Collaborative arrangement method and system for communication and inductance integration, electronic equipment and storage medium

Abstract

The invention provides a communication and sensing integrated collaborative arrangement method and system, electronic equipment and a storage medium. The method comprises the following steps: acquiring basic state information of a communication and sensing integrated optical network; generating a sensing task descriptor according to a sensing task demand; determining a multi-dimensional interference vector based on the basic state information and the perception task descriptor, and generating a candidate configuration set based on the optical layer physical resource state and the multi-dimensional interference vector; in the candidate configuration set, solving by taking minimization of service level protocol deviation degree of the communication service, minimization of network resource occupation and maximization of perception task confidence as a joint optimization target to obtain an optimal communication slicing strategy and an optimal perception slicing strategy; issuing an execution instruction to the underlying network equipment based on the optimal communication slicing strategy and the sensing slicing strategy, and obtaining the fluctuation slope of the communication service performance index; and when the fluctuation slope exceeds a preset security threshold, recovering the network configuration to the last stable version.

Description

Technical Field

[0001] This application relates to the technical field of sensory integrated networks, and more particularly to a sensory integrated collaborative orchestration method, system, electronic device, and storage medium. Background Technology

[0002] The core objective of the integrated optical network is to fully utilize existing optical fiber resources to integrate diverse sensing technologies such as OTDR (Optical Time Domain Reflectometry), DAS (Distributed Acoustic Sensing), and DTS (Distributed Temperature Sensing) while ensuring communication functions, thereby achieving efficient collaborative operation that integrates communication transmission, security monitoring, and network maintenance.

[0003] Current sensing solutions (such as traditional optical time domain reflectometers (OTDRs) typically only focus on whether the physical power of the sensing pulse meets the standard. However, in actual 100G / 400G high-speed transmission links, high-power pulses generated by sensing devices may induce stimulated Brillouin scattering (SBS) or cross-phase modulation (XPM) phenomena. Even if the transmitting power at the sensing end is within the safe threshold, the bit error rate (BER) on the communication side will still fluctuate significantly and periodically with the scanning pulse, causing instantaneous off-target errors or packet loss in existing network services. Summary of the Invention

[0004] This application provides a sensor-integrated collaborative orchestration method, system, electronic device, and storage medium to address the problems existing in related technologies. The technical solution is as follows: In a first aspect, embodiments of this application provide a synergistic integrated collaborative orchestration method, including: Acquire basic status information of the integrated sensing and communication optical network. The basic status information includes the service level agreement indicators of communication services, the status of optical layer physical resources, and the requirements of sensing tasks to be performed. Based on the requirements of the perception task, generate a perception task descriptor that includes task priority and maximum interference budget; Based on basic state information and sensing task descriptors, a multidimensional interference vector is determined. The multidimensional interference vector is used to quantify the nonlinear physical layer impact of the detection signal generated by the sensing task on the communication service signal. Based on the physical resource status of the optical layer and multi-dimensional interference vectors, feasible domain pruning is performed on the combination of configuration parameters for sensing tasks to generate a candidate configuration set that meets the service level protocol indicators of communication services. In the candidate configuration set, the optimal communication slicing strategy and the optimal perception slicing strategy are obtained by solving the joint optimization objective of minimizing the service level protocol deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks. Based on the optimal communication slicing strategy and perception slicing strategy, execution instructions are issued to the underlying network devices to obtain the fluctuation slope of communication service performance indicators. When the fluctuation slope exceeds the preset safety threshold, the network configuration is restored to the previous stable version. When the fluctuation slope does not exceed the safety threshold and the perception confidence does not meet the standard, a step-by-step enhancement strategy is executed to adjust the perception slicing strategy.

[0005] Secondly, embodiments of this application provide a sensor-integrated collaborative orchestration system, including: The first acquisition module is used to acquire the basic status information of the integrated optical network, including the service level agreement indicators of communication services, the status of optical layer physical resources, and the requirements of the sensing tasks to be executed. The first generation module is used to generate a perception task descriptor that includes task priority and maximum interference budget according to the perception task requirements. The first determining module is used to determine the multi-dimensional interference vector based on the basic state information and the sensing task descriptor. The multi-dimensional interference vector is used to quantify the nonlinear physical layer impact of the detection signal generated by the sensing task on the communication service signal. The first generation module is used to perform feasible domain pruning on the combination of configuration parameters for sensing tasks based on the physical resource status of the optical layer and multi-dimensional interference vectors, and generate a candidate configuration set that meets the service level protocol indicators of communication services. The first module is used to solve the candidate configuration set with the joint optimization objectives of minimizing the service level protocol deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks, so as to obtain the optimal communication slicing strategy and perception slicing strategy. The second acquisition module is used to issue execution instructions to the underlying network devices based on the optimal communication slicing strategy and perception slicing strategy to obtain the fluctuation slope of communication service performance indicators; when the fluctuation slope exceeds the preset safety threshold, the network configuration is restored to the previous stable version; when the fluctuation slope does not exceed the safety threshold and the perception confidence does not meet the standard, a step-by-step enhancement strategy is executed to adjust the perception slicing strategy.

[0006] Thirdly, embodiments of this application provide an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to execute the aforementioned synergistic orchestration method.

[0007] Fourthly, embodiments of this application provide a computer-readable storage medium that stores computer instructions, wherein when the computer instructions are executed on a computer, the methods in any of the above-described embodiments are performed.

[0008] The advantages or beneficial effects of the above technical solutions include at least the following: In this embodiment, the method addresses the core challenge of high-power pulses from sensing signals easily inducing fiber nonlinear effects (such as cross-phase modulation, XPM), thereby degrading communication quality. By using multi-dimensional interference vectors, the abstract and complex physical layer nonlinear interference is decoupled into specific mathematical vectors such as the predicted bit error rate increment and the optical signal-to-noise ratio degradation value. This allows the orchestration system to accurately quantify the potential physical damage of sensing tasks to the Communication Service Level Agreement (SSLA), achieving a qualitative leap from the traditional, broad-based power threshold to refined interference budget management. Furthermore, it effectively solves the problem of high-power pulses generated by sensing devices potentially triggering stimulated Brillouin scattering (SBS) or cross-phase modulation (XPM) phenomena. Even if the transmitting power at the sensing end is within a safe threshold, the bit error rate (BER) on the communication side will still exhibit significant periodic fluctuations with the scanning pulse, causing instantaneous off-target errors or packet loss in existing network services.

[0009] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of this application will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0010] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the various drawings denote the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments disclosed in this application and should not be construed as limiting the scope of this application.

[0011] Figure 1 This is a flowchart of a synesthetic integrated collaborative orchestration method according to an embodiment of this application.

[0012] Figure 2 This is a block diagram of a synesthetic system for a synesthetic integrated collaborative orchestration method according to an embodiment of this application.

[0013] Figure 3 This is a diagram illustrating an example of a synergistic integrated service in a synergistic integrated collaborative orchestration method according to an embodiment of this application.

[0014] Figure 4 This is a block diagram of an electronic device according to an embodiment of the present application. Detailed Implementation

[0015] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of this application. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0016] In the relevant technical field, existing attempts at "integrated sensing and communication" in the actual maintenance of metropolitan area networks and trunk optical cables are mostly concentrated in the laboratory or small-scale pilot stages. Once they enter the complex Wavelength Division Multiplexing (WDM) network environment, many problems that do not conform to the actual situation will be exposed: The "invisible killer" effect of nonlinear interference: Existing sensing solutions, such as traditional optical time domain reflectometers (OTDRs), typically only focus on whether the physical power of the sensing pulse meets the standard. However, in actual 100G / 400G high-speed transmission links, high-power pulses generated by sensing devices are highly susceptible to stimulated Brillouin scattering (SBS) or cross-phase modulation (XPM). Studies have found that even if the transmit power at the sensing end is within the so-called "safe threshold" range, the bit error rate (BER) on the communication side will still fluctuate drastically and periodically with the scanning pulse, leading to momentary off-target or packet loss in current network services. This nonlinear crosstalk cannot be avoided by traditional static isolation schemes.

[0017] The disconnect between perception strategies and high-risk areas: Most existing inspections adopt a "uniform scanning across the entire line" approach. However, at construction sites, high-risk areas (such as subway construction zones, entrances to core Internet Data Center (IDC) server rooms, or cross-road sections with a history of frequent failures) urgently require high-frequency, high-confidence focused protection. Due to a lack of adaptive capabilities, the system wastes a significant amount of ineffective scanning resources in safe areas, while failing to provide focused and enhanced perception capabilities in truly critical areas.

[0018] Lack of automated "oil cut-off protection" mechanism: Currently, the sensing system and communication network management are basically operating in an "open-loop" state. Once the business (KPI, Key Performance Indicator) deteriorates, maintenance personnel often can only take the most extreme "manual shutdown" method or directly disconnect the sensing jumper. There is a lack of a "rush-off prevention" mechanism that can automatically and smoothly roll back according to the slope of business fluctuations. This seriously limits the large-scale deployment of the sensing system in the core production network.

[0019] Figure 1 A flowchart illustrating a synesthetic integrated collaborative orchestration method according to an embodiment of this application is shown. Figures 1-3As shown, a synesthetic integrated collaborative orchestration method may include: S110: Obtain basic status information of the integrated optical network, including service level agreement indicators of communication services, status of optical layer physical resources, and requirements for sensing tasks to be performed. S120: Generate a perception task descriptor that includes task priority and maximum interference budget based on the perception task requirements; S130: Based on the basic state information and the sensing task descriptor, determine the multi-dimensional interference vector. The multi-dimensional interference vector is used to quantify the nonlinear physical layer impact of the detection signal generated by the sensing task on the communication service signal. S140: Based on the physical resource status of the optical layer and multi-dimensional interference vectors, perform feasible domain pruning on the combination of configuration parameters for the sensing task to generate a candidate configuration set that meets the service level protocol indicators of the communication service. S150: In the candidate configuration set, the optimal communication slicing strategy and the optimal perception slicing strategy are obtained by solving the problem with the joint optimization objectives of minimizing the service level protocol deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks. S160: Based on the optimal communication slicing strategy and perception slicing strategy, issue execution instructions to the underlying network devices to obtain the fluctuation slope of communication service performance indicators; when the fluctuation slope exceeds the preset safety threshold, restore the network configuration to the previous stable version; when the fluctuation slope does not exceed the safety threshold and the perception confidence does not meet the standard, execute the step enhancement strategy to adjust the perception slicing strategy.

[0020] In this embodiment, addressing the core challenge of high-power pulses from sensing signals easily inducing fiber nonlinear effects (such as cross-phase modulation, XPM) and thus degrading communication quality, the method employs multi-dimensional interference vectors to decouple abstract and complex physical layer nonlinear interference into specific mathematical vectors such as bit error rate increment prediction and optical signal-to-noise ratio degradation. This allows the orchestration system to accurately quantify the potential physical damage of sensing tasks to the Communication Service Level Agreement (SSLA), achieving a qualitative leap from the traditional, broad-based power threshold to refined interference budget management. Furthermore, it effectively solves the problem of high-power pulses generated by sensing devices potentially triggering stimulated Brillouin scattering (SBS) or cross-phase modulation (XPM). Even if the transmitting power at the sensing end is within a safe threshold, the bit error rate (BER) on the communication side will still exhibit significant periodic fluctuations with the scanning pulse, causing instantaneous off-target errors or packet loss in existing network services.

[0021] To address the issue of static model failure caused by dynamic fluctuations in the optical network channel environment over time (such as temperature drift and aging), the embodiments of this application employ a dynamic safety margin calculation based on historical fluctuation characteristics and feasible region pruning. This utilizes statistical variance to dynamically adjust the protection boundary (smaller margin for stable channels, larger margin for chaotic channels), and pre-eliminates physically infeasible solutions. This logic not only significantly reduces the computational complexity of subsequent multi-objective optimization algorithms (improving convergence speed), but also endows the scheduling strategy with strong robustness to environmental jitter, avoiding the risk of Service-Level Agreement (SLA) default due to model overfitting.

[0022] Finally, to compensate for the inevitable error between the theoretical calculation model and the real physical environment, the embodiments of this application construct a real-time closed-loop control mechanism that combines fluctuation slope monitoring with adaptive rollback / step enhancement. By utilizing the extremely high sensitivity of fluctuation slope (derivative) to the nonlinear avalanche effect, the lag of single numerical threshold monitoring is overcome, enabling safe step-by-step probing at the physical limits of communication SLA. Once the channel nonlinear characteristics are detected (slope change), the rollback is performed immediately at the millisecond level. Thus, under the premise of absolutely ensuring the security of communication services, the power carrying potential of the optical fiber link is maximized, and the signal-to-noise ratio and detection distance of the sensing task are significantly improved.

[0023] In the embodiments of this application, the aim is to achieve deep coexistence and intelligent operation and maintenance of sensing tasks by the following specific objectives, while ensuring zero-degradation of communication services: 1. Construct a multi-dimensional joint resource abstract model: Establish a unified abstract model covering business KPIs, optical layer physical resources and sensing task parameters. In particular, introduce an interference vector quantization mechanism to make the impact of sensing pulses on the communication side, such as power crosstalk, signal-to-noise ratio degradation and bit error rate, calculable, predictable and constrainable.

[0024] 2. Implement collaborative scheduling based on feasible region pruning: By using the task descriptor (TSD) and bidirectional slicing mechanism (C-Slice / S-Slice), feasible region automatic pruning is performed under multi-dimensional interference budget constraints to ensure that the generated scheduling scheme avoids the risk of nonlinear mutual interference from the physical layer.

[0025] 3. Establish a closed-loop self-healing mechanism with risk adaptation capabilities: Introduce step-wise enhancement and fail-safe rollback strategies to ensure that the intensity of perception tasks can be dynamically adjusted according to risk levels such as construction windows, and achieve millisecond-level adaptive disturbance reduction when the slope of business indicator fluctuations is abnormal.

[0026] 4. Form an end-to-end automated closed loop for production and operation: Establish a closed-loop orchestration process of "scheduling - execution - multi-dimensional monitoring - feedback - recalculation" and output executable actions with physical-geographic mapping relationship (such as GIS accurate mapping, operation and maintenance work orders with version numbers, etc.) to improve the auditability of live network operation and maintenance.

[0027] like Figure 2 As shown in the embodiments of this application, the overall software architecture can be abstracted as a "five-layer logical architecture", which consists of the following layers from bottom to top: state acquisition layer, resource abstraction layer, collaborative orchestration layer, execution and feedback layer, and visualization decision layer.

[0028] The system is divided into three layers: the status acquisition layer, the optical layer KPI, and the status of sensing devices; the resource abstraction layer, which establishes a unified resource model; the collaborative orchestration layer, which generates TSD and calculates S-Slice / C-Slice; the execution and feedback layer, which is responsible for issuing configurations, rolling back, and evaluating the effects; and the visualization decision-making layer, which is responsible for linking the network management, GIS, and work order systems.

[0029] The system should include at least the following functional modules: Service SLA parsing and service orchestration module: Parses service priorities, SLA metrics, protection levels, and scheduling constraints.

[0030] Optical layer resource abstraction module: abstracts WDM / OTN / ROADM resources, wavelength / time slot occupancy, power budget and OSNR margin, and provides a programmable interface.

[0031] Perception Task Management Module: Generates TSD, maintains task queues, and manages priorities and triggering conditions (baseline inspection / risk trigger / manual instructions).

[0032] Collaborative scheduling and slice generation engine: Solve for S-Slice and necessary C-Slice adjustment schemes under business constraints.

[0033] Execution control and rollback module: Distributes configurations to sensing and network devices, records versions, and supports rollback in case of anomalies.

[0034] Monitoring and Quality Assessment Module: Continuously monitor business KPIs, optical layer KPIs, and perception KPIs (such as confidence level, coverage, and effective scan rate).

[0035] Visualization / Network Management / GIS / Work Order Interface Module: Maps mileage segments, events, and tiling strategies to GIS and work order fields to achieve an executable closed loop.

[0036] The method in this embodiment is as follows: In step S110, the basic status information of the integrated sensing and communication optical network is obtained. The basic status information includes the service level agreement indicators of communication services, the status of optical layer physical resources, and the requirements of sensing tasks to be performed.

[0037] In one embodiment of this application, by interacting with the northbound interface of the software-defined networking (SDN) controller and the optical layer network management system, the current physical and logical state of the network can be fully perceived, and the set of information obtained is called basic state information.

[0038] Service Level Agreement (SLA) metrics for communication services are the safety baselines that must be strictly adhered to when performing any sensing operations. In this embodiment, SLA metrics not only refer to traditional bandwidth commitments, but also focus on quantitative constraints on the quality of optical signal transmission, specifically including bit error rate (BER) thresholds, optical signal-to-noise ratio (OSNR) baselines, and service priority indicators.

[0039] The Bit Error Rate (BER) threshold refers to the maximum percentage of erroneous bits allowed in a service signal during transmission. For high-priority services such as financial leased lines, the BER tolerance is extremely low (e.g., less than 10 to the power of -12), which means that extreme caution must be exercised when applying sensing and probing signals to links hosting such services to prevent interference.

[0040] The optical signal-to-noise ratio (OSNR) baseline refers to the minimum SNR level required for an optical receiver to correctly demodulate a signal. This metric directly determines the signal's ability to resist noise interference and is a key benchmark for subsequent calculations of dynamic safety margin.

[0041] Service priority identifiers are used to distinguish the importance of different service flows. When network resources are strained or interference conflicts are unavoidable, service level agreements (SLAs) for high-priority services will be prioritized based on this identifier, while low-priority services will be strategically downgraded or rerouted.

[0042] The physical resource status of the optical layer is a digital mapping of the underlying optical network hardware environment, i.e., a digital twin view. This status information describes in detail the physical medium and channel environment for optical signal transmission, specifically including network topology and routing information, wavelength channel occupancy, fiber optic link physical parameters, and real-time optical power distribution. Network topology and routing information describes the connection relationships of various network element nodes (such as ROADM nodes and amplifier sites) in the optical network, as well as the specific optical cable segments and port locations traversed by the currently activated optical paths.

[0043] The wavelength channel occupancy record shows which wavelengths (frequency slots) within the C-band or L-band are occupied by communication services and which wavelengths are idle. This is to find a legitimate transmission frequency point for the sensing pulse in subsequent steps and avoid directly occupying communication wavelengths.

[0044] The physical parameters of the fiber optic link are the most crucial data for calculating nonlinear interference. These include the type of fiber (such as G.652 standard single-mode fiber or G.655 non-zero dispersion-shifted fiber), the length of the fiber span, the dispersion coefficient, and the nonlinear coefficient. Because different types of fibers have varying sensitivities to nonlinear effects (such as cross-phase modulation) generated by high-power signals, these parameters must be used to calibrate the interference calculation model.

[0045] Real-time optical power distribution obtains the total power of existing optical signals in each transmission segment. This is because the nonlinear effect of optical fiber has an exponential relationship with the total power entering the fiber. Only by knowing the current background power can we accurately assess whether adding a new sensing pulse will trigger the nonlinear threshold.

[0046] Requirements for the perception task to be performed The perception task requirements are the "trigger inputs" driving this collaborative orchestration, that is, what the user or upper-layer application wants to perceive through the optical network. Specifically, this includes the description of the perception target, the perception type and accuracy requirements, and the expected confidence level. The target description specifies the physical object that needs to be monitored, such as a specific segment of optical cable link (corresponding to a segment of pipe gallery or submarine cable in geography).

[0047] The sensing type and accuracy requirements must be clearly defined to determine whether the task belongs to distributed acoustic sensing (DAS), distributed temperature sensing (DTS), or fiber optic condition monitoring. Simultaneously, its specific requirements for spatial resolution (e.g., accurate to 5 meters or 10 meters) and sampling frequency (e.g., 1 kHz or 10 kHz) must be determined.

[0048] Confidence expectation refers to the accuracy requirements of the task for the perceived results. High confidence usually means that higher transmit power or longer pulse-coded sequences are needed, which is often accompanied by stronger signal interference capability, and is therefore the core basis for cost-benefit trade-offs.

[0049] By acquiring the service level agreement indicators, optical layer physical resource status, and sensing task requirements of the aforementioned communication services, a complete holographic view was constructed, including constraints (SLA), environmental boundaries (physical resources), and driving objectives (sensing requirements). This enables subsequent collaborative orchestration models to no longer blindly allocate resources, but rather to perform precise mathematical deduction and optimization within a defined solution space with known physical characteristics and service baselines.

[0050] In step S120, a perception task descriptor containing task priority and maximum interference budget is generated according to the perception task requirements.

[0051] In embodiments of this application, a perceptual task descriptor (TSD) is used to structurally describe object 1 of the perceptual task. It is defined as a vector. Among them, the newly added items This is the maximum allowable disturbance budget vector for this task.

[0052] The perception task descriptor is a standardized data packet or digital file used in this application to carry and transfer the core parameters of the perception task. It is equivalent to the digital identity of the perception task in the collaborative orchestration system.

[0053] The perceptual task descriptor not only contains the basic attributes of the task (such as detection duration and detection area), but also encapsulates two key constraint fields that guide the system in resource scheduling: task priority and maximum interference budget. By generating perceptual task descriptors, unstructured user requirements are transformed into structured data objects that can be read, compared, and computed by the algorithm.

[0054] Task priority determines the order of the sensing task in the scheduling queue and its preemption rights over other services when network resources (such as optical power and spectrum resources) are insufficient or conflicts occur. Priority labels are generated based on the urgency of the task and the type of service, using a lookup table or logical mapping. High Priority (Critical Assurance Level): Corresponds to security tasks with extremely high real-time requirements, such as fiber optic break detection and physical intrusion alarms. Descriptors generated for these tasks have preemptive properties, allowing for the compression of resources for other lower-priority tasks when necessary, and even temporarily squeezing out a small amount of communication redundancy within a safe range to ensure the immediacy of detection.

[0055] Medium priority (performance maintenance level): Corresponds to routine maintenance tasks such as fiber optic stress analysis and health checks. These tasks are executed when resources are plentiful and queued when resources are scarce.

[0056] Low priority (best-effort level): Corresponds to long-term statistical tasks with low accuracy requirements or low time sensitivity. These tasks are executed by filling in gaps using only fragmented remaining network resources (such as idle time slots).

[0057] The maximum interference budget defines the behavioral boundary of the sensing task. The maximum interference budget is not the power budget of the sensing task itself, but rather the maximum amount of additional interference that the communication service can tolerate. Its generation logic is based on the principle of residual tolerance. The maximum interference budget is the portion of the communication signal that is allowed to be consumed by the sensing task during transmission, where the maximum interference budget quality metric (such as optical signal-to-noise ratio, OSNR) is typically higher than the minimum threshold for an interruption.

[0058] The system reads the current real-time optical signal-to-noise ratio (OSNR) and bit error rate (BER) of the communication service and compares them with the minimum alarm threshold specified in the SLA agreement. The difference between the two is then converted into the maximum interference budget for the sensing task after being adjusted by a safety factor (e.g., reserving 20% ​​as a buffer).

[0059] The maximum interference budget is a multi-dimensional constraint. For example, it stipulates that the increase in bit error rate caused by the sensing task cannot exceed 10 to the power of -8, or the decrease in optical signal-to-noise ratio caused by it cannot exceed 0.5 dB.

[0060] The maximum interference budget is written into the perception task descriptor and serves as the ceiling for subsequent feasible domain pruning steps. Any perception task under any configuration that predicts interference values ​​exceeding the maximum interference budget will be deemed invalid and discarded directly.

[0061] In summary, the process of generating a perception task descriptor is essentially a process of qualitative classification (determining priority) and boundary delineation (calculating interference budget). It ensures that the perception task clearly defines its identity and behavioral boundaries before entering a complex orchestration model, thus laying the logical foundation for subsequent algorithms to find the optimal solution while protecting the security of communication services.

[0062] In step S130, a multidimensional interference vector is determined based on the basic state information and the sensing task descriptor. The multidimensional interference vector is used to quantify the nonlinear physical layer impact of the detection signal generated by the sensing task on the communication service signal.

[0063] In this embodiment, due to the complex physical mechanisms of fiber nonlinear effects (especially cross-phase modulation XPM), direct numerical calculations are not only time-consuming but also difficult to perform in real time. The specific implementation of this application is as follows: It is necessary to eliminate the data differences caused by heterogeneous underlying hardware, perform multi-dimensional resource abstraction on the physical resource status of the optical layer, and build a standardized unified resource view.

[0064] Multidimensional resource abstraction refers to mapping scattered physical data (such as fiber type, span length, amplifier gain) into logical objects that can be read by algorithms.

[0065] The unified resource view is a global digital map that includes the following key dimensions: Network topology layer: Dynamically displays the node connections of the optical network and the route of the optical cable links.

[0066] Wavelength channel status layer: Marks the occupancy status (idle, occupied, or protected) of each wavelength grid in the C-band or L-band, clarifying the available spectral gaps for the sensing signal.

[0067] Power budget layer: records the remaining gain capability of the optical amplifier in the current link and the upper limit of the nonlinear power threshold of the optical fiber.

[0068] Optical Signal-to-Noise Ratio (OSNR) Margin: This is a key dynamic parameter that represents the difference between the actual OSNR value of the current communication signal and the minimum OSNR threshold required for demodulation errors to occur. The larger this margin, the stronger the signal's resistance to interference.

[0069] Based on the unified resource view, the modulation format used by the communication service is identified, and different modulation formats have drastically different resistance to interference.

[0070] The nonlinear sensitivity coefficient is used to quantify the efficiency of external interference of unit strength in converting into damage to the quality of communication signals.

[0071] Mapping relationship analysis is used to analyze the modulation format of services in the current transmission link. For example, for Quadrature Phase Shift Keying (QPSK) modulation, the constellation point spacing is relatively large, making it insensitive to phase noise, thus its nonlinear sensitivity coefficient is low. However, for Hexadecimal Quadrature Amplitude Modulation (16QAM) or higher levels of modulation, the constellation diagram is extremely dense, and even small phase jitter can lead to decision errors, thus its nonlinear sensitivity coefficient is extremely high. Using this coefficient, the collaborative orchestration model can differentiate the anti-interference capabilities of different services.

[0072] A collaborative orchestration model is constructed based on a unified resource view and nonlinear sensitivity coefficients. This model is used as the logical computation engine for simulating and predicting optical layer physical interference.

[0073] The collaborative orchestration model uses key parameters in the perception task descriptor as driving inputs, mainly including transmit power parameters (which determine the intensity of the interference) and duty cycle parameters (which determine the duration of the interference).

[0074] The collaborative orchestration model uses nonlinear sensitivity coefficients as core weights.

[0075] The internal logic of the cooperative orchestration model simulates the nonlinear coupling process of optical signals during transmission. Specifically, based on the input variable (sensor signal strength) and the physical parameters of the link, the cooperative orchestration model calculates the refractive index disturbance generated in the transmission optical fiber, multiplies it by a sensitivity coefficient, and then derives the specific damage value transmitted by the disturbance to the communication signal. This process is performed entirely at the mathematical logic level and does not require destructive testing on the physical network.

[0076] Finally, the specific parameters of the perception task to be performed are substituted into the above collaborative orchestration model to calculate the multidimensional interference vector.

[0077] Interference Vector ): Used to quantify the multi-dimensional impact of sensing tasks on communication services. Defined as:

[0078] in, Represents instantaneous power crosstalk. This represents the degradation value of the optical signal-to-noise ratio. This represents the predicted increment of the bit error rate.

[0079] A multidimensional interference vector is not a single numerical value, but a dataset describing the interference characteristics, containing at least the following three dimensions: Instantaneous power crosstalk: Characterizes the amplitude of instantaneous power fluctuations that occur when the high energy of a sensing pulse is coupled to the communication wavelength through nonlinear effects. It reflects the severity of energy transfer at the physical layer.

[0080] Optical Signal-to-Noise Ratio (OSNR) Degradation Prediction: This represents the specific numerical decrease in OSNR (e.g., a decrease of 0.5 dB) caused by the increase in background noise level in the communication channel due to the introduction of sensing signals. It is a direct indicator for assessing signal quality degradation.

[0081] Predicted Bit Error Rate (BER) Increment: This is the final operational layer metric. The multidimensional interference vector calculates the expected increase in the bit error rate based on the predicted optical signal-to-noise ratio (OSNR) degradation and the specific coding error correction capability.

[0082] This embodiment transforms the abstract sensing task into a concrete, quantifiable, multi-dimensional interference vector. This allows subsequent scheduling algorithms to clearly understand how much the OSNR of the communication service will decrease and how much the bit error rate will increase if this sensing task is executed, thus providing precise data support for subsequent feasible region pruning and multi-objective optimization.

[0083] In step S140, feasible domain pruning is performed on the combination of configuration parameters for the sensing task based on the optical layer physical resource status and multi-dimensional interference vectors to generate a candidate configuration set that meets the service level protocol indicators of the communication service.

[0084] In the embodiments of this application, after generating the perception task descriptor (containing the maximum interference budget) and constructing the multi-dimensional interference vector, the collaborative orchestration model faces a huge parameter solution space. To improve optimization efficiency and ensure absolute communication security, this embodiment introduces a pre-screening mechanism based on dynamic security margin, namely feasible domain pruning. This step aims to physically eliminate high-risk configuration combinations, retaining only compliant candidate sets.

[0085] The collection and quantification of historical fluctuation characteristics require an assessment of the "stability" of the current optical network environment. The more unstable the environment, the higher the risk of nonlinear effects.

[0086] Optical power sampling sequence: Optical power values ​​at each monitoring point in the optical link are collected at a millisecond frequency according to a preset time window (e.g., the past 15 minutes), forming a data sequence that changes over time. This sequence faithfully records the recent jitter of the optical signal.

[0087] Statistical variance: The statistical variance is obtained by performing statistical calculations on the above optical power sampling sequence. In this embodiment, the statistical variance is not merely a mathematical indicator; it is given a clear physical meaning—a characteristic value that characterizes the level of nonlinear noise in the link environment.

[0088] When the statistical variance is small, it indicates that the link is in a stable state and the channel characteristics are relatively linear. When the statistical variance is large, it indicates that there are rapid power transients or polarization state rotations in the link, which indicates that the channel is in a nonlinear sensitive region and any new small interference may be amplified.

[0089] To address the uncertainties of the aforementioned environment, the construction and calculation of the Dynamic Safety Margin no longer uses a fixed protection threshold, but instead calculates a dynamically changing safety margin.

[0090] Construction of the dynamic function: A monotonically increasing function with statistical variance as the independent variable is constructed. The core logic of this function is risk hedging. Based on this dynamic function, the dynamic safety margin is calculated. Its core logic is: the larger the input statistical variance (the more chaotic the environment), the larger the calculated dynamic safety margin; conversely, the smaller it is.

[0091] Compressing available budget space: The physical function of dynamic safety margin is to encroach on the budget space originally belonging to the perception task. If we consider the total allowable disturbances as a container, the dynamic safety margin is the gap that must be reserved. The harsher the environment, the larger the reserved gap, and the more the actual operational space available to the perception task is compressed. This reflects the intelligent strategy of the system to automatically converge and act conservatively in harsh environments.

[0092] Define cumulative disturbance constraints. After determining the size of the safety cushion, define an insurmountable red line, which is the cumulative disturbance constraint.

[0093] At any given moment, the multidimensional interference vectors generated by all currently activated sensing tasks are vector-superimposed with the interference vector of the proposed new task to obtain the cumulative interference value. This condition stipulates that the sum of the cumulative interference value and the dynamic safety margin must be strictly less than or equal to the maximum interference budget generated in the preceding steps. This means that sensing tasks must not only ensure that the interference they generate does not exceed the SLA limit, but also pay an additional safety tax determined by current environmental fluctuations (i.e., the dynamic safety margin). Only the space remaining after paying this tax represents the physical resources that the sensing task can truly utilize.

[0094] Perform specific filtering actions to generate a candidate configuration set.

[0095] Traversing potential parameter combinations: The configuration parameters of the sensing task (such as transmit power from 1mW to 10mW, duty cycle from 1% to 10%) constitute a vast space of parameter combinations. These combinations are scanned one by one at the logical level. For each parameter combination, its interference value is calculated using the interference model and verified using the aforementioned cumulative interference constraints. Any combination that results in "interference + margin > budget" is judged as "physically infeasible" or "high-risk," and the system directly removes it from the solution space.

[0096] Generating a candidate configuration set: After this rigorous selection process, all remaining parameter combinations that meet the constraints constitute the candidate configuration set. Each configuration in this set is theoretically safe and meets the Service Level Agreement (SLA) requirements of the communication service, thus providing a safe search range for the subsequent "optimization" step.

[0097] In this embodiment, statistical variance is introduced to perceive the environment, dynamic safety margins are used to mitigate risks, and pruning is employed to narrow down the scope. This coherent logic ensures that the system does not tread the edge of danger, achieving a technological leap from blind search to optimization within safe boundaries.

[0098] In step S150, the optimal communication slicing strategy and perception slicing strategy are obtained by solving the candidate configuration set with the joint optimization objectives of minimizing the service level agreement deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks.

[0099] In this embodiment, after obtaining a set of safe candidate configurations through feasible region pruning, the challenge is to select the most cost-effective option from these safe choices. To address this, this embodiment constructs a complex multi-objective cooperative scheduling function that unifies communication quality, resource consumption, and perceived benefits within a single mathematical framework for weighing these factors. The specific implementation process and the meanings of each factor are explained below: The architecture of the multi-objective cooperative scheduling function is the core carrier of the decision-making logic of this invention. It consists of three mutually interacting sub-items: a Service Level Agreement (SLA) offset penalty term, a network resource cost term, and a perceived confidence gain term. The design goal of this function is to find an equilibrium point that minimizes the total value (i.e., minimizes cost and maximizes gain).

[0100] The service SLA offset penalty term (risk control dimension) aims to impose a very high mathematical penalty on behaviors that may jeopardize communication quality, forcing the algorithm away from the SLA red line. It calculates the difference between the predicted performance metric (such as bit error rate BER) of the communication service under a certain configuration and the benchmark metric specified in the SLA protocol, i.e., the offset.

[0101] This embodiment employs a non-linear penalty strategy. Specifically, when the prediction performance metric is better than the SLA metric, the value of this penalty term is very small or even close to zero; however, once the prediction performance metric approaches or falls below the SLA metric, the function value of this term increases exponentially. This design simulates the cliff effect, meaning that if the sensing task slightly touches the security threshold of the communication service, the total cost of the entire scheduling function will become infinitely large, thus directly deterring such dangerous configuration attempts at the algorithm level and ensuring that communication security has the highest weight.

[0102] The network resource cost component (operating cost dimension) aims to quantify the occupancy of valuable optical network resources by sensing tasks, advocating for resource-efficient scheduling. It primarily considers two types of scarce resources: wavelength bandwidth resources, i.e., the spectrum width occupied by the sensing signal; and time slot resources, i.e., the execution time of the sensing task. Based on the current network load, different weighting coefficients are assigned to these two types of resources (e.g., spectrum is more expensive during peak periods, so the weighting is higher), and their weighted sum is calculated. The purpose of the weighted sum is to suppress unnecessary resource waste. When sensing effects are similar, the algorithm will automatically tend to choose the configuration scheme with narrower spectrum occupation and shorter execution time, thereby reducing operating costs.

[0103] The perceived confidence benefit term (business value dimension) represents the positive value brought about by performing the perceived task. Since the overall goal is to minimize the function value, this embodiment typically includes the negative value of the benefit in the overall function or treats it as the denominator.

[0104] The benefits are primarily measured by the spatial resolution (how much detail you see) and sampling rate (how fast you see) of the perception task.

[0105] The logarithmic function form and the diminishing marginal returns effect are crucial technical features. This embodiment uses the logarithmic function form to describe the returns. Its physical meaning is that in the initial stage, investing a small amount of resources to improve resolution yields enormous benefits; however, when the resolution is already high, investing a large amount of resources to pursue even the smallest improvements results in slow growth in actual value (i.e., diminishing marginal returns).

[0106] The logarithmic function prevents the system from blindly pursuing excessive precision. It guides the algorithm to find a reasonable stopping point between sufficiently good perception and excessive resource input, which conforms to the cost-effectiveness principle in practical engineering.

[0107] After constructing a multi-objective cooperative scheduling function, a heuristic search algorithm (such as a genetic algorithm or particle swarm optimization) is used to optimize the candidate configuration set. The algorithm iteratively tries different parameter combinations and calculates the corresponding multi-objective scheduling function values. Its goal is to find the parameter combination that minimizes SLA penalties, moderates resource costs, and provides reasonable perceived benefits—that is, the point that minimizes the total function value. The parameter combination finally found is confirmed as the optimal communication slicing strategy and perception slicing strategy. It includes not only the execution parameters of the perception task, such as transmit power, frequency, and duration, but also the fine-tuning strategies that the communication service may need to adapt to the task (such as enabling guard slots).

[0108] In the embodiments of this application, through three cleverly designed function terms, a complex game is mathematically realized, involving a veto for communication security, careful calculation of resource usage, and knowing when to stop for perceived benefits. The final output is a globally optimal solution that takes into account the interests of all parties.

[0109] In S160, execution instructions are issued to the underlying network devices based on the optimal communication slicing strategy and the perception slicing strategy to obtain the fluctuation slope of the communication service performance indicators. When the fluctuation slope exceeds the preset safety threshold, the network configuration is restored to the previous stable version. When the fluctuation slope does not exceed the safety threshold and the perception confidence does not meet the standard, a step-by-step enhancement strategy is executed to adjust the perception slicing strategy.

[0110] In this embodiment, the theoretically optimal communication slicing strategy and sensing slicing strategy were solved. However, considering the complexity and unpredictability of the fiber optic channel environment (such as sudden external vibrations or instantaneous polarization mode dispersion), it is not simply a matter of transmitting and then stopping; a millisecond-level real-time closed-loop feedback mechanism must be established. This embodiment further explores the potential of sensing performance by combining tentative enhancement with slope-level monitoring, while ensuring absolute communication security. The specific implementation process and terminology are as follows: After the initial strategy was implemented, it was not satisfied with the status quo, but tried to explore whether there was room to further improve the quality of perception.

[0111] First, read the real-time optical signal-to-noise ratio of the current communication service and calculate the difference between it and the minimum SLA requirement, which is the remaining optical signal-to-noise ratio margin.

[0112] The preset enhancement threshold is a safety judgment value. It determines whether the remaining optical signal-to-noise ratio margin is greater than the preset enhancement threshold. The enhancement mechanism is only allowed to be activated when the quality of the communication service is very good and far exceeds the safety threshold; otherwise, the current state is maintained and no risk is taken.

[0113] Once sufficient safety margin is confirmed, a step-by-step probing strategy will be implemented, aiming to gradually approach the physical limits in very small increments.

[0114] The preset power / duty cycle step value is the defined minimum adjustment unit. For example, the power step value might be set to 0.1 dBm, and the duty cycle step value to 0.1%.

[0115] Building upon the current strategy, the transmit power of the sensing task is increased by adding a power step value, or the pulse duration is increased by adding a duty cycle step value. This fine-tuned strategy is called the enhanced sensing slicing strategy. The enhanced sensing slicing strategy is then distributed to the underlying network devices, enabling the sensing hardware to operate according to the enhanced parameters.

[0116] Within a very short period after the enhancement strategy takes effect (i.e., the preset monitoring time), the underlying network devices enter a highly sensitive monitoring state. Error rate data from the communication receiver is captured frequently, forming a dense time series.

[0117] The slope of the bit error rate (BER) data over time is calculated to obtain the fluctuation slope. The fluctuation slope represents the acceleration or rate of BER degradation. Compared to directly looking at the absolute value of the BER (which often has a lag), the fluctuation slope can more sensitively reflect the trend of the problem. If the fluctuation slope is gentle, it indicates that the impact is controllable; if the fluctuation slope is steep, it indicates that the system is undergoing severe quality degradation.

[0118] The calculated slope is used to diagnose the physical condition of the fiber optic channel.

[0119] The preset nonlinear inflection point threshold is a critical indicator characterizing the entry of optical fiber into the nonlinear avalanche effect. When the optical power exceeds a certain limit, nonlinear interference (such as cross-phase modulation) will cause the bit error rate to no longer increase linearly, but will exhibit an exponential burst. The characteristic of this burst point is that the fluctuation slope is extremely high.

[0120] The absolute value of the calculated fluctuation slope is compared with the nonlinear inflection point threshold.

[0121] Scenario A (Safe): If the absolute value of the fluctuation slope is less than the nonlinear inflection point threshold, it indicates that although the current enhanced sensing slicing strategy introduces a small amount of interference, the channel is still in the linear or quasi-linear region, and communication services are safe. This enhancement configuration can be retained, and even the next round of improvements can be implemented.

[0122] Scenario B (Dangerous): If the absolute value of the fluctuation slope exceeds the nonlinear inflection point threshold, the communication channel is determined to have entered the nonlinear region. This means that interference is spreading avalanche-like, and if this state continues, communication will be interrupted within milliseconds. Once the dangerous region is determined, circuit breaker measures must be taken immediately to terminate the step-up enhancement strategy, immediately stop any further attempts to increase power or duty cycle, execute the adaptive rollback mechanism, and immediately send a command to force the network device's configuration parameters to be restored to the previous stable version (i.e., the previous successful step or the initial optimal strategy state).

[0123] The rollback operation eliminates the excessive interference that was just introduced, bringing the optical network back to a safe operating range. This mechanism ensures that even when the system is testing its physical limits, it can reverse dangerous operations before users perceive any service disruption, thus achieving exploratory optimization with zero SLA breach.

[0124] In this embodiment, a complete closed-loop logic is employed: cautious probing (stepping) – keen observation (slope) – immediate retreat upon encountering danger (rollback). It utilizes the extremely high sensitivity of the fluctuation slope to nonlinear effects, overcomes the lag of traditional monitoring methods, and enables the integrated sensing system to safely operate at the edge of physical limits, maximizing sensing performance.

[0125] In one embodiment of this application, determining the multidimensional interference vector based on basic state information and a perception task descriptor includes: Based on basic state information and sensing task descriptors, a collaborative orchestration model for communication services and sensing tasks is constructed; the collaborative orchestration model is calculated to determine the multidimensional interference vector.

[0126] In one embodiment of this application, how can we predict the specific interference that a sensing task may cause without actually disrupting communication services? This embodiment achieves this goal by constructing a collaborative orchestration model and solving for multi-dimensional interference vectors.

[0127] To eliminate the differences between heterogeneous optical network devices, the acquired physical resource status of the optical layer is first standardized to establish a global digital foundation, namely a unified resource view.

[0128] Multidimensional resource abstraction refers to the process of extracting physical parameters (such as fiber type, span length, amplifier gain, etc.) that are scattered in different network elements and equipment from different manufacturers, and mapping them into logical objects that can be uniformly called by algorithms.

[0129] The Unified Resource View is not a simple database, but a dynamically updated holographic map that mainly includes network topology, wavelength channel status, power budget, and optical signal-to-noise ratio (OSNR) margin. Network topology describes the node connections and physical routing of optical fiber links, establishing the basis for interference propagation paths. Wavelength channel status clearly marks the occupancy of each wavelength grid within the C-band or L-band, distinguishing between the communication-occupied spectrum and the sensing-available spectrum. Power budget records the remaining gain capacity of the optical amplifiers in the current link and the upper limit of the input power that the fiber can withstand. Optical signal-to-noise ratio (OSNR) margin is a key dynamic indicator, representing the margin by which the actual OSNR value of the current communication signal exceeds the minimum threshold required for demodulation errors. The larger this margin, the stronger the link's resilience.

[0130] After establishing a unified resource view, it is necessary to set conversion rules between physical interference and signal quality. This embodiment introduces a nonlinear sensitivity coefficient to identify the signal modulation scheme (such as QPSK, 16QAM, or 64QAM) used in the current communication service. As is well known, the higher the modulation order, the closer the distance between symbols in the constellation diagram, and the more sensitive it is to phase noise and amplitude jitter.

[0131] The non-linear sensitivity coefficient is a conversion parameter used to characterize susceptibility. It quantifies the efficiency with which a unit intensity of external interference energy (such as cross-phase modulation caused by a sensing pulse) is converted into communication signal quality degradation (such as an increase in the EVM value).

[0132] The nonlinear sensitivity coefficient establishes a function mapping: for low-order modulations such as QPSK, the nonlinear sensitivity coefficient is small, meaning that the sensing pulse can be appropriately amplified; for high-order modulations such as 16QAM, the nonlinear sensitivity coefficient is extremely large, meaning that even a very small sensing pulse may cause communication failure. Through the nonlinear sensitivity coefficient, the collaborative orchestration model can assess the impact of interference in a context-specific manner.

[0133] Based on the unified resource view and nonlinear sensitivity coefficients described above, a collaborative orchestration model is constructed at the logical layer. This model is a mathematical engine for calculating physical interference at the optical layer.

[0134] The collaborative orchestration model directly calls key parameters from the sensing task descriptor generated in the preceding steps, specifically including transmit power parameters (determining the strength of the interference source) and duty cycle parameters (determining the temporal density of the interference source). The collaborative orchestration model loads nonlinear sensitivity coefficients as core weights to transform the physical layer's energy parameters into the signal layer's quality parameters. Essentially, the collaborative orchestration model simulates the nonlinear coupling process of optical signals in optical fibers. Under the topology and physical parameter constraints provided by the unified resource view, it calculates the specific impact of the input variable (sensing signal) on the communication wavelength after transformation by the nonlinear sensitivity coefficients. Substituting the specific parameters of the sensing task to be executed into the collaborative orchestration model for calculation, the output is not a single numerical value, but a multi-dimensional interference vector describing the overall interference situation.

[0135] Multidimensional interference vectors are used to comprehensively characterize the datasets that may have a potential negative impact on communication services from sensing tasks, and must include predicted values ​​in at least the following three dimensions: Instantaneous power crosstalk: Characterizes the amplitude of instantaneous power fluctuations in the communication signal caused by the high energy of the sensing pulse coupled to the communication wavelength through fiber nonlinear effects (such as cross-phase modulation XPM). This is a direct interference indicator of the physical layer.

[0136] Optical Signal-to-Noise Ratio (OSNR) degradation prediction: This represents the expected decrease in the OSNR at the receiver of the communication channel due to the introduction of the sensed signal and the associated noise floor increase (e.g., a predicted decrease of 0.5 dB). This is a quality degradation indicator at the link layer.

[0137] Predicted Bit Error Rate (BER) Increment: This is the most critical service layer metric. The model calculates the expected increase in the bit error rate based on the OSNR degradation value and the coding error correction capability of the communication service.

[0138] This embodiment transforms abstract perception requirements into concrete, quantifiable, multi-dimensional interference vectors. This allows subsequent scheduling algorithms to clearly predict how much communication quality will be degraded when the task is executed, thus providing precise digital basis for feasible region pruning and multi-objective optimization, achieving a technological leap from empirical judgment to quantitative calculation.

[0139] In one embodiment of this application, the sensing task descriptor includes transmit power parameters and duty cycle parameters; based on basic state information and the sensing task descriptor, a collaborative orchestration model for communication services and sensing tasks is constructed; the collaborative orchestration model is calculated to determine the multi-dimensional interference vector, including: Multidimensional resource abstraction is performed on the optical layer physical resource status of basic state information to establish a unified resource view that includes network topology, wavelength channel status, power budget and optical signal-to-noise ratio margin. Based on the unified resource view, the modulation format of the transmission link where the communication service is located is analyzed, and a nonlinear sensitivity coefficient is defined according to the modulation format. The nonlinear sensitivity coefficient is used to characterize the mapping relationship between the sensing pulse intensity and the degradation of the communication signal quality. In the unified resource view, a collaborative orchestration model is constructed using the transmit power parameter and duty cycle parameter in the perception task descriptor as input variables and the nonlinear sensitivity coefficient as the transformation parameter. The parameters of the sensing task to be performed are substituted into the collaborative orchestration model to obtain a multidimensional interference vector. The multidimensional interference vector includes at least the instantaneous power crosstalk value, the optical signal-to-noise ratio degradation prediction value, and the bit error rate increment prediction value.

[0140] In integrated optical networks, sensing signals (typically high-power pulses) and communication signals (typically continuous waves) are transmitted along the same fiber, which can easily induce fiber nonlinear effects (especially cross-phase modulation, XPM). To achieve precise interference control, this embodiment no longer relies on fuzzy empirical values, but instead constructs a physical layer-based collaborative orchestration model. By analyzing the entire network resources and introducing nonlinear sensitivity coefficients, specific multidimensional interference vectors are calculated.

[0141] Optical networks suffer from severe heterogeneity in their underlying devices and complex physical parameters. To provide standardized input for upper-layer algorithms, a multi-dimensional resource abstraction process is first performed to establish a unified resource view.

[0142] Network topology includes not only node connections but also physical attribute data of fiber optic links, such as fiber type (G.652 / G.655) and span length L. span And the dispersion coefficient D. These parameters determine the cumulative distance of nonlinear disturbances and the walk-off effect.

[0143] Wavelength channel status: Records the current occupancy status of the spectrum grid. The system identifies the communication wavelength λ. C With the intended sensing wavelength λ S The frequency domain spacing Δf = |λc-λs| between channels directly determines the efficiency of nonlinear crosstalk between channels.

[0144] Power budget and OSNR margin: This is a dynamic ledger that records the remaining gain capacity of the optical amplifiers (EDFAs) in the current link, as well as the current actual optical signal-to-noise ratio (OSNR) of the communication signal. curr ) higher than its minimum SLA threshold (OSNR) th (The remaining amount).

[0145] The tolerance of communication services to interference depends on their modulation format. This embodiment introduces a nonlinear sensitivity coefficient η to quantify the sensitivity of different modulation formats to external phase noise. The high power of the sensing pulse, through the Kerr effect in the optical fiber, causes a transient change in the refractive index, which in turn leads to a phase shift in the communication optical signal.

[0146] Modulation format parsing: Reads the modulation format of the communication service (such as QPSK, 16QAM, 64QAM).

[0147] QPSK: It has only 4 constellation points, a large phase interval (90 degrees), and strong anti-interference ability.

[0148] 16QAM / 64QAM: The constellation points are densely packed, and the phase intervals are extremely small. Even a tiny phase jitter can lead to incorrect judgments.

[0149] The nonlinear sensitivity coefficient η is defined as a function of the modulation order M. In this embodiment, the mapping relationship can be expressed as:

[0150] Where M is the modulation order (e.g., M=16 for 16QAM). and This is a constant preset based on the fiber nonlinearity coefficient η. This mapping relationship shows that the higher the modulation order, the larger the value of η, which means that the signal quality degradation caused by per unit intensity of perceived interference is more severe. This provides a mathematical basis for the collaborative orchestration model to treat different services differently.

[0151] The collaborative orchestration model is a computational engine whose core task is to transform input sensing parameters into communication impairment indicators. This model uses a unified resource view as the environmental constraint, sensing task descriptors as input, and nonlinear sensitivity coefficients as the core of the transformation.

[0152] Input variables: Sensing transmit power (P) s ): determines the upper limit of the intensity of nonlinear disturbance.

[0153] Duty cycle parameter (D) cycle The duty cycle determines the density of interference in the time domain. Sensing signals are usually in pulse form; the lower the duty cycle, the smaller the average interference, but instantaneous impulses still exist.

[0154] The collaborative orchestration model is constructed based on cross-phase modulation (XPM) theory. The nonlinear phase shift variance generated by the sensing signal... It is proportional to the square of the sensing power. The cooperative orchestration model introduces a walk-off effect factor H(Δf) to describe the weakening of the interaction when the sensing pulse slips over the communication pulse due to the different group velocities.

[0155] By substituting the parameters of the sensing task to be performed into the collaborative orchestration model, a multidimensional interference vector containing three key components is calculated. : First dimension: Instantaneous power crosstalk value X XPM This is direct interference at the physical layer. The cooperative orchestration model calculates the noise power coupled to the communication wavelength by the sensed pulse through the XPM effect. Its calculation logic can be described as follows:

[0156] Where γ is the fiber nonlinearity coefficient, L eff It is the effective duration. This value quantifies how much noise the sensing pulse adds to the communication channel.

[0157] Second dimension: Predicted optical signal-to-noise ratio degradation (ΔOSNR) This is a quality degradation metric for the link layer. This is due to the introduction of additional interference noise X. XPM As a result, the effective optical signal-to-noise ratio of the communication signal will inevitably decrease. The formula for calculating the degradation value is:

[0158] in, For communication signal power, This refers to the original spontaneous radiated noise. The formula for calculating the degradation value reflects: the interference X generated by the sensed power. XPM After the nonlinear sensitivity coefficient η and duty cycle D cycle After weighted correction, it is added to the system as new noise, thus lowering the overall signal-to-noise ratio.

[0159] Third dimension: Predicted bit error rate increment ΔBER By using the complementary error function erfc, the degradation of OSNR is mapped to the increase of bit error rate.

[0160]

[0161] The predicted increment of the bit error rate tells the scheduling system directly how much the bit error rate of the communication service will increase if this sensing task is performed.

[0162] This embodiment introduces a nonlinear sensitivity coefficient to identify the inherent differences in the capabilities of various communication services. For QPSK services with strong anti-interference capabilities, the cooperative orchestration model allows sensing tasks to use higher transmit power, thereby detecting targets at greater distances; while for vulnerable 16QAM services, constraints are automatically tightened. This tailored approach greatly unleashes the sensing potential of optical networks.

[0163] By calculating a multi-dimensional interference vector that includes instantaneous power crosstalk, OSNR degradation, and BER increment, the system can accurately predict the consequences before the task is issued. This means that the system no longer needs to conduct dangerous trial and error in the real network, but can intercept high-risk configurations that may lead to excessive bit error rate through model simulation in advance, thereby improving the security and reliability of the integrated sensing system to the telecom-grade standard (i.e., 99.999% reliability).

[0164] The collaborative orchestration model incorporates the calculation of walk-off effect factors and duty cycle parameters, enabling the system to leverage frequency domain spacing (wavelengths being further apart) and time domain gaps (pulse transmissions being more sparse) to gain greater power space. This multi-dimensional resource decoupling allows communication and sensing services to coexist more closely in congested optical networks, significantly improving the overall utilization rate of fiber optic spectrum resources.

[0165] In one embodiment of this application, feasible domain pruning is performed on the combination of configuration parameters for the sensing task based on the optical layer physical resource status and multi-dimensional interference vectors to generate a candidate configuration set that meets the service level protocol indicators of communication services, including: Collect optical power sampling sequences of optical layer physical resources within a preset historical time window; The statistical variance is calculated from the optical power sampling sequence and used as a feature value characterizing the nonlinear noise level of the link environment. Construct a dynamic function with statistical variance as the independent variable; The dynamic safety margin is calculated based on the dynamic function. The larger the statistical variance, the larger the dynamic safety margin, so as to compress the available disturbance budget space. Define a cumulative disturbance constraint condition, which is the sum of the cumulative value of the multidimensional disturbance vectors of all active tasks at any given time and the dynamic safety margin, and it must not exceed the maximum disturbance budget. By traversing all potential parameter combinations for the perception task and eliminating parameter combinations that do not satisfy the cumulative interference constraint, a candidate configuration set is obtained.

[0166] In this embodiment, after constructing the collaborative orchestration model and possessing the ability to calculate multi-dimensional interference vectors, a vast parameter solution space is encountered (e.g., countless combinations of transmit power from -10dBm to +10dBm and duty cycles from 1% to 20%). To improve the efficiency of subsequent optimization and ensure safety under extreme environments, this embodiment introduces a feasible region pruning step. This step utilizes the statistical characteristics of historical data to dynamically adjust the safety boundary and eliminate high-risk combinations.

[0167] Optical fiber channels are not static; they are affected by temperature changes, mechanical vibrations, polarization rotation, and other factors, resulting in a random link state. The first step is to quantify this environmental instability.

[0168] Optical Power Sampling Sequence: Defines a preset historical time window (e.g., past T). win =15 minutes). Within this window, the Optical Performance Monitoring (OPM) module collects optical power data from each monitoring point in the link at a millisecond-level sampling rate, forming a time-series dataset. .

[0169] The statistical variance σ is obtained by calculating the above sequence. 2 In this embodiment, statistical variance is given a clear physical meaning: it is a characteristic value that characterizes the level of nonlinear noise in the link environment.

[0170]

[0171] Because the nonlinear effects of optical fibers (such as the Kerr effect) are closely related to the instantaneous intensity of optical power, a large statistical variance indicates severe fluctuations in optical power and the presence of numerous spikes. These spikes are highly susceptible to triggering avalanche-like nonlinear interference. Therefore, the statistical variance directly reflects the potential risk of channel-induced nonlinear problems.

[0172] To address the aforementioned risks, fixed protective harnesses cannot be used; instead, a dynamic safety margin (M) must be established that changes with the environment. safe ).

[0173] Construct a statistical variance σ 2 This is a monotonically increasing function of the independent variable. Its basic logic is: the more chaotic the environment (the larger the variance), the wider the buffer zone (margin) should be. In this embodiment, this function can be designed as a linear or non-linear growth model:

[0174] in, That is, the standard deviation, which directly corresponds to the amplitude of power fluctuation; k is the risk factor, which is usually taken as 3 (corresponding to 3 times the standard deviation covering 99.7% of the fluctuation probability). The minimum protective noise level inherent to the equipment.

[0175] Through calculation, abstract environmental instabilities are transformed into specific power / signal-to-noise ratio subtraction values. When the statistical variance σ... 2 When M increases, safe This increases the workload, which means that the operational space left for sensing tasks must be forcibly compressed, sacrificing sensing performance in exchange for communication security.

[0176] After determining the "safety cushion" required for the current environment, the system performs the core screening action.

[0177] The cumulative disturbance constraint is an inequality criterion. It is the cumulative value of the multidimensional disturbance vectors generated by all activated sensing tasks at any given time. In addition to the current dynamic safety margin Their sum must not exceed the system's preset maximum interference budget. .

[0178] The mathematical expression is:

[0179] The left side of the cumulative disturbance constraint represents the actual disturbance plus the virtual disturbance that must be reserved to cope with fluctuations. The configuration is physically safe only when the sum of these two parts is still less than the minimum allowable limit of the SLA (maximum disturbance budget).

[0180] Iterate through all potential parameter combinations for the sensing task to be performed (e.g., the transmit power gradually increases from 1mW to 10mW in steps of 0.1mW).

[0181] Substitute each combination into the aforementioned collaborative model to calculate its interference vector. .

[0182] Substituting the interference vector into the above inequality for verification, any parameter combination that causes the inequality to fail is considered a high-risk configuration (i.e., potentially leading to communication interruption during signal jitter) and is directly removed from the solution space. The remaining parameter combinations that satisfy the conditions constitute the candidate configuration set. This set is an absolutely safe zone; any subsequent optimization algorithm only needs to search within this set to guarantee that the resulting solution is safe.

[0183] Existing technologies typically use a fixed safety margin. If this margin is too small, when the optical fiber is subjected to external vibrations (such as a subway passing by) causing power fluctuations (increased variance), the fixed margin is insufficient to offset the fluctuations, leading to a momentary interruption of communication services. This embodiment, by monitoring and statistically analyzing variance, can automatically increase the dynamic safety margin within milliseconds of environmental deterioration, tightening the sensed power in advance, thereby effectively avoiding the risk of network outages caused by "black swan" events.

[0184] The parameter space for perception tasks is extremely large (continuous variables). Directly performing multi-objective optimization across the entire space (such as using genetic algorithms) would be computationally intensive and difficult to converge. This embodiment uses a pruning step, leveraging physical constraints (cumulative interference + margin ≤ budget), to pre-eliminate the vast majority (possibly over 80%) of infeasible solutions. This allows subsequent optimization algorithms to run only on a small set of candidate configurations with good convexity, reducing decision time from seconds to milliseconds and meeting the real-time control requirements of optical networks.

[0185] During periods of environmental stability (minimal variance), the dynamic safety margin automatically decreases to near C. base This means the system will boldly allocate more interference budget to the sensing task, allowing it to operate at higher power. Compared to a conservative approach that always reserves a large margin, this embodiment can significantly improve the detection range and accuracy of the sensing task during stable network periods, achieving a balance between conservatism when necessary and aggressiveness when appropriate.

[0186] In one embodiment of this application, the optimal communication slicing strategy and perception slicing strategy are obtained by solving a candidate configuration set with the joint optimization objectives of minimizing the service level agreement deviation of communication services, minimizing network resource consumption, and maximizing the confidence of the perception task. Construct a multi-objective cooperative scheduling function, which includes a service level protocol offset penalty term, a network resource cost term, and a perceived confidence benefit term. The service level agreement offset penalty term is configured such that when the predicted performance metric of the communication service is worse than the service level agreement metric, the function value increases exponentially. The network resource cost item is configured as: the weighted sum of wavelength bandwidth resources and time slot resources occupied by the quantified sensing task; The perception confidence benefit term is configured as follows: the benefits brought by the spatial resolution and sampling rate of the perception task are described in the form of a logarithmic function to reflect the diminishing marginal effect of perception resource investment. By using a heuristic search algorithm to find the parameter combination that minimizes the value of the multi-objective cooperative scheduling function in the candidate configuration set, the optimal communication slicing strategy and perception slicing strategy are obtained.

[0187] To find the optimal balance among the three mutually constraining objectives of communication service quality, network resource consumption, and sensing task performance, this embodiment no longer relies on optimizing a single objective (such as simply pursuing the strongest sensing capability). Instead, it constructs a composite multi-objective collaborative scheduling function (J). The final slicing strategy is determined by minimizing the value of this function.

[0188] The multi-objective cooperative scheduling function J is designed as a weighted combination of three core sub-items, mathematically represented by the superposition of penalty and cost terms, and the deduction of benefit terms. Its general form is as follows:

[0189] Among them, W1, W2, and W3 are the weight coefficients of each sub-item, which are used to adjust the system's strategy preferences in different scenarios (for example, increasing W2 during congestion to emphasize resource conservation).

[0190] Service Level Agreement (SLA) Offset Penalty Item P SLA Its purpose is to ensure that no strategy choice is made at the expense of the promised quality of communication services. It is a predictive performance metric for communication services (such as the predicted bit error rate, BER). pred ) and SLA agreement metrics (such as SLA threshold BER) th A function of the difference between ().

[0191] This embodiment innovatively uses an exponential function to describe the penalty term.

[0192]

[0193] in, : SLA offset penalty term value, meaning: a component in the objective function. The larger the value, the greater the threat the strategy poses to communication services.

[0194] A: Penalty base, a preset constant (e.g., A=1), used to adjust the basic magnitude of the penalty term in the overall objective function to ensure that it is numerically comparable to other terms (e.g., cost terms).

[0195] e: The natural constant, approximately equal to 2.718. It is the base of exponential growth, determining that the growth pattern of the penalty value is explosive.

[0196] k p Penalty slope factor, a large positive number (such as k) p =100). It determines the steepness of the wall. k p The larger the BER, the more pred Approaching BER th The faster the penalty value increases, the stronger the deterrent effect on the algorithm.

[0197] BER pred Predicted bit error rate: The expected bit error rate level of the communication service after performing the sensing task, calculated by a collaborative orchestration model.

[0198] BER thSLA (Service Level Agreement) bit error rate threshold: the inviolable red line specified in the Service Level Agreement (SLA). When the prediction error rate is much lower than the threshold (BER) pred <BER th When the threshold is reached, the exponential term is extremely small, indicating that communication is very safe and the penalty is almost zero. Once the predicted value approaches or attempts to exceed the threshold, the function value explodes (tends to infinity). This design constructs a soft wall. It allows the algorithm to explore freely away from the wall, but once it attempts to hit the wall (default), the huge penalty forces the optimization algorithm to turn back immediately, thus mathematically guaranteeing the priority of communication services.

[0199] Network resource cost item C Res Its purpose is to quantify the occupancy of valuable optical network resources by sensing tasks, preventing resource waste. Network resource cost item C Res Wavelength bandwidth resource B configured for the sensing task width and time slot resources T slot The linear weighted sum.

[0200]

[0201] Wavelength bandwidth resource B width The wider the spectrum bandwidth (GHz) occupied by the sensing task, the less space is left for communication expansion, and the higher the cost. Time slot resource T slot The duration or duty cycle of the pulse emitted by the sensing task. Longer emission times not only consume more power but also occupy a time slice. Network resource cost item C. Res This encourages the algorithm to prioritize strategies that complete tasks using minimal spectrum and time, thus optimizing spectral efficiency. α: Bandwidth weighting factor, used to convert spectrum resources (GHz) into a uniform cost unit. If network spectrum is scarce, the α value can be increased. β: Time weighting factor, used to convert time resources (ms) into a uniform cost unit. If the task is latency-sensitive, the β value can be increased.

[0202] Perceived confidence benefit term R Sense Its role is to quantify the value brought by the perception task (such as detection accuracy and resolution). In minimizing the objective function J, the perception confidence gain term R Sense A negative sign means that the greater the return, the smaller the total function value (the better).

[0203] This embodiment uses a logarithmic function to describe the returns.

[0204]

[0205] The signal-to-noise ratio (SNR) of the sensing echo is the ratio of the intensity of the target reflected signal received by the sensing receiver to the background noise. It is a core indicator that determines the probability of detection.

[0206] Spatial resolution: The minimum distance at which two adjacent targets can be distinguished. It is usually proportional to the bandwidth of the transmitted signal. ln: Natural logarithm, a mathematical operator that simulates diminishing marginal utility. The benefit of increasing from 10 to 20 is far greater than the benefit of increasing from 100 to 110.

[0207] Signal-to-noise ratio gain coefficient: the degree of importance the adjustment system places on the target that is clearly detected.

[0208] δ: Resolution gain coefficient, which adjusts the system's emphasis on achieving detailed resolution.

[0209] When the perceived signal-to-noise ratio (SNR) is low, increasing the power even slightly results in a rapid increase in the logarithmic function value. This means that investing resources at this point is extremely cost-effective, significantly improving the probability of visibility. However, when the SNR is already high, further increasing the power flattens the logarithmic curve. This indicates that investing massive amounts of resources at this point yields negligible performance improvements (e.g., from 99.9% accuracy to 99.91%). This design prevents the algorithm from recklessly consuming network resources in pursuit of minimal performance gains, guiding the system to remain in the region with the highest cost-effectiveness.

[0210] After constructing function J, a parameter combination (x) needs to be found in the candidate configuration set generated in the previous steps. opt ), such that J(x) opt Minimize. Because optical network parameters (power, frequency, time) have a mixture of nonlinear and discrete characteristics, traditional gradient descent methods are prone to getting trapped in local optima. This embodiment employs a heuristic search algorithm (e.g., Genetic Algorithm GA, Particle Swarm Optimization (PSO), or Simulated Annealing Algorithm SA). Solution process: Initialization: Randomly scatter points in the candidate configuration set (to generate the initial population).

[0211] Iterative evaluation: Calculate the J value for each individual.

[0212] Evolution / Update: Based on the J value, retain superior individuals (such as combinations with low penalty, low cost, and high return), eliminate inferior individuals, and perform crossover mutation or position update.

[0213] Convergence Output: When the number of iterations reaches the upper limit or the J value no longer decreases significantly, output the current global optimal solution.

[0214] Output results: The final parameter combination is the optimal communication slicing strategy (such as wavelength planning) and sensing slicing strategy (such as transmit power and duty cycle).

[0215] This embodiment introduces an exponentially growing SLA offset penalty term to ensure that no matter how urgent the sensing task is, the algorithm will never choose strategies that would cause the communication bit error rate to exceed the limit, perfectly solving the service pain point that operators are most concerned about: sensing affecting communication. Utilizing a logarithmic sensing benefit term effectively avoids the energy waste and nonlinear interference risks caused by the system blindly pushing the transmit power to the limit when channel conditions are good. It automatically selects strategies that are sufficient for the needs of the system, achieving green energy saving.

[0216] In one embodiment of this application, when the fluctuation slope does not exceed a safety threshold and the perception confidence level does not meet the standard, executing a step-enhancement strategy to adjust the perception slicing strategy includes: Determine whether the remaining optical signal-to-noise ratio margin of the current communication service is greater than the preset enhancement threshold; When the remaining optical signal-to-noise ratio margin of the current communication service is greater than the preset enhancement threshold, the transmit power of the sensing task is increased by a preset power step value, or the pulse duty cycle of the sensing task is increased by a preset duty cycle step value, thereby generating an enhanced sensing slice strategy. The enhanced perception slicing strategy is issued to the underlying network devices, and the bit error rate data of communication services is continuously collected within the preset monitoring time after issuance. The slope of the bit error rate data over time is calculated to obtain the specified slope; The absolute value of the slope is compared with a preset nonlinear inflection point threshold. If the absolute value of the slope is greater than the nonlinear inflection point threshold, it is determined that the communication channel has entered the nonlinear region, the step enhancement strategy is terminated and the adaptive rollback mechanism is executed to restore the network configuration to the previous stable version.

[0217] After executing the initial strategy generated by the multi-objective cooperative scheduling function, the system enters the real-time closed-loop control phase. This embodiment, instead of relying on static theoretical optimization, attempts to further explore the network's sensing potential through a step-by-step enhancement mechanism, while simultaneously utilizing a slope-based fast rollback mechanism to ensure the absolute security of communication services.

[0218] First, monitor the current health status of communication services in real time, with the core indicator being the remaining optical signal-to-noise ratio margin M. OSNR The calculation formula is as follows:

[0219] in, Optical signal-to-noise ratio (OSR): The ratio of optical signal power to noise power measured in real time at the receiver by the Optical Performance Monitoring (OPM) module, expressed in dB. This is acquired in real time by an optical spectrum analyzer (OSA) deployed at the optical network node or a coherent receiver with built-in OPM functionality.

[0220] The required optical signal-to-noise ratio (OSR) is the theoretical minimum OSR required to maintain current communication services at a specific bit error rate with error-free transmission. This parameter is typically provided in the optical module's equipment manual or determined by the forward error correction (FEC) limit threshold.

[0221] Preset enhancement threshold T enh It's a decision switch. Only when... For example, a channel is considered very healthy and has the potential to further increase sensing power only when the margin is greater than 3dB. This avoids the risks of forcibly enhancing the channel when it is inherently vulnerable. Once the activation conditions are met, the current state is no longer maintained; instead, an enhanced sensing slice strategy is generated.

[0222] The power step increases the transmit power of the sensing signal by a preset power step value ΔP (e.g., 0.1dB). Increasing the power directly improves the sensing range and echo intensity.

[0223] The duty cycle is increased by a preset duty cycle step value ΔD (e.g., 0.5%) to increase the density of sensing pulses. Increasing the duty cycle can improve the energy accumulation at the receiver and improve the signal-to-noise ratio.

[0224] Choose one or a combination of the above methods based on the current resource bottleneck (whether it is power-constrained or time-constrained). Issue the adjusted configuration to the underlying Optical Line Terminal (OLT) or Reconfigurable Optical Add-Drop Multiplexer (ROADM).

[0225] Preset monitoring time T mon It is an extremely short time window (e.g., 100ms - 500ms). After the command is issued, the system does not check the final result, but enters a high-frequency microscope mode to continuously collect bit error rate (BER) data of communication services at an extremely high sampling rate (e.g., once every 10ms).

[0226] Existing technologies typically only consider the absolute value of the bit error rate (BER), but this is often lagging; by the time the BER is exceeded, service may have already been interrupted. This embodiment instead calculates the slope of the BER's fluctuation over time. The formula for calculating the slope is:

[0227] Alternatively, the slope of the logarithmic field can be used to accommodate changes in the order of magnitude of the BER:

[0228] The slope K represents the acceleration or trend of signal quality deterioration.

[0229] If K≈0, it means that the bit error rate is stable and the interference is within a controllable range.

[0230] If K suddenly increases, it indicates that the bit error rate is rising rapidly, foreshadowing an impending avalanche effect.

[0231] K log The logarithmic slope of the bit error rate characterizes how quickly the order of magnitude of the bit error rate changes. If K... log A positive and large value indicates that the bit error rate is deteriorating exponentially. It is the best indicator for identifying the avalanche effect.

[0232] : Bit error rate at the current time, the bit error rate before correction (Pre-FEC BER) calculated by the FEC chip of the optical transmission equipment at the current sampling time t.

[0233] : Bit error rate at the previous moment, the bit error rate value recorded in the previous sampling period.

[0234] Δt: Preset monitoring time interval, the small time difference between two samples (e.g., 10ms or 20ms). Δt must be short enough; if the time is too long (e.g., on the order of seconds), the system will not be able to capture the transient changes that occur when fiber optic nonlinear effects occur, resulting in delayed protection actions.

[0235] The absolute value of the calculated slope |K| is compared with the preset nonlinear inflection point threshold (K). th (Compare)

[0236] Fiber nonlinear threshold effect: Nonlinear interference in optical fibers (such as stimulated Brillouin scattering (SBS) or cross-phase modulation (XPM) exhibits threshold characteristics. When the power is below the threshold, the interference increases linearly and weakly; once the threshold is exceeded, the interference explodes exponentially (waterfall curve).

[0237] Judgment condition: If |K|>K th If this happens, the communication channel has crossed the linear region and entered the non-linear region. At this point, although the absolute bit error rate may not have exceeded the limit, the collapse is irreversible.

[0238] Once the above judgment is triggered, the system responds in milliseconds, terminates the current step enhancement attempt, and immediately performs a rollback operation to restore the transmit power and duty cycle to the previous stable version.

[0239] This step approximates the physical limits of optical fibers in a real-world environment through real-time, incremental probing. This allows the detection distance of sensing tasks to often exceed theoretical calculations by 10%–20% in actual operation, maximizing performance. Furthermore, by setting a nonlinear inflection point threshold specifically for detecting such abrupt changes, the nonlinear region can be identified most quickly, preventing the system from falling into prolonged oscillations or locking up, thus ensuring the engineering stability of the integrated sensing system.

[0240] In one embodiment of this application, the method further includes: Perform the physical location calibration step of the optical link, and use the correspondence between the feature points of the optical time domain reflectometer curve and the actual optical cable junction box location to generate a mapping table between the optical logical distance of the optical cable link and the GIS coordinates. Feature extraction is performed on the sensing results. When event features characterizing abnormal vibration or light decay are extracted, the relative optical distance of the event features on the optical cable link is determined. Based on the mapping table, a linear interpolation algorithm is used to convert the relative optical distance into the GIS latitude and longitude coordinates of the fault point; The configuration version number of the currently effective perceptual slicing strategy is obtained through parsing. Based on the configuration version number, obtain the corresponding set of perception parameters; The severity level is obtained by classifying the severity of event features using a set of perception parameters. Based on the GIS latitude and longitude coordinates and the severity level, generate maintenance work orders.

[0241] In this embodiment, the optical cable may be coiled or cross terrain during the laying process, resulting in the optical logical distance often being greater than the geographical physical distance. To achieve accurate positioning, a calibration step is performed in advance.

[0242] Optical feature points on the optical time domain reflectometer (OTDR) curve (such as Fresnel reflection peaks at junction boxes or loss steps at fusion joints) are used as landmarks. These optical feature points are mapped one-to-one with the latitude and longitude coordinates of actual facilities (such as manholes and junction boxes) recorded in the geographic information system (GIS) to generate an optical distance-GIS coordinate mapping table.

[0243] Table structure example:

[0244] Where L iLet be the optical distance of the i-th junction box. It is its corresponding latitude and longitude coordinate vector.

[0245] Abnormal event feature extraction and relative positioning are performed, and signal processing is carried out on the real-time received sensing echo data.

[0246] Feature extraction: Identify signal waveforms that characterize abnormal vibrations (such as excavator operation) or light attenuation (such as fiber optic cable bending).

[0247] Relative optical distance determination: Calculate the optical transmission delay from the point where the abnormal event occurred to the central office equipment, and then convert it into a relative optical distance x (e.g., 15.435 km from the central office).

[0248] Consult the optical distance-GIS coordinate mapping table to find the two known calibration points (forward reference point i and backward reference point i+1) closest to the fault point x, which satisfy the following conditions.

[0249] Subsequently, a linear interpolation algorithm was used to calculate the precise geographic coordinates of the fault point. in, The target point's geographic coordinate vector represents the precise location of the fault point on the Earth's surface, typically expressed as a vector form of (longitude, latitude).

[0250] x: Relative optical distance of the event, measured by OTDR or distributed fiber optic sensing technology, the fiber optic length (in meters) from the fault point to the measuring end.

[0251] Interpolation ratio: Represents the relative position ratio of the fault point between two reference points (between 0 and 1).

[0252] : Correction for fiber coiling, a factor close to 1 but slightly less than 1 (e.g., 0.98). This factor is used to correct for the excess fiber length that is often coiled within the splice closure. This excess fiber length does not correspond to the actual geographical displacement. This factor helps eliminate positioning errors caused by the coiled fiber, preventing "virtual length" from causing geographical location deviations.

[0253] The system may dynamically adjust the transmit power or duty cycle based on network busy / idle status (i.e., change the "sensory slicing policy"). Therefore, the same signal strength represents different levels of damage under different policies. Read the configuration version ID of the currently effective sensing slicing policy.

[0254] Based on the version number, retrieve the corresponding sensing parameter set from the database (including the transmit power P at that time). tx Integration time T int Theoretical signal-to-noise ratio (SNR) benchmark ref ).

[0255] The intensity of the extracted event features is normalized using a set of perceptual parameters to obtain the true severity level S. level ).

[0256] Calculation logic:

[0257] Severity level meaning: The final threat level determined (e.g., Level 1 alarm - immediate interruption; Level 2 alarm - potential risk; Level 3 alarm - minor disturbance). Classify(⋅): A hierarchical mapping function that maps normalized values ​​to discrete levels of a ladder function.

[0258] I event Event characteristic intensity: The intensity of the raw signal directly extracted by the sensor (such as the change in Rayleigh scattering light intensity, usually expressed as change in optical power in dB or amplitude).

[0259] P tx Transmit power (from configuration version), this value must be read from the currently active configuration version.

[0260] G sys System gain includes inherent hardware parameters such as optical amplifier gain and receiver sensitivity.

[0261] This factor actually reflects the true physical energy of external disturbances (such as excavator impacts). It eliminates the distortion caused by the system adjusting its own transmit power P. tx And the interference it brings to the measurement results.

[0262] If the current mode is low power (P) tx (Small), a faint vibration was detected. After normalization (divided by the smaller P) tx Afterwards, it may be discovered that the actual magnitude is much larger, but appears smaller only because the probe light is weak, therefore it must be classified as severe. The reverse is also true.

[0263] Finally, the system integrates GIS latitude and longitude coordinates G x and severity level The system generates digital maintenance work orders that include navigation routes, fault types, urgency levels, and suggested tools, and pushes them to the handheld terminals of field personnel.

[0264] Existing technologies often calculate distances directly using the speed of light formula, ignoring fiber optic cable routing and terrain undulations, resulting in positioning errors often reaching hundreds of meters. This embodiment employs a three-level positioning algorithm—feature point mapping, linear interpolation, and fiber optic cable correction—to accurately anchor the virtual distance in the optical world to the latitude and longitude of the physical world. This allows maintenance personnel to directly navigate to the fault location, significantly reducing the time to report a fault (MTTR). The sensing parameters (power, duty cycle) of the integrated sensing system dynamically change with the communication load. If a fixed alarm threshold is used, serious intrusions may be missed due to weak echoes when the system reduces its sensing power. This embodiment achieves intelligent tiering by parsing the configuration version number, judging the size of the target based on the current parameter set, ensuring that the rating of external threats remains objective and accurate regardless of whether the system is in high-performance or low-power mode. From signal acquisition, feature analysis, coordinate transformation to work order generation, the entire process requires no manual intervention. In particular, work orders combined with GIS information directly break down information barriers between network equipment and field personnel, greatly improving the operational efficiency and intelligence level of the optical network.

[0265] like Figure 3 As shown, taking a key urban trunk line of Guangdong Mobile as an example, this section of the line... Currently in a peak period of municipal construction, daily There are risks associated with excavation operations during this period: Pain Point: This trunk line carries multiple high-priority IDC inter-port traffic, and its OSNR margin is already tightly balanced. If a regular-intensity DAS or OTDR scan is initiated during peak daytime traffic, it can easily cause the business to miss its target instantly.

[0266] The processing logic of this application is as follows: Off-peak scheduling: Based on the traffic model of the trunk line over the past week, the system identifies 2:00-4:00 AM as the business traffic trough and automatically deploys high-intensity "risk scanning tasks" to this time period.

[0267] Spatial Focusing and Parameter Stepping: The system automatically narrows the scanning range from the entire 80km line to the critical 5km area before and after the construction point (focused scanning), and adopts a stepping method to increase the pulse duty cycle.

[0268] Real-time rollback monitoring: During the enhanced sensing process, the collaborative engine monitors the BER slope reported by the wavelength division multiplexing (WDM) network management system in real time. Once a change in the slope indicates that the communication channel is at risk of entering the nonlinear region, the system immediately triggers parameter rollback within 100ms, ensuring uninterrupted sensing in the construction section while maintaining zero packet loss in communication services throughout the entire process.

[0269] Work order automation logic: After detecting micro-vibration anomalies, the system directly plots the corresponding latitude and longitude coordinates on the GIS map and automatically pushes a maintenance work order with a version number to the local maintenance team, noting "It is recommended to conduct on-site verification and enable backup route protection".

[0270] Secondly, embodiments of this application provide a sensor-integrated collaborative orchestration system, including: The first acquisition module is used to acquire the basic status information of the integrated optical network, including the service level agreement indicators of communication services, the status of optical layer physical resources, and the requirements of the sensing tasks to be executed. The first generation module is used to generate a perception task descriptor that includes task priority and maximum interference budget according to the perception task requirements. The first determining module is used to determine the multi-dimensional interference vector based on the basic state information and the sensing task descriptor. The multi-dimensional interference vector is used to quantify the nonlinear physical layer impact of the detection signal generated by the sensing task on the communication service signal. The first generation module is used to perform feasible domain pruning on the combination of configuration parameters for sensing tasks based on the physical resource status of the optical layer and multi-dimensional interference vectors, and generate a candidate configuration set that meets the service level protocol indicators of communication services. The first module is used to solve the candidate configuration set with the joint optimization objectives of minimizing the service level protocol deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks, so as to obtain the optimal communication slicing strategy and perception slicing strategy. The second acquisition module is used to issue execution instructions to the underlying network devices based on the optimal communication slicing strategy and perception slicing strategy to obtain the fluctuation slope of communication service performance indicators; when the fluctuation slope exceeds the preset safety threshold, the network configuration is restored to the previous stable version; when the fluctuation slope does not exceed the safety threshold and the perception confidence does not meet the standard, a step-by-step enhancement strategy is executed to adjust the perception slicing strategy.

[0271] In this embodiment, addressing the core challenge of high-power pulses from sensing signals easily inducing fiber nonlinear effects (such as cross-phase modulation, XPM) and thus degrading communication quality, a multi-dimensional interference vector is used. This decouples the abstract and complex physical layer nonlinear interference into specific mathematical vectors such as the predicted bit error rate increment and the optical signal-to-noise ratio degradation value. This allows the orchestration system to accurately quantify the potential physical damage of sensing tasks to the Communication Service Level Agreement (SSLA), achieving a qualitative leap from the traditional, broad-based power threshold to refined interference budget management. Furthermore, it effectively solves the problem of high-power pulses generated by sensing devices potentially triggering stimulated Brillouin scattering (SBS) or cross-phase modulation (XPM). Even if the transmitting power at the sensing end is within a safe threshold, the bit error rate (BER) on the communication side will still fluctuate significantly and periodically with the scanning pulse, causing instantaneous off-target errors or packet loss in existing network services.

[0272] In one embodiment of this application, determining the multidimensional interference vector based on basic state information and a perception task descriptor includes: Based on basic state information and sensing task descriptors, a collaborative orchestration model for communication services and sensing tasks is constructed; the collaborative orchestration model is calculated to determine the multidimensional interference vector.

[0273] In one embodiment of this application, the sensing task descriptor includes transmit power parameters and duty cycle parameters; based on basic state information and the sensing task descriptor, a collaborative orchestration model for communication services and sensing tasks is constructed; the collaborative orchestration model is calculated to determine the multi-dimensional interference vector, including: Multidimensional resource abstraction is performed on the optical layer physical resource status of basic state information to establish a unified resource view that includes network topology, wavelength channel status, power budget and optical signal-to-noise ratio margin. Based on the unified resource view, the modulation format of the transmission link where the communication service is located is analyzed, and a nonlinear sensitivity coefficient is defined according to the modulation format. The nonlinear sensitivity coefficient is used to characterize the mapping relationship between the sensing pulse intensity and the degradation of the communication signal quality. In the unified resource view, a collaborative orchestration model is constructed using the transmit power parameter and duty cycle parameter in the perception task descriptor as input variables and the nonlinear sensitivity coefficient as the transformation parameter. The parameters of the sensing task to be performed are substituted into the collaborative orchestration model to obtain a multidimensional interference vector. The multidimensional interference vector includes at least the instantaneous power crosstalk value, the optical signal-to-noise ratio degradation prediction value, and the bit error rate increment prediction value.

[0274] In one embodiment of this application, feasible domain pruning is performed on the combination of configuration parameters for the sensing task based on the optical layer physical resource status and multi-dimensional interference vectors to generate a candidate configuration set that meets the service level protocol indicators of communication services, including: Collect optical power sampling sequences of optical layer physical resources within a preset historical time window; The statistical variance is calculated from the optical power sampling sequence and used as a feature value characterizing the nonlinear noise level of the link environment. Construct a dynamic function with statistical variance as the independent variable; The dynamic safety margin is calculated based on the dynamic function. The larger the statistical variance, the larger the dynamic safety margin, so as to compress the available disturbance budget space. Define a cumulative disturbance constraint condition, which is the sum of the cumulative value of the multidimensional disturbance vectors of all active tasks at any given time and the dynamic safety margin, and it must not exceed the maximum disturbance budget. By traversing all potential parameter combinations for the perception task and eliminating parameter combinations that do not satisfy the cumulative interference constraint, a candidate configuration set is obtained.

[0275] In one embodiment of this application, the optimal communication slicing strategy and perception slicing strategy are obtained by solving a candidate configuration set with the joint optimization objectives of minimizing the service level agreement deviation of communication services, minimizing network resource consumption, and maximizing the confidence of the perception task. Construct a multi-objective cooperative scheduling function, which includes a service level protocol offset penalty term, a network resource cost term, and a perceived confidence benefit term. The service level agreement offset penalty term is configured such that when the predicted performance metric of the communication service is worse than the service level agreement metric, the function value increases exponentially. The network resource cost item is configured as: the weighted sum of wavelength bandwidth resources and time slot resources occupied by the quantified sensing task; The perception confidence benefit term is configured as follows: the benefits brought by the spatial resolution and sampling rate of the perception task are described in the form of a logarithmic function to reflect the diminishing marginal effect of perception resource investment. By using a heuristic search algorithm to find the parameter combination that minimizes the value of the multi-objective cooperative scheduling function in the candidate configuration set, the optimal communication slicing strategy and perception slicing strategy are obtained.

[0276] In one embodiment of this application, when the fluctuation slope does not exceed a safety threshold and the perception confidence level does not meet the standard, executing a step-enhancement strategy to adjust the perception slicing strategy includes: Determine whether the remaining optical signal-to-noise ratio margin of the current communication service is greater than the preset enhancement threshold; When the remaining optical signal-to-noise ratio margin of the current communication service is greater than the preset enhancement threshold, the transmit power of the sensing task is increased by a preset power step value, or the pulse duty cycle of the sensing task is increased by a preset duty cycle step value, thereby generating an enhanced sensing slice strategy. The enhanced perception slicing strategy is issued to the underlying network devices, and the bit error rate data of communication services is continuously collected within the preset monitoring time after issuance. The slope of the bit error rate data over time is calculated to obtain the specified slope; The absolute value of the slope is compared with a preset nonlinear inflection point threshold. If the absolute value of the slope is greater than the nonlinear inflection point threshold, it is determined that the communication channel has entered the nonlinear region, the step enhancement strategy is terminated and the adaptive rollback mechanism is executed to restore the network configuration to the previous stable version.

[0277] In one embodiment of this application, the method further includes: Perform the physical location calibration step of the optical link, and use the correspondence between the feature points of the optical time domain reflectometer curve and the actual optical cable junction box location to generate a mapping table between the optical logical distance of the optical cable link and the GIS coordinates. Feature extraction is performed on the sensing results. When event features characterizing abnormal vibration or light decay are extracted, the relative optical distance of the event features on the optical cable link is determined. Based on the mapping table, a linear interpolation algorithm is used to convert the relative optical distance into the GIS latitude and longitude coordinates of the fault point; The configuration version number of the currently effective perceptual slicing strategy is obtained through parsing. Based on the configuration version number, obtain the corresponding set of perception parameters; The severity level is obtained by classifying the severity of event features using a set of perception parameters. Based on the GIS latitude and longitude coordinates and the severity level, generate maintenance work orders.

[0278] The functions of each module in each device in the embodiments of this application can be found in the corresponding descriptions in the above methods, and will not be repeated here.

[0279] Figure 4 A structural block diagram of an electronic device according to an embodiment of this application is shown. Figure 4As shown, the electronic device includes a memory 410 and a processor 420. The memory 410 stores instructions that can be executed on the processor 420. When the processor 420 executes the instructions, it implements the synergistic orchestration method of the above embodiments. The number of memories 410 and processors 420 can be one or more. This electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or claimed herein.

[0280] The electronic device may also include a communication interface 430 for communicating with external devices and exchanging data. The devices are interconnected using different buses and can be mounted on a common motherboard or otherwise as needed. The processor 420 can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In other embodiments, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple electronic devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). The bus can be divided into address buses, data buses, control buses, etc. For ease of illustration, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0281] Optionally, in a specific implementation, if the memory 410, processor 420 and communication interface 430 are integrated on a single chip, the memory 410, processor 420 and communication interface 430 can communicate with each other through an internal interface.

[0282] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor can be a processor supporting Advanced Reduced Instruction Set Machines (ARM) architecture.

[0283] This application provides a computer-readable storage medium (such as the memory 410 described above) that stores computer instructions, which, when executed by a processor, implement the method provided in this application.

[0284] Optionally, memory 410 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device, etc. Furthermore, memory 410 may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 410 may optionally include memory remotely located relative to processor 420, and these remote memories can be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0285] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0286] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0287] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all 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.

Claims

1. A synesthetic integrated collaborative orchestration method, characterized in that, include: Acquire basic status information of the integrated sensing and communication optical network, including service level agreement indicators of communication services, status of optical layer physical resources, and requirements for sensing tasks to be performed. Based on the perception task requirements, a perception task descriptor containing task priority and maximum interference budget is generated; Based on the basic state information and the sensing task descriptor, a multi-dimensional interference vector is determined. The multi-dimensional interference vector is used to quantify the nonlinear physical layer impact of the detection signal generated by the sensing task on the communication service signal. Based on the physical resource status of the optical layer and the multidimensional interference vector, feasible domain pruning is performed on the combination of configuration parameters for the sensing task to generate a candidate configuration set that meets the service level protocol indicators of the communication service. In the candidate configuration set, the optimal communication slicing strategy and the optimal perception slicing strategy are obtained by solving the problem with the joint optimization objectives of minimizing the service level agreement deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks. Based on the optimal communication slicing strategy and the perception slicing strategy, execution instructions are issued to the underlying network devices to obtain the fluctuation slope of the communication service performance indicators; when the fluctuation slope exceeds the preset safety threshold, the network configuration is restored to the previous stable version; when the fluctuation slope does not exceed the safety threshold and the perception confidence does not meet the standard, a step-by-step enhancement strategy is executed to adjust the perception slicing strategy.

2. The method according to claim 1, characterized in that, The determination of the multidimensional interference vector based on the basic state information and the perception task descriptor includes: Based on the basic state information and the sensing task descriptor, a collaborative orchestration model for communication services and sensing tasks is constructed; the collaborative orchestration model is calculated to determine the multidimensional interference vector.

3. The method according to claim 2, characterized in that, The sensing task descriptor includes transmit power parameters and duty cycle parameters; based on the basic state information and the sensing task descriptor, a collaborative orchestration model for communication services and sensing tasks is constructed. The multidimensional interference vector is determined by calculating the collaborative orchestration model, including: Multidimensional resource abstraction is performed on the optical layer physical resource status of the basic state information to establish a unified resource view that includes network topology, wavelength channel status, power budget and optical signal-to-noise ratio margin. Based on the unified resource view, the modulation format of the transmission link where the communication service is located is analyzed, and a nonlinear sensitivity coefficient is defined according to the modulation format. The nonlinear sensitivity coefficient is used to characterize the mapping relationship between the sensing pulse intensity and the degradation of the communication signal quality. In the unified resource view, a collaborative orchestration model is constructed using the transmit power parameter and duty cycle parameter in the perception task descriptor as input variables and the nonlinear sensitivity coefficient as transformation parameter. The parameters of the sensing task to be performed are substituted into the collaborative orchestration model to obtain the multidimensional interference vector. The multidimensional interference vector includes at least the instantaneous power crosstalk value, the optical signal-to-noise ratio degradation prediction value, and the bit error rate increment prediction value.

4. The method according to claim 3, characterized in that, The step of performing feasible domain pruning on the combination of configuration parameters for the sensing task based on the optical layer physical resource status and the multi-dimensional interference vector to generate a candidate configuration set that meets the service level protocol indicators of the communication service includes: Collect optical power sampling sequences of the optical layer physical resources within a preset historical time window; The statistical variance is calculated from the optical power sampling sequence and used as a feature value characterizing the nonlinear noise level of the link environment. Construct a dynamic function with the statistical variance as the independent variable; The dynamic safety margin is calculated based on the dynamic function, wherein the larger the statistical variance, the larger the dynamic safety margin, so as to compress the available disturbance budget space. Define a cumulative disturbance constraint condition, which is the sum of the cumulative value of the multidimensional disturbance vector of all active tasks at any time and the dynamic safety margin, which shall not exceed the maximum disturbance budget. By traversing all potential parameter combinations for the perception task and eliminating parameter combinations that do not satisfy the cumulative interference constraint, a candidate configuration set is obtained.

5. The method according to claim 4, characterized in that, The optimal communication slicing strategy and perception slicing strategy are obtained by solving the candidate configuration set with the joint optimization objectives of minimizing the service level agreement deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks. Construct a multi-objective cooperative scheduling function, which includes a service level agreement offset penalty term, a network resource cost term, and a perceived confidence benefit term; The service level agreement offset penalty term is configured such that when the predicted performance index of the communication service is worse than the service level agreement index, the function value increases exponentially. The network resource cost item is configured as: a weighted sum of wavelength bandwidth resources and time slot resources occupied by the quantized sensing task; The perception confidence benefit term is configured to describe the benefits brought by the spatial resolution and sampling rate of the perception task in the form of a logarithmic function, so as to reflect the diminishing marginal effect of perception resource investment. A heuristic search algorithm is used to find the parameter combination that minimizes the value of the multi-objective cooperative scheduling function in the candidate configuration set, thereby obtaining the optimal communication slicing strategy and perception slicing strategy.

6. The method according to claim 5, characterized in that, The step-by-step enhancement strategy to adjust the perception slice strategy when the fluctuation slope does not exceed the safety threshold and the perception confidence level does not meet the standard includes: Determine whether the remaining optical signal-to-noise ratio margin of the current communication service is greater than the preset enhancement threshold; When the remaining optical signal-to-noise ratio margin of the current communication service is greater than the preset enhancement threshold, the transmit power of the sensing task is increased by a preset power step value, or the pulse duty cycle of the sensing task is increased by a preset duty cycle step value, thereby generating an enhanced sensing slice strategy. The enhanced perception slicing strategy is issued to the underlying network devices, and the bit error rate data of communication services is continuously collected within a preset monitoring time after issuance. The slope of the bit error rate data changing over time is calculated to obtain a specified slope; The absolute value of the slope is compared with a preset nonlinear inflection point threshold; If the absolute value of the slope is greater than the nonlinear inflection point threshold, it is determined that the communication channel has entered the nonlinear region, the step enhancement strategy is terminated and the adaptive rollback mechanism is executed to restore the network configuration to the previous stable version.

7. The method according to claim 6, characterized in that, The method further includes: Perform the physical location calibration step of the optical link, and use the correspondence between the feature points of the optical time domain reflectometer curve and the actual optical cable junction box location to generate a mapping table between the optical logical distance of the optical cable link and the GIS coordinates. Feature extraction is performed on the sensing results. When event features characterizing abnormal vibration or light decay are extracted, the relative optical distance of the event features on the optical cable link is determined. Based on the mapping table, a linear interpolation algorithm is used to convert the relative optical distance into the GIS latitude and longitude coordinates of the fault point; The configuration version number of the currently effective perceptual slicing strategy is obtained through parsing. Based on the configuration version number, the corresponding set of perception parameters is obtained; The severity level is obtained by classifying the event features using the set of perception parameters. A maintenance work order is generated based on the GIS latitude and longitude coordinates and the severity level.

8. A synesthetic integrated collaborative orchestration system, characterized in that, include: The first acquisition module is used to acquire the basic status information of the integrated optical network, which includes the service level agreement indicators of communication services, the status of optical layer physical resources, and the requirements of the sensing tasks to be executed. The first generation module is used to generate a perception task descriptor containing task priority and maximum interference budget according to the perception task requirements. The first determining module is used to determine a multi-dimensional interference vector based on the basic state information and the sensing task descriptor. The multi-dimensional interference vector is used to quantify the nonlinear physical layer impact of the detection signal generated by the sensing task on the communication service signal. The first generation module is used to perform feasible domain pruning on the combination of configuration parameters for the sensing task based on the optical layer physical resource status and the multi-dimensional interference vector, and generate a candidate configuration set that meets the service level protocol indicators of the communication service. The first obtaining module is used to solve the candidate configuration set with the joint optimization objectives of minimizing the service level protocol deviation of communication services, minimizing network resource consumption, and maximizing the confidence of perception tasks, so as to obtain the optimal communication slicing strategy and perception slicing strategy. The second acquisition module is used to issue execution instructions to the underlying network devices based on the optimal communication slicing strategy and the perception slicing strategy to acquire the fluctuation slope of the communication service performance indicators; when the fluctuation slope exceeds the preset safety threshold, the network configuration is restored to the previous stable version; when the fluctuation slope does not exceed the safety threshold and the perception confidence does not meet the standard, a step-by-step enhancement strategy is executed to adjust the perception slicing strategy.

9. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A computer-readable storage medium storing computer instructions that, when executed by a processor, implement the method as described in any one of claims 1-7.

Images