration and convergence. The current position of each particle is added to its new velocity to obtain the new position, thus yielding the next generation of candidate PID parameters. Then, the algorithm returns to step S44 to recalculate its fitness.

[0037] S46. At the end of each iteration, the particle with the smallest time-weighted integral is selected as the global optimum for that round, and its (Kp, Ki, Kd) is written into the discrete PID controller. The PID controller maps the current driving error to the "transmit power adjustment" of each optical path, which is then used as the control output for the next step. The specific mapping process is as follows: In each discrete update cycle, the control surface first selects the three globally optimal PID gains (proportional, integral, and derivative) using the particle swarm optimization algorithm, and then sends the "current driving error" into the discrete PID controller: the proportional term directly corrects the current deviation, the integral term accumulates historical deviations to eliminate steady-state residuals, and the derivative term "brakes" in advance according to the trend of error change. The three are superimposed to form the "transmit power adjustment" for each optical path. The selection objective is to minimize the time-weighted integral, so that the obtained (Kp, Ki, Kd) can be self-tuned online according to the link status and disturbances.

[0038] S47, the control output is manifested as updating the power setpoint of the transmitter, thereby keeping the drive quantity error mapped to the transmit power adjustment within an adaptive dynamic range: At each discrete update moment, the particle swarm algorithm first provides the optimal proportional, integral, and derivative gains for the current round. The discrete PID controller takes the "current error sequence" as input and maps the error signal to the transmit power adjustment of each optical path through its closed-loop linear power controller, generating a new transmitter power setpoint accordingly. This setpoint is then implemented at the transmitter as an execution command, forming a direct adjustment of the optical path's transmit power, thus projecting the mapping of "drive quantity error → transmit power adjustment" onto the executable parameters on the device side in real time. This process operates independently on a link-by-link basis within a distributed closed-loop framework: the control plane explicitly includes the forward delay from the transmitter to the link output, the backward delay from the receiver back to the control / execution side, and the processing and execution delay; after the control command completes "measurement—calculation—issuance—execution," it enters the next sampling update. To ensure the safety and feasibility of execution, the new transmit power setting must meet hard constraints imposed by both the algorithm and physics, including the target optical signal-to-noise ratio, minimum guaranteed code rate, and upper and lower limits of transmit power. This ensures that the adjustment output by the PID can both adapt to the link status and always remain within the allowable operating range, avoiding pushing the system into overdrive or saturation. At the engineering level, the device management and execution link is implemented through the network management system and network management card (NMU) and optical monitoring board (OSC): the network management card is used to collect board status and issue control information, and the optical monitoring board carries management signaling at 1490 / 1510nm to access the unified network management system, making the update of the transmitter power setting and subsequent monitoring and alarms visible, controllable, and traceable across multiple sites.

[0039] An optical power adaptive system for fiber optic transmission includes an ISAP261 optical transmission platform, a service access card (BD-OTU10G), a multiplexing / demultiplexing card (Mux), a gain-adjustable erbium-doped fiber amplifier card (EDFA), a network-managed electrically adjustable dispersion compensation card (TDCM), a network management card (NUM), and an optical monitoring card (OSC). The optical power adaptive system uses the ISAP261 optical transmission platform as its core. The service access card at the front end performs multi-standard access and wavelength conversion. The optical layer uses the multiplexing / demultiplexing card for multiplexing / demultiplexing. The gain-adjustable erbium-doped fiber amplifier card ensures power and noise margins across distances. The network-managed electrically adjustable dispersion compensation card suppresses dispersion in high-speed links. The network-managed electrically adjustable dispersion compensation card and the optical monitoring card provide unified management and alarms across different sites.

[0040] The optical power adaptive method and system proposed in this invention take "meeting the QoT constraint + minimizing transmit power" as the control objective. Addressing the time-varying and uncertainties of link states caused by amplifier gain fluctuations, channel power coupling, group velocity dispersion, and nonlinearity, a closed-loop optical performance monitoring system based on a coherent receiver is constructed. The SNR of each optical path is periodically collected and converted to a consistent OSNR according to the modulation scheme, symbol rate, and reference bandwidth, forming a QoT time series as a unified metric on the control side. The control layer employs discrete PID control with online self-tuning using a particle swarm optimization algorithm. Using the time-weighted absolute error integral as the cost function, it significantly improves the recovery speed after disturbances, suppresses overshoot, and reduces steady-state residuals. Under coupled multi-optical-path conditions, it maintains rapid convergence and robustness of transmit power adjustment, thereby ensuring the minimum guaranteed code rate. To reduce conservative system margins and unnecessary power consumption while meeting the target bit error rate threshold, the system is implemented with a fiber-core combining platform at its core. The front end uses service access and multiplexing / demultiplexing boards to handle multi-standard access and multiplexing / demultiplexing. The span is provided by a gain-adjustable EDFA, offering power and noise margins (noise figure less than 6dB, flatness less than 1.5dB). High-speed, long-distance links utilize a network-managed electrically adjustable dispersion compensation board to achieve adjustable compensation within the C-band range of 0–60km at ±800ps / nm. A unified network management card and optical monitoring board handle cross-site measurements, command issuance, and traceable alarms. This achieves efficient reuse and energy efficiency improvement of existing fiber resources, reduces the probability of service interruption, and enhances robustness and maintainability under multi-service concurrency, demonstrating practical value for smooth deployment in existing networks.

[0041] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0042] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An adaptive optical power method for optical fiber transmission, characterized in that, Includes the following steps: S1, fiber-to-fiber combining platforms are deployed at both ends of the data center to carry out access services; S2, obtain the signal-to-noise ratio and optical signal-to-noise ratio of each optical path, and convert the service carrying capacity into a QoT indicator time series; S3. Based on the obtained signal-to-noise ratio and optical signal-to-noise ratio of each optical path, an adaptive reference power is generated, and the driving quantity error is obtained by using the adaptive reference power and the adaptive reference transmit power. S4 employs a discrete PID controller, tuned online using a particle swarm optimization algorithm, and uses the time-weighted absolute integral of the driving quantity error as the cost function to achieve adaptive optical power. The particle swarm optimization algorithm adaptively updates the three gains of the PID controller as the link state, coupling strength, and disturbance level change, continuously maintaining the mapping between "drive quantity error → transmit power adjustment" within an adaptive dynamic range. This includes the following steps: S41, finding the most suitable PID gain for the current link is regarded as an optimization problem that rolls over time; S42 uses the time-weighted integral of the absolute value of the driving error as the cost function; S43 maps a set of particles to candidate PID gain triplet sets; S44, use the latest segment of the driving quantity error sequence obtained in S41 to calculate the time-weighted integral value of each particle; S45 updates the velocity of each particle at every moment using a triple superposition of three terms; S46, select the particle with the smallest time-weighted integral as the global optimum for the current round; S47, the control output is manifested as updating the power setting value of the transmitter, so that the drive quantity error is mapped to the transmit power adjustment quantity and kept within the adaptive dynamic range.

2. The optical power adaptive method for optical fiber transmission according to claim 1, characterized in that, In S47, within each discrete update cycle, the particle swarm optimization algorithm first determines the current proportional, integral, and differential gains online; the current error timing is input into the discrete PID controller to generate the transmit power adjustment for each optical path and form a new transmit power setting value accordingly. This setting value is implemented at the transmitting end; the control flow explicitly includes the forward delay from the transmitting end to the link output, the return delay from the receiving end to the control / execution side, and the processing and execution delay, and enters the next sampling cycle after completing "measurement-calculation-transmission-execution"; the transmit power setting value is jointly constrained by the target transmission quality, the minimum guaranteed code rate, and the upper and lower limits of the transmit power; through unified network management, network management card, and optical monitoring channel, multi-site parameter transmission and status return are realized on the 1490 / 1510nm management channel, thereby ensuring the visibility, controllability, and traceability of the transmit power setting update.

3. The optical power adaptive method for optical fiber transmission according to claim 1, characterized in that, The multi-channel services of the fiber core combining platform described in S1 are multiplexed into the fiber via multiplexing and demultiplexing boards, and demultiplexed via remote network management cards. Long-distance or multi-segment links are inserted with gain-adjustable erbium-doped fiber amplifier boards. High-speed long-distance links are equipped with network management electrically adjustable dispersion compensation boards, which adopt G.652 slope matching electrically adjustable dispersion compensation to cover the entire C-band.

4. The optical power adaptive method for optical fiber transmission according to claim 1, characterized in that, In S2, the signal-to-noise ratio (SNR) and optical signal-to-noise ratio (SNR) are obtained through an optical performance monitoring module. The optical performance monitoring module is integrated into a coherent receiver based on digital signal processing. The coherent receiver collects and reports data at a fixed period and converts the SNR into optical SNR when the modulation format, bit rate, and reference bandwidth are known.

5. The optical power adaptive method for optical fiber transmission according to claim 1, characterized in that, The steps for converting the signal-to-noise ratio (SNR) to the optical SNR in S2 are as follows: based on the modulation format and whether dual polarization is used, the bit rate is mapped to the symbol rate to determine the equivalent noise bandwidth; the electrical noise power is normalized to the 12.5 GHz reference window according to the bandwidth; the roll-off of the receiver / optical filter and the bandwidth shape are corrected; and the spontaneous emission of the link amplification, channel crosstalk, nonlinearity and receiver noise are taken into account. By applying modulation format-related mapping constants and calibration coefficients, the time-stamped optical signal-to-noise ratio is output to the control plane.

6. The optical power adaptive method for optical fiber transmission according to claim 1, characterized in that, The adaptive reference power in S3 is generated by multiplying the ratio of the target optical signal-to-noise ratio to the measured optical signal-to-noise ratio by the transmission power at the previous moment; the adaptive reference transmission power is obtained by scaling the ratio of the target optical signal-to-noise ratio to the current optical signal-to-noise ratio to the actual transmission power at the previous moment; the driving quantity error is the difference between the adaptive reference power and the adaptive reference transmission power.

7. The optical power adaptive method for optical fiber transmission according to claim 1, characterized in that, The working process of the discrete PID controller described in S4 is as follows: the proportional element multiplies the driving error with the proportional gain to generate an instantaneous power correction, the direction of which is opposite to the error; the integral element accumulates the driving error over time to eliminate steady-state deviation; the derivative element calculates the difference between "current error and previous error" to suppress overshoot and oscillation; the correction is constrained by the upper and lower limits of the transmit power before being superimposed.

8. The optical power adaptive method for optical fiber transmission according to claim 1, characterized in that, The target optical signal-to-noise ratio (SNR) described in S3 is set as the minimum SNR that satisfies the nominal bit rate, minimum guaranteed bit rate, and required bit error rate of each optical path, and is used as a hard constraint parameter of the control plane.

9. An adaptive optical power system for optical fiber transmission, characterized in that, An optical power adaptive method for optical fiber transmission as described in any one of claims 1-8 includes: an ISAP261 type optical transmission platform, a service access board, a multiplexing / demultiplexing board, a gain-adjustable erbium-doped fiber amplifier board, a network management electrically adjustable dispersion compensation board, a network management card, and an optical monitoring board.

Images