Self-adaptive separation method, device and equipment for power carrier signals and medium

Abstract

The invention relates to the technical field of power systems, and provides a power carrier signal adaptive separation method, device and equipment and a medium, and the method comprises the steps: sampling a power carrier signal, and decomposing the power carrier signal to obtain a multi-mode signal; dividing an interference component in the multi-mode signal into a plurality of frequency modes, and dynamically adjusting each frequency mode based on the intra-mode parameters to obtain a decoupling mode; performing time shift compensation on disturbance in the decoupling mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain a steady-state feature; and adaptively calculating the modal energy of each interference modal in the steady-state characteristics, and dynamically adjusting the contribution proportion of each interference modal to obtain a multi-modal mask so as to realize the adaptive separation of the power carrier signals. According to the invention, high-precision and robust separation and dynamic recovery of the power line carrier signal can be realized, and the method is suitable for a complex power line channel environment.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to an adaptive separation method for power line carrier signals, an adaptive separation device for power line carrier signals, a corresponding electronic device, and a corresponding computer-readable storage medium. Background Technology

[0002] In the related technologies of power line carrier signal separation and recovery, traditional signal processing methods, time-frequency analysis and mode decomposition methods, and intelligent optimization and deep learning methods can be used. Although these three types of methods can improve the separation accuracy and noise resistance of power signals to varying degrees, they are not suitable for complex power line channel environments with multimodal nonlinear disturbances and strong time variations. Summary of the Invention

[0003] This application provides an adaptive separation method, apparatus, device, and medium for power line carrier signals. It is a novel algorithm that simultaneously possesses disturbance decoupling capability, time-frequency structure preservation characteristics, and adaptive learning mechanism to achieve high-precision, robust separation and dynamic recovery of power line carrier signals, and is suitable for complex power line channel environments.

[0004] In one aspect, this application provides an adaptive separation method for power line carrier signals, the method comprising:

[0005] The power line carrier signal is sampled, and the power line carrier signal is decomposed to obtain a multimode signal;

[0006] The interference components in the multimodal signal are divided into multiple frequency modes, and each frequency mode is dynamically adjusted based on the intramodal parameters to obtain the decoupled mode.

[0007] Time-shift compensation is performed on the disturbances in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain steady-state characteristics;

[0008] The modal energy of each interference mode in the steady-state characteristics is adaptively calculated, and the contribution ratio of each interference mode is dynamically adjusted to obtain a multimode mask, thereby realizing the adaptive separation of the power line carrier signal.

[0009] On the other hand, this application provides an adaptive separation device for power line carrier signals, the device comprising:

[0010] The signal processing module is used to sample the power line carrier signal and decompose the power line carrier signal to obtain a multimode signal;

[0011] A multi-modal disturbance decoupling module is used to divide the interference components in the multi-modal signal into multiple frequency modes, and dynamically adjust each frequency mode based on the intramodal parameters to obtain decoupled modes;

[0012] The time-frequency feature extraction module is used to perform time-shift compensation on the disturbances in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain steady-state features;

[0013] The mode weight update module is used to adaptively calculate the modal energy of each interference mode in the steady-state characteristics and dynamically adjust the contribution ratio of each interference mode to obtain a multimode mask, thereby realizing the adaptive separation of the power line carrier signal.

[0014] In another aspect, this application also provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the adaptive separation method for power line carrier signals as described in any one of the claims.

[0015] In another aspect, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the adaptive separation method for power line carrier signals as described in any one of the claims.

[0016] In another aspect, this application also provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the adaptive separation method for power line carrier signals described in the above aspects.

[0017] The adaptive separation method, apparatus, device, and storage medium for power line carrier signals provided in this application sample the power line carrier signal, decompose the power line carrier signal to obtain a multimode signal, divide the interference components in the multimode signal into multiple frequency modes, and dynamically adjust each frequency mode based on the intramodal parameters to obtain decoupled modes. Then, time-shift compensation can be performed on the disturbances in the decoupled modes based on time compensation coefficients and frequency scaling coefficients to obtain steady-state characteristics. Finally, the modal energy of each interference mode in the steady-state characteristics can be adaptively calculated, and the contribution ratio of each interference mode can be dynamically adjusted to obtain a multimode mask, thereby realizing the adaptive separation of power line carrier signals. By dividing the interference components into multiple frequency modes and obtaining decoupled modes, multi-mode disturbances are decoupled, enabling high-level differentiation of interference from different sources in the modal domain. Furthermore, time-shift compensation is applied to the disturbances in the decoupled modes based on time compensation coefficients and frequency scale adjustment coefficients, achieving steady-state feature extraction of time-frequency invariant transformations. This ensures the carrier signal maintains a robust time-frequency structure under frequency drift and nonlinear disturbances. Moreover, adaptive mode weight optimization is achieved through adaptive calculation of mode energy and dynamic adjustment of the contribution ratio of each interference mode, realizing adaptive separation of the power line carrier signal and providing technical support for the subsequent prominent recovery of the target carrier signal. Therefore, the adaptive separation scheme for power line carrier signals provided in this application is applicable to the complex channel environment of multi-mode nonlinear disturbances and highly time-varying power lines. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the steps of an adaptive separation method for power line carrier signals provided in an embodiment of this application.

[0019] Figure 2 This is a structural block diagram of an adaptive separation device for power line carrier signals provided in an embodiment of this application;

[0020] Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of this application;

[0021] Figure 4 This is a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] In the related technologies for power line carrier signal separation and recovery, they can be divided into three categories: traditional signal processing methods, time-frequency analysis and mode decomposition methods, and intelligent optimization and deep learning methods. Although these three types of methods can improve the separation accuracy and noise resistance of power lines to varying degrees, they still have obvious technical bottlenecks and insufficient adaptability in the context of multimodal nonlinear disturbances and strong time-varying conditions.

[0024] Specifically, the first category comprises traditional methods based on linear filtering and statistical separation, such as bandpass filtering, Independent Component Analysis (ICA), and Principal Component Analysis (PCA). These methods typically assume that the target signal and noise have separable spectra or statistical independence, and extract the signal through linear projection. However, in real-world power line carrier environments, interference signals and carrier signals often overlap in the spectrum and change dynamically over time. Algorithms like ICA and PCA cannot accurately capture their non-stationary characteristics, limiting their separation accuracy. Furthermore, filter banks experience a sharp performance degradation when faced with higher-order harmonics and nonlinear distortions, leading to carrier signal waveform distortion and amplitude attenuation, severely impacting the reliability of communication resolution.

[0025] The second category comprises improved algorithms based on mode decomposition and time-frequency reconstruction, such as Empirical Mode Decomposition (EMD), Ensemble Empirical Mode Decomposition (EEMD), Variational Mode Decomposition (VMD), and Hilbert-Huang Transform (HHT). These methods achieve component extraction at different time scales through adaptive signal decomposition. For example, EMD-type algorithms have good nonlinear processing capabilities but are extremely sensitive to noise and prone to mode aliasing and endpoint effects. VMD improves decomposition stability by introducing variational constraints, but its frequency band division requires a preset center frequency, making it insufficient for non-stationary signals in complex power disturbance environments. While time-frequency reconstruction methods such as HHT can obtain instantaneous frequency information, their energy distribution is easily distorted under strong noise, making it difficult to maintain structural consistency in the time and frequency domains, leading to feature shifts and misseparation in subsequent signal recognition stages.

[0026] The third category comprises intelligent signal separation methods based on deep neural networks and adaptive optimization, such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), self-attention mechanisms, and Generative Adversarial Networks (GANs). These methods can establish complex nonlinear mapping relationships through end-to-end learning, exhibiting strong expressive power for signal features. However, they rely on large-scale labeled samples and specific training scenarios; their model performance significantly degrades once the channel environment or disturbance characteristics change. Furthermore, the network structure lacks interpretability, making it difficult to theoretically guarantee the stability and consistency of signal separation. In addition, deep network algorithms still suffer from feature overlap and information loss in low signal-to-noise ratio, spectral drift, or asynchronous interference environments, limiting their widespread application in practical power line carrier communication.

[0027] As mentioned above, while power line carrier signal separation and recovery technologies have made some progress in the field of power line signal separation and enhancement, their algorithms are mostly based on static signal models, making it difficult to capture the nonlinear evolution characteristics of multimodal dynamic disturbances. Furthermore, the lack of time-frequency invariant constraints makes the signal prone to instability under spectral drift and multi-source interference. Additionally, the high computational complexity and data dependency make them unsuitable for real-time deployment on embedded or edge devices. In short, they are not applicable to the complex power line channel environment characterized by multimodal nonlinear disturbances and strong time-varying characteristics.

[0028] This application presents a novel algorithm designed for complex power line channel environments. It simultaneously possesses disturbance decoupling capabilities, time-frequency structure preservation characteristics, and an adaptive learning mechanism to achieve high-precision, robust separation and dynamic recovery of power line carrier signals. This addresses the problems of traditional power line carrier signal separation methods in complex power distribution environments, such as susceptibility to multi-source interference, instability of time-frequency characteristics, low separation accuracy, and insufficient adaptive capabilities. The technical solution provided by this application enables power line carrier communication systems to achieve high-precision, millisecond-level adaptive separation under conditions of multi-source interference and nonlinear coupling. It also provides interference signal extraction for transformer area monitoring and anomaly analysis, effectively improving communication signal reliability, data transmission quality, and the monitoring capabilities of intelligent distribution transformer areas. This demonstrates significant engineering application value and patent protection potential.

[0029] Reference Figure 1 The diagram illustrates a flowchart of an adaptive separation method for power line carrier signals according to an embodiment of this application, which may specifically include the following steps:

[0030] Step S101: Sample the power line carrier signal and decompose the power line carrier signal to obtain a multimode signal.

[0031] In the signal acquisition and preprocessing stage, multi-scale bandpass filtering, normalization, and window function weighting methods can be used to perform high-fidelity sampling and edge stabilization processing on power line carrier signals and superimposed disturbance signals to obtain multi-mode signals.

[0032] Among them, by using multi-scale processing technology, it is possible to retain the high-frequency dynamic characteristics of the signal while removing DC offset, fundamental interference and initial noise, providing a reliable input for subsequent multi-mode disturbance decoupling, and significantly improving the signal quality and engineering feasibility of the initial separation.

[0033] Step S102: Divide the interference components in the multimode signal into multiple frequency modes, and dynamically adjust each frequency mode based on the intramodal parameters to obtain the decoupled mode.

[0034] In the multimodal disturbance decoupling stage, specifically, the disturbance components are decoupled in the time and frequency domains.

[0035] Optionally, the multimode signal is a mixed signal. An adaptive decomposition method can be used to divide the mixed signal into multiple frequency modes. Then, a decoupling factor can be introduced to dynamically adjust each frequency mode. This dynamic adjustment of each frequency mode is achieved by constructing a dynamically changing mode decoupling matrix in the time-frequency domain. The decoupling factor can be represented as the intrinsic parameters of the mode decoupling matrix. Specifically, it can adjust the amplitude and frequency distribution of each mode during the decomposition process, enhance the energy components of the target carrier mode, make the target carrier signal dominate in the mode domain, and simultaneously suppress multi-source interference such as load fluctuations, harmonics, and switching transients. This achieves dynamic adjustment of each frequency mode, resulting in decoupled modes.

[0036] Decoupling modes eliminate the main modal coupling interference in the time and frequency domain. The embodiments of this application target the nonlinear coupling characteristics of power systems, achieving efficient differentiation of multi-source disturbances and laying the foundation for high-precision signal separation.

[0037] Step S103: Time-shift compensation is performed on the disturbance in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain the steady-state characteristics.

[0038] In the time-frequency feature extraction stage, time-shift compensation can be performed on the disturbances in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to ensure the consistency of the time-frequency structure of the target signal, so that the carrier signal remains stable in the energy concentration area under frequency drift or nonlinear disturbance conditions, and achieves the purpose of time-frequency invariant transformation.

[0039] Optionally, the above-mentioned time-frequency invariant transformation strategy can be used to extract steady-state features of the decoupled modes, thereby realizing the extraction of reliable features under complex disturbances, providing robust support for adaptive separation, and providing accurate basis for subsequent mode weight optimization.

[0040] Step S104: Adaptively calculate the modal energy of each interference mode in the steady-state characteristics, and dynamically adjust the contribution ratio of each interference mode to obtain a multimode mask, thereby realizing the adaptive separation of the power line carrier signal.

[0041] In the mode weight optimization stage, the modal energy and correlation are adaptively calculated, and the contribution ratio of the mode in the reconstruction is dynamically adjusted by combining iterative optimization methods to achieve outstanding recovery of the subsequent target carrier signal while suppressing interference residues.

[0042] Optionally, adaptive iterative weight adjustment can be combined with multimodal decoupling closed loop to enable faster signal reconstruction, higher separation accuracy, and maintain stability under dynamic disturbance environments. This application does not impose any restrictions on this.

[0043] In the embodiments of this application, by constructing a complete signal processing closed loop and adaptive optimization system, the target carrier signal and interference components form a continuous and controllable processing chain in the multi-mode domain and time-frequency domain, thereby achieving high-precision, adaptive separation and dynamic recovery of the power line carrier signal.

[0044] This application proposes an Adaptive-Multimodal-Disturbance Time-Frequency-Invariant-Separation (AMD-TIS) algorithm for power line carrier signal multimodal disturbance decoupling. Addressing the challenges of complex channel interference, strong randomness of time-varying noise, and difficulty in eliminating multimodal coupling in existing carrier communication, this algorithm constructs a cross-time-frequency domain, end-link interference identification and signal reconstruction method. First, at the signal modeling level, a multi-component hybrid model of the power line carrier signal is established. Through short-time Fourier transform and time-frequency distribution operators, the carrier signal, harmonic interference, impulse interference, and noise components are jointly characterized, achieving local separability and structured description of the signal in the time-frequency domain, laying a theoretical foundation for subsequent interference modeling and separation.

[0045] In some embodiments of this application, step S101 may specifically include the following sub-steps:

[0046] Sub-step S11 involves performing time-frequency transformation on the mixed signal to obtain the signal time-frequency distribution function of each interference component in the time-frequency domain;

[0047] Sub-step S12 involves arranging the complex amplitude terms of each interference component in the signal time-frequency distribution function to obtain the joint time-frequency distribution matrix.

[0048] In actual power line carrier environments, interference signals and carrier signals often overlap in the spectrum. The actual sampled power line carrier signal is a complex signal, which is usually the original signal that contains the real carrier signal and various noise-related disturbance signals superimposed.

[0049] Optionally, the actual sampled power line carrier signal can be represented as a mixed signal via a multi-component mixing model. For example, the fundamental mathematical model of the power line carrier signal, i.e., the multi-component mixing model, can be specifically represented as follows:

[0050]

[0051] In the formula, This refers to the actual sampled power line carrier signal; The actual carrier signal represents the effective component of the carrier communication data after modulation, which is the target signal that ultimately needs to be extracted from the interference; This represents the i-th type of interference signal component, mainly used to represent disturbances from different sources (such as electromagnetic noise, interference from household appliance switches, harmonic interference, etc.). Gaussian white noise refers to the random background noise component in the modeled circuit; M is the total number of interference modes, used to define the number of interference categories identified in the current system.

[0052] The aforementioned multi-component mixing model decomposes the original signal into a mixed signal containing carrier components, minor interference components, and background noise, defining the composition of the signal and thus providing structured input for subsequent analysis.

[0053] For ease of processing, a time-frequency transformation can be performed on the mixed signal to obtain the time-frequency distribution function of each interference component in the time-frequency domain, as shown below:

[0054]

[0055] In the formula, The result is the Short-Time Fourier Transform (STFT), which represents the complex amplitude distribution of the signal in the time and frequency domain and is the basis for subsequent calculations of coherence and interference decoupling. ω is the time shift parameter, mainly used for sliding time windows; ω is the angular frequency variable, mainly used for analyzing signal components in different frequency bands. This is the expression for the rotation factor in a complex function; The window function (e.g., Gaussian window or Hanning window) is mainly used to ensure time-frequency localization analysis.

[0056] The signal time-frequency distribution function can be used to indicate the complex amplitude distribution of a signal in the time-frequency domain.

[0057] To achieve subsequent multimodal disturbance decoupling, embodiments of this application can define a joint time-frequency distribution matrix for multimodal disturbances. Specifically, the complex amplitude terms of each disturbance component in the signal time-frequency distribution function can be jointly arranged to obtain the joint time-frequency distribution matrix.

[0058] The joint time-frequency distribution matrix is ​​in matrix form of multimode signals and can record the amplitude of all interference components in the time-frequency domain.

[0059] In practical implementation, the energy distributions of each disturbance component in the complex amplitude form of the time-frequency domain can be jointly arranged to represent the energy coupling and coherence relationships of different modes in the same time-frequency plane in a matrix manner. For example, the defined joint time-frequency distribution matrix of the multimodal disturbance can be as follows:

[0060]

[0061] In the formula, The interference mode matrix records the amplitudes of all interference components in the time-frequency domain, which is used for subsequent multi-mode perturbation decoupling, feature extraction, and weight estimation. It is a joint energy representation structure naturally constructed from the complex amplitude distribution of the signal's time-frequency domain. Each element of the matrix originates from the complex amplitude term (i.e., a complex form containing amplitude and phase) in the signal's time-frequency distribution function, thus preserving both the energy intensity and phase consistency information of different modes in the joint domain.

[0062] By using time-frequency transformation, different interference characteristics can be distinguished in the local time-frequency domain, making the interference components separable and modelable on the time-frequency plane, and defining the mathematical space for subsequent variables.

[0063] It should be noted that interference components are specific useless signals / energy superimposed on a signal that hinder target analysis or the normal operation of the system; interference modes are mode components containing only interference elements separated from a mixed signal after mode decomposition. Interference components are the source of interference modes, and interference modes are the existing form of interference components after decomposition.

[0064] Based on the above, the algorithm of this application introduces a time-frequency coherence analysis mechanism. By calculating the coherence function and correlation matrix between different interference modes, it achieves decoupling and feature orthogonalization of multi-source interference, thereby separating interference components that originally overlapped in the same frequency band.

[0065] In some embodiments of this application, step S102 may specifically include the following sub-steps:

[0066] Sub-step S21: Based on the energy coupling and coherence relationship of different interference modes in the same time-frequency plane, define the perturbation coherence function between each interference mode and obtain the coherence coefficient between each interference mode;

[0067] Sub-step S22: Construct the mode decoupling matrix based on the coherence coefficients between each interference mode;

[0068] Sub-step S23 involves linearly transforming the signal time-frequency distribution function of each interference component in the time-frequency domain based on the mode decoupling matrix to obtain the signal representation after preliminary decoupling.

[0069] The embodiments of this application can perform modal decoupling of interference components in the time-frequency domain. Multimodal disturbance decoupling enables efficient differentiation of interference from different sources in the modal domain.

[0070] The joint time-frequency distribution matrix can represent the energy coupling and coherence relationship of different interference modes in the same time-frequency plane in matrix form.

[0071] Dynamic adjustment of each frequency mode is based on modal intrinsic parameters, achieved by constructing a modal decoupling matrix. The modal intrinsic parameters can be the coherence coefficients between various interfering modes, enabling dynamic construction of the modal decoupling matrix at each time-frequency point. Specifically, firstly, based on the energy coupling and coherence relationship between different interfering modes in the same time-frequency plane, the perturbation coherence function between each interfering mode is defined in real time. The time-frequency dependent coherence coefficients are obtained, and then the modal decoupling matrix can be constructed based on these time-frequency dependent coherence coefficients. Since the coherence relationship of the interference components is non-stationary and time-varying, the matrix... At each time frequency point All of the above are dynamically constructed and updated in real time. This dynamic decoupling matrix... It was then used to perform a linear transformation on the signal time-frequency distribution function, thereby dynamically removing the main mode coupling interference in the time-frequency domain, enhancing the energy components of the target carrier mode, and finally obtaining the decoupled mode that reflects the dynamic adjustment result.

[0072] In practical implementation, firstly, the coherence function between different types of interference can be calculated to distinguish various types of interference and quantitatively characterize the coupling strength between interference components. The definition of the perturbation coherence function can be as follows:

[0073]

[0074] In the formula, The coherence coefficient between the i-th and j-th interference modes can be used to measure the correlation between interference components; the smaller the value, the more independent they are. This is the inner product operator, used to calculate the similarity between two time-frequency signals; The vector norm is used to normalize the inner product, ensuring... Within the range of [0,1].

[0075] Furthermore, a mode decoupling matrix can be constructed. Optionally, the mode decoupling matrix can be constructed by subtracting the coherence matrix from the identity matrix, thereby achieving independence between interfering modes. For example, the constructed mode decoupling matrix can be as follows:

[0076]

[0077] In the formula, is the mode decoupling matrix; I is the identity matrix, belonging to the identity amplitude term, which can be used to describe the standard amplitude distribution of a signal when it is undisturbed.

[0078] The construction of the coherence matrix, i.e. the mode decoupling matrix, enables the algorithm to identify highly correlated interference modes. At this time, the energy of these interference modes can be separated by the constructed mode decoupling matrix A, thereby obtaining a more independent signal characterization Z, and achieving efficient differentiation of interference from different sources in the modal domain.

[0079] For example, the decoupled modes can be represented based on the signal representation after initial decoupling. This initial decoupled signal representation can be obtained by performing a linear transformation on the mode decoupling matrix, as shown below:

[0080]

[0081] In the formula, This is the initial decoupled signal representation, equivalent to removing the main modal coupling interference in the time-frequency domain, and serves as the input for the next time-frequency transformation. The linear transformation process performed on the modal decoupling matrix is ​​equivalent to performing an orthogonalization operation in the feature space, which is a prerequisite for subsequent time-frequency invariant transformation and adaptive estimation.

[0082] The algorithm proposed in this application uses a Time-Frequency Invariant Transform (TFIT) operator to perform time-varying compensation and energy normalization on power line carrier signals by utilizing a coordinated modulation mechanism of time offset and frequency scaling. This maintains the stable structure of the signal in the time and frequency domain and eliminates time-varying drift caused by factors such as load switching and nonlinear harmonics.

[0083] In some embodiments of this application, step S103 may specifically include the following sub-steps:

[0084] Sub-step S31: Time-shift compensation is performed on the disturbance in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain the transformed signal;

[0085] Sub-step S32: Perform energy normalization on the transformed signal to obtain an energy-normalized signal (used to ensure that the energy scale of all modes is consistent, so that signals of different frequency bands are comparable in energy); the energy-normalized signal has steady-state characteristics.

[0086] To suppress non-stationary impulse noise, embodiments of this application employ a time-frequency invariant transform to achieve adaptive translation compensation for non-stationary disturbances in power line carriers.

[0087] In the embodiments of this application, time-shift compensation can be performed on the disturbance in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to ensure the consistency of the time-frequency structure of the target signal, so that the energy concentration area of ​​the carrier signal remains stable under the conditions of frequency drift or nonlinear disturbance, thereby achieving the purpose of time-frequency invariant transformation.

[0088] For example, a specific signal transformation can be shown below:

[0089]

[0090] In the formula, The signal after time-frequency invariant transformation is used to compensate for time-frequency drift and non-stationarity; α is the time compensation coefficient, mainly used to control the compensation intensity for time-domain drift; β is the frequency scale adjustment coefficient, mainly used to control the frequency sensitivity of the compensation. ω is the time shift parameter, mainly used for sliding time windows; ω is the angular frequency variable, mainly used for analyzing signal components in different frequency bands. This is a bit compensation term, mainly used to ensure phase consistency during time-shift transformation.

[0091] Optionally, time shift compensation for high-frequency / low-frequency disturbances can be achieved by controlling parameters α and β. The compensation strength indicated by parameter α can specifically determine the offset of the signal on the time axis, and the control compensation frequency sensitivity indicated by parameter β can specifically control the weight of signals in different frequency bands during time shift.

[0092] It should be noted that the settings of parameters α and β can be adaptively determined based on the local energy gradient of the signal in the time-frequency domain and the frequency band distribution of the disturbance. Specifically, α controls the compensation intensity for time-domain drift, and β controls the frequency sensitivity of the compensation. Together, they achieve differentiated time-shift correction for high-frequency transient disturbances and low-frequency slowly varying disturbances. For specific time-shift compensation strategies, parameters α and β can be set accordingly with the goal of improving time-frequency domain stability and disturbance decoupling accuracy. This application does not impose any limitations on this.

[0093] The transformed signal maintains a stable energy structure in the time-frequency domain, thus avoiding the stretching and drift effects of impulse interference or time-varying noise on signal characteristics. Furthermore, the transformed result can be energy normalized to ensure consistent energy scale across all modes and prevent certain strong noise components from dominating subsequent estimations.

[0094] For example, the energy normalization performed can be expressed as:

[0095]

[0096] In the formula, The energy-normalized signal can be used to ensure the comparability of signals in different frequency bands in terms of energy and reduce the problem of excessive weighting of high-frequency interference; the integral term is used to calculate the energy in the time domain and can be used to normalize the denominator.

[0097] The energy-normalized signal described above represents the extracted steady-state feature. The steady-state feature extraction using time-frequency invariant transform enables the carrier signal to maintain a robust time-frequency structure under frequency drift and nonlinear disturbances.

[0098] The energy-normalized signal is used as the input signal for subsequent estimation, which enables stability and scale consistency in the time and frequency domains, and is the core source of the algorithm's robustness.

[0099] This application further introduces a Bayesian adaptive weight estimation layer to perform probabilistic modeling and dynamic weight updates for various interference components. Through posterior inference and likelihood optimization, the algorithm can adaptively evaluate the impact of different interference types at each time-frequency point and generate corresponding soft mask functions, achieving accurate interference localization and amplitude suppression.

[0100] In some embodiments of this application, step S104 may specifically include the following sub-steps:

[0101] Sub-step S41 introduces a Bayesian adaptive weight estimation layer to perform probability modeling on each interference mode and calculate the degree of influence of different interference modes at each time frequency point;

[0102] Sub-step S42: Dynamically update the weights of each interference mode to obtain adaptive weights;

[0103] Sub-step S43: Generate the corresponding soft mask function based on the obtained adaptive weights to obtain the multimodal mask.

[0104] Adaptive mode weight optimization and iterative adjustment can achieve the highlighting and recovery of the target carrier signal, as well as closed-loop optimization of interference suppression.

[0105] The embodiments of this application introduce a Bayesian inference mechanism to perform probabilistic modeling and weighted estimation of the contribution of each type of interference.

[0106] First, probabilistic modeling can be performed for each interference mode. For example, the probabilistic modeling can be as follows:

[0107]

[0108] In the formula, The weighting coefficient for the i-th type of interference represents the degree of influence of this type of interference on the signal at the current time and frequency point. Given an observed signal, this is the posterior probability of the i-th type of interference, which can be used to quantify the likelihood of interference occurring.

[0109] The degree of influence of different interference modes at various time and frequency points can be specifically reflected by calculating the posterior probability. That is, by calculating the posterior probability... The algorithm can dynamically determine which type of interference has the strongest impact at each time frequency point.

[0110] There is a positive correlation between the degree of influence and the contribution ratio; that is, the greater the contribution ratio, the greater the degree of influence. The contribution ratio of each interference mode can be expressed as the weight of each interference mode.

[0111] Furthermore, the weights of each interfering mode can be dynamically updated to adjust the contribution ratio of each mode in the reconstruction. Specifically, the weights of each interfering mode can be determined based on the posterior probability calculated using Bayes' theorem, and are entirely dependent on the currently observed signal. .

[0112] For example, using Bayesian methods to determine posterior probabilities The dynamic update can be performed as follows:

[0113]

[0114] In the formula, The likelihood probability can be used to represent the probability of observing the current signal in the presence of the i-th type of interference; The prior probability of the disturbance can be derived from historical statistics or empirical models; k is an index variable used to traverse all M disturbance modes; This is the prior probability of traversing all disturbances (the kth one).

[0115] in, Historical statistical modeling can be used to model:

[0116]

[0117] This formula assumes that the disturbance amplitude follows a Gaussian distribution and uses an exponential function to establish a probability model. In the formula, Let be the average amplitude of the i-th type of interference, used to characterize its energy distribution center; Let be the variance of the i-th type of interference, used to describe its energy dispersion. This formula is a standard Bayesian posterior probability normalization form, used to estimate the interference contribution of each mode at a given time-frequency point. The algorithm can directly implement adaptive interference weight update based on this structure.

[0118] It should be noted that Bayesian update is itself an iterative optimization method. When processing data at each time frequency point, the algorithm incorporates prior knowledge. and current signal evidence An optimal weight estimate is calculated iteratively, thereby achieving adaptive optimization of the modal contribution ratio.

[0119] The soft masking function for the interference components can be expressed as:

[0120]

[0121] In the formula, For interference suppression masking, the closer the value is to 1, the more likely the frequency point is mainly the target signal, and the closer it is to 0, the more likely the interference component is dominant.

[0122] Generated adaptive weights and soft mask This will be a core variable in the subsequent optimization function, directly affecting the signal recovery effect. This step is equivalent to constructing a probabilistic interpretation layer, giving the algorithm statistically adaptive capabilities.

[0123] To achieve global optimization and energy balance, the algorithm proposed in this application also constructs a joint optimization objective function, simultaneously constraining signal reconstruction error, mask smoothness, and modal correlation. Optimal weights and mask parameters are obtained analytically, thus ensuring a balance between interference intensity and signal fidelity. In this application embodiment, by constructing a comprehensive objective function, the algorithm simultaneously considers multiple performance indicators such as signal reconstruction accuracy, interference suppression strength, and mask continuity. Analytically solving the constructed comprehensive objective function yields a closed-form expression for the optimal weights and mask, improving computational efficiency and ensuring mathematical interpretability. The essence of optimization is to minimize the impact of interference while maintaining the signal structure, providing optimal time-frequency mask parameters for final signal reconstruction.

[0124] In some embodiments of this application, a comprehensive objective function can be constructed based on the signal reconstruction term, the mask gradient in the time direction, and the modal coherence penalty term. Then, the partial derivatives in the comprehensive objective function can be set to zero to obtain the optimal weights. Finally, the optimal interference mask can be calculated based on the optimal weights.

[0125] The signal reconstruction term can use the signal time-frequency distribution function of each interference component in the time-frequency domain, the signal after energy normalization, and the weight coefficients of various interferences; the mask gradient in the time direction can use the soft mask of each interference component.

[0126] For example, the comprehensive objective function constructed by combining the three definitions of energy fidelity, coherence suppression, and masking smoothing can be:

[0127]

[0128] In the formula, The objective function is a global optimization function, mainly used to simultaneously minimize signal distortion, mask non-smoothing, and modal coherence. These are weighting factors used to control the influence intensity of the three parts: reconstruction error, smoothing constraint, and modal decoupling, respectively. This is a signal reconstruction term used to constrain the fidelity between the recovered signal and the observed signal; The mask gradient in the time direction is used to penalize abrupt changes and improve smoothness; This is a modal coherence penalty term used to constrain the correlation between different interference masks through trace operations.

[0129] The first term maintains signal reconstruction fidelity, the second term restricts mask smoothing, and the third term enables trace term smoothing. Further suppress modal coherence. Setting the partial derivatives to zero yields the optimal weights:

[0130]

[0131]

[0132] In the formula, To analyze the optimal weights, we can minimize right The partial derivative is obtained; As the complex conjugate transpose, it can be used to calculate the optimal linear projection between signals; It is a small constant and can be used to prevent the denominator from being zero; The optimal interference mask can be directly calculated from the optimal weights and is used for subsequent signal recovery.

[0133] At the output layer, the algorithm proposed in this application can adaptively superimpose various interference suppression results in the time-frequency domain through a multi-modal mask fusion mechanism, and achieve high-fidelity reconstruction of the target signal using inverse time-frequency transform. This fusion process takes into account both global energy constraints and local feature preservation, enabling stable recovery of the power line carrier signal even in high-noise environments.

[0134] In some embodiments of this application, after obtaining the optimal interference mask for various types of interference, the optimal interference mask can be adaptively superimposed, and the target signal can be obtained by inverse time-frequency transformation, thereby effectively suppressing interference while recovering the clear original signal to the maximum extent.

[0135] Specifically, interference masks selectively suppress interference components while preserving useful carrier characteristics at the spectral level. Various interference masks can be fused to form a comprehensive time-frequency domain filter. Specifically, it can be expressed as:

[0136]

[0137] In the formula, The fusion mask can represent the time-frequency weights after comprehensive suppression of all interference types. The product form can simultaneously reduce the superposition effect of multimodal interference.

[0138] Furthermore, the purified power line carrier signal can be obtained by multiplying the fusion mask with the input time-frequency signal and then performing an inverse transform, which is the recovered target signal.

[0139] Specifically, the purified time-frequency domain signal can be represented as:

[0140]

[0141] In the formula, The purified time-frequency domain signal mainly consists of the main carrier components remaining after mask filtering. For input time-frequency signals.

[0142] The signal is then reconstructed using an inverse transform, which can be expressed as:

[0143]

[0144] In the formula, The reconstructed time-domain signal is the purified carrier signal that is the final output of the algorithm, which can be used for subsequent demodulation or communication data recovery. It is the inverse Fourier transform kernel function, which can be used to return from the time-frequency domain to the time domain.

[0145] The above process enables a closed loop from interference modeling, separation, estimation to signal recovery, and is a key link in the algorithm output layer.

[0146] The embodiments of this application enable signal mixing modeling. Modal decoupling Time-frequency transformation stabilization Adaptive weight estimation Joint optimization solution The closed-loop processing chain of the time-domain reconstruction output ensures the continuity of the algorithm's mathematical logic and the completeness of its computational structure. This method can achieve accurate recovery and real-time separation of power line carrier signals under conditions of strong interference, time-varying channels, and non-Gaussian noise, significantly improving the reliability, robustness, and noise immunity of the carrier communication link. Through multimodal collaboration and adaptive inference mechanisms, this invention has significant engineering application value and patent innovation potential in areas such as distribution substation communication enhancement, smart meter box network stability improvement, and power Internet of Things signal analysis.

[0147] Optionally, based on the adaptive separation method for power line carrier signals provided in the embodiments of this application, an intelligent identification and dynamic control algorithm system for power line carrier communication signals can be constructed. This system can integrate a multi-layered algorithm structure including Dynamic Time Warping (DTW), attention mechanism, dynamic Bayesian network, fuzzy clustering, and closed-loop feedback optimization. It aims to solve problems such as timing mismatch, noise interference, state drift, and feature attenuation in existing carrier signal identification processes, achieving high-precision signal modeling, robust identification, and adaptive optimization. This algorithm can use a combination of data-driven and model-driven approaches as its core idea, forming progressively related logical links in the time domain, feature domain, and probability domain, constructing a closed-loop system from signal preprocessing to mode feedback.

[0148] For example, in the signal alignment stage, an improved dynamic time warping algorithm is used to dynamically register and reconstruct multi-channel power line carrier signals. This overcomes the problem of accuracy degradation caused by time drift and nonlinear mismatch in traditional algorithms, and realizes a unified mapping of carrier signals in the time domain and morphological domain, providing a stable input basis for subsequent feature extraction.

[0149] In the feature modeling stage, by introducing a fusion structure of multi-layer attention mechanism and dynamic Bayesian network, the key feature regions are focused and weighted through attention weight function to improve the discrimination performance of the model in complex electromagnetic interference environment. At the same time, dynamic Bayesian network constructs the evolution probability model of signal state in time series, enabling the system to predict and infer the dynamic changes of signal under non-stationary carrier characteristics, which significantly enhances the time-varying adaptability and noise robustness of the algorithm.

[0150] In the pattern recognition and clustering optimization stage, a fuzzy clustering algorithm is used to flexibly divide and identify the modes of the power line carrier signal. Furthermore, adaptive adjustment of fuzzy membership degrees enables dynamic updates of signal modes under different operating conditions, achieving continuous evolution and adaptive updates of the signal modes. Further, a closed-loop feedback mechanism is introduced at the global level to dynamically correct and adjust the parameters of the aforementioned model output. Gain adaptation and error gradient feedback strategies are used to ensure that the algorithm maintains convergence and stability under long-term operation and external disturbances.

[0151] This application's embodiments cover a complete innovative mechanism from signal modeling to feedback optimization, including a power line carrier signal timing alignment method based on dynamic time warping, a multi-dimensional feature adaptive weighting structure based on attention mechanisms, a signal state evolution modeling framework based on dynamic Bayesian networks, a flexible signal pattern partitioning and dynamic update strategy based on fuzzy clustering, and an adaptive control mechanism based on closed-loop feedback. The above modules achieve a strongly coupled mathematical link through parameter sharing and symbol mapping, forming a unified temporal logic and structural progression relationship. This application's embodiments do not impose limitations on this.

[0152] Through the coordinated operation of this technical system, the embodiments of this application can achieve accurate identification, dynamic prediction, and adaptive optimization control of power line carrier signals in complex noise and multi-source disturbance environments, significantly improving the robustness, identification accuracy, and real-time response performance of the communication system. It should be noted that this method not only has significant engineering application potential in power line carrier communication, transformer area signal identification, and power consumption pattern sensing, but also demonstrates high patent protection value and research innovation significance in the fields of smart power distribution and demand-side response; the embodiments of this application do not impose any limitations on this.

[0153] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of this application.

[0154] Reference Figure 2 The diagram shows a structural block diagram of an adaptive separation device for power line carrier signals provided in an embodiment of this application, which may specifically include the following modules:

[0155] Signal processing module 201 is used to sample the power line carrier signal and decompose the power line carrier signal to obtain a multimode signal;

[0156] The multi-mode disturbance decoupling module 202 is used to divide the interference components in the multi-mode signal into multiple frequency modes, and dynamically adjust each frequency mode based on the intramodal parameters to obtain the decoupled mode;

[0157] The time-frequency feature extraction module 203 is used to perform time-shift compensation on disturbances in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain steady-state features;

[0158] The mode weight update module 204 is used to adaptively calculate the mode energy of each interference mode in the steady-state characteristics and dynamically adjust the contribution ratio of each interference mode to obtain a multi-mode mask, thereby realizing the adaptive separation of the power line carrier signal.

[0159] In some embodiments of this application, the power line carrier signal is represented as a mixed signal via a multi-component mixing model; the signal processing module 201 may include the following sub-modules:

[0160] The multimodal signal generation submodule is used to perform time-frequency transformation on the mixed signal to obtain the signal time-frequency distribution function of each interference component in the time-frequency domain; the signal time-frequency distribution function is used to indicate the complex amplitude distribution of the signal in the time-frequency domain; the complex amplitude terms of each interference component in the signal time-frequency distribution function are combined and arranged to obtain the joint time-frequency distribution matrix; the joint time-frequency distribution matrix is ​​in matrix form of the multimodal signal.

[0161] In some embodiments of this application, the dynamic adjustment of each frequency mode based on the modal intrinsic parameters is achieved by constructing a modal decoupling matrix, where the modal intrinsic parameters are the coherence coefficients between each interference mode; the joint time-frequency distribution matrix is ​​used to represent the energy coupling and coherence relationship of different interference modes in the same time-frequency plane; the decoupled modes are represented based on the signal representation after preliminary decoupling; the multi-mode disturbance decoupling module 202 may include the following sub-modules:

[0162] The signal decoupling submodule is used to define the perturbation coherence function between different interference modes based on the energy coupling and coherence relationship between different interference modes in the same time-frequency plane, and obtain the coherence coefficient between each interference mode; based on the coherence coefficient between each interference mode, a mode decoupling matrix is ​​constructed; wherein, the mode decoupling matrix is ​​dynamically constructed at each time-frequency point; based on the mode decoupling matrix, the signal time-frequency distribution function of each interference component in the time-frequency domain is linearly transformed to obtain the signal representation after preliminary decoupling.

[0163] In some embodiments of this application, the time-frequency feature extraction module 203 may include the following sub-modules:

[0164] The steady-state feature extraction submodule is used to perform time-shift compensation on disturbances in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain the transformed signal. The time compensation coefficient is used to control the compensation intensity for time-domain drift, and the frequency scale adjustment coefficient is used to control the frequency sensitivity of the compensation. The transformed signal is then normalized to obtain the energy-normalized signal. The energy-normalized signal is the steady-state feature.

[0165] In some embodiments of this application, the contribution ratio of each interference mode is represented as the weight of each interference mode; the mode weight update module 204 may include the following sub-modules:

[0166] The multimodal masking submodule is used to introduce a Bayesian adaptive weight estimation layer to perform probabilistic modeling of each interference mode, calculate the influence of different interference modes at each time and frequency point, dynamically update the weights of each interference mode to obtain adaptive weights, and generate the corresponding soft masking function based on the obtained adaptive weights to obtain the multimodal mask.

[0167] In some embodiments of this application, the modal weight update module 204 may include the following sub-modules:

[0168] The modal weight optimization submodule is used to construct a comprehensive objective function based on the signal reconstruction term, the mask gradient in the time direction, and the modal coherence penalty term. The signal reconstruction term uses the signal time-frequency distribution function of each interference component in the time-frequency domain, the energy-normalized signal, and the weight coefficients of various interferences. The mask gradient in the time direction uses the soft mask of each interference component. The partial derivatives in the comprehensive objective function are set to zero to obtain the optimal weights. The optimal interference mask is calculated based on the optimal weights.

[0169] In some embodiments of this application, the apparatus provided in this application may further include the following modules:

[0170] The signal recovery module is used to adaptively superimpose the optimal interference mask and obtain the target signal using inverse time-frequency transformation.

[0171] In this embodiment, by sampling the power line carrier signal and decomposing it into a multimode signal, the interference components in the multimode signal are divided into multiple frequency modes. Based on the intrinsic parameters of each frequency mode, the decoupled modes are dynamically adjusted to obtain decoupled modes. Then, time-shift compensation can be performed on the disturbances in the decoupled modes based on the time compensation coefficient and the frequency scaling coefficient to obtain steady-state characteristics. Finally, the modal energy of each interference mode in the steady-state characteristics can be adaptively calculated, and the contribution ratio of each interference mode can be dynamically adjusted to obtain a multimode mask, thereby realizing the adaptive separation of the power line carrier signal. By dividing the interference components into multiple frequency modes and obtaining decoupled modes, multi-mode disturbances are decoupled, enabling high-level differentiation of interference from different sources in the modal domain. Furthermore, time-shift compensation is applied to the disturbances in the decoupled modes based on time compensation coefficients and frequency scale adjustment coefficients, achieving steady-state feature extraction of time-frequency invariant transformations. This allows the carrier signal to maintain a robust time-frequency structure under frequency drift and nonlinear disturbances. Moreover, adaptive mode weight optimization is achieved through adaptive calculation of modal energy and dynamic adjustment of the contribution ratio of each interference mode, realizing adaptive separation of the power line carrier signal and providing technical support for the subsequent prominent recovery of the target carrier signal. Therefore, the adaptive separation scheme for power line carrier signals provided in this application is applicable to the complex power line channel environment characterized by multi-mode nonlinear disturbances and strong time-varying characteristics.

[0172] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0173] This application also provides an electronic device, see embodiments thereof. Figure 3 The provided electronic device 300 includes a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and capable of running on the processor 320. When the computer program 311 is executed by the processor, it implements the various processes of the above-described adaptive separation method embodiment for power line carrier signals and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0174] This application also provides a computer-readable storage medium, see embodiments thereof. Figure 4 The computer-readable storage medium 400 provided stores a computer program 311. When the computer program 311 is executed by the processor, it implements the various processes of the above-described adaptive separation method embodiment for power line carrier signals and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0175] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0176] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of the embodiments of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or modules is not necessarily limited to those steps or modules explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. The division of modules in the embodiments of this application is merely a logical division; in actual applications, there may be other division methods. For example, multiple modules may be combined into or integrated into another system, or some features may be ignored or not performed. Additionally, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interface, and the indirect coupling or communication connection between modules may be electrical or other similar forms, none of which are limited in the embodiments of this application. Furthermore, the modules or sub-modules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed among multiple circuit modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of this application.

[0177] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0178] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0179] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, apparatuses, or modules, and may be electrical, mechanical, or other forms.

[0180] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0181] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0182] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0183] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0184] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0185] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes; these computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0186] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0187] Finally, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0188] The technical solutions provided in the embodiments of this application have been described in detail above. Specific examples have been used in the embodiments of this application to illustrate the principles and implementation methods of the embodiments of this application. The description of the above embodiments is only for the purpose of helping to understand the methods and core ideas of the embodiments of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the embodiments of this application. Therefore, the content of this specification should not be construed as a limitation on the embodiments of this application.

Claims

1. An adaptive separation method for power line carrier signals, characterized in that, The method includes: The power line carrier signal is sampled, and the power line carrier signal is decomposed to obtain a multimode signal; The interference components in the multimodal signal are divided into multiple frequency modes, and each frequency mode is dynamically adjusted based on the intramodal parameters to obtain the decoupled mode. Time-shift compensation is performed on the disturbances in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain steady-state characteristics; The modal energy of each interference mode in the steady-state characteristics is adaptively calculated, and the contribution ratio of each interference mode is dynamically adjusted to obtain a multimode mask, thereby realizing the adaptive separation of the power line carrier signal.

2. The method according to claim 1, characterized in that, The power line carrier signal is represented as a mixed signal via a multi-component mixing model; the decomposition of the power line carrier signal to obtain a multi-mode signal includes: The mixed signal is subjected to time-frequency transformation to obtain the signal time-frequency distribution function of each interference component in the time-frequency domain; the signal time-frequency distribution function is used to indicate the complex amplitude distribution of the signal in the time-frequency domain; The complex amplitude terms of each interference component in the time-frequency distribution function of the signal are combined to obtain a joint time-frequency distribution matrix; the joint time-frequency distribution matrix is ​​in matrix form of a multimode signal.

3. The method according to claim 2, characterized in that, The dynamic adjustment of each frequency mode based on the modal intrinsic parameters is achieved by constructing a modal decoupling matrix, where the modal intrinsic parameters are the coherence coefficients between each interference mode; the joint time-frequency distribution matrix is ​​used to represent the energy coupling and coherence relationship of different interference modes in the same time-frequency plane; the decoupling mode is reflected based on the signal representation after preliminary decoupling; The step of dividing the interference components in the multimodal signal into multiple frequency modes and dynamically adjusting each frequency mode based on the intramodal parameters to obtain decoupled modes includes: Based on the energy coupling and coherence relationship of the different interference modes in the same time-frequency plane, the perturbation coherence function between each interference mode is defined, and the coherence coefficient between each interference mode is obtained. Based on the coherence coefficients between the various interference modes, a mode decoupling matrix is ​​constructed; wherein, the mode decoupling matrix is ​​dynamically constructed at each time-frequency point; Based on the modal decoupling matrix, the signal time-frequency distribution function of each interference component in the time-frequency domain is linearly transformed to obtain the signal representation after preliminary decoupling.

4. The method according to claim 1, characterized in that, The step of performing time-shift compensation on disturbances in the decoupled mode based on time compensation coefficients and frequency scale adjustment coefficients to obtain steady-state characteristics includes: The disturbance in the decoupled mode is time-shifted and compensated based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain the transformed signal; wherein, the time compensation coefficient is used to control the compensation intensity for time-domain drift, and the frequency scale adjustment coefficient is used to control the frequency sensitivity of the compensation. The transformed signal is then normalized to obtain an energy-normalized signal; the energy-normalized signal exhibits steady-state characteristics.

5. The method according to claim 1, characterized in that, The contribution ratio of each interference mode is represented as the weight of each interference mode; the adaptive calculation of the modal energy of each interference mode in the steady-state feature and the dynamic adjustment of the contribution ratio of each interference mode to obtain a multimode mask include: A Bayesian adaptive weight estimation layer is introduced to perform probabilistic modeling of each interference mode and calculate the degree of influence of different interference modes at each time frequency point; The weights of each interference mode are dynamically updated to obtain adaptive weights; Based on the obtained adaptive weights, a corresponding soft mask function is generated to obtain the multimodal mask.

6. The method according to claim 5, characterized in that, The method further includes: A comprehensive objective function is constructed based on the signal reconstruction term, the mask gradient in the time direction, and the modal coherence penalty term. The signal reconstruction term adopts the signal time-frequency distribution function of each interference component in the time-frequency domain, the energy-normalized signal, and the weight coefficients of various interferences. The mask gradient in the time direction adopts the soft mask of each interference component. Setting the partial derivatives of the comprehensive objective function to zero yields the optimal weights; The optimal interference mask is calculated based on the optimal weights.

7. The method according to claim 6, characterized in that, Also includes: The optimal interference mask is adaptively superimposed, and the target signal is obtained by inverse time-frequency transformation.

8. An adaptive separation device for power line carrier signals, characterized in that, The device includes: The signal processing module is used to sample the power line carrier signal and decompose the power line carrier signal to obtain a multimode signal; A multi-modal disturbance decoupling module is used to divide the interference components in the multi-modal signal into multiple frequency modes, and dynamically adjust each frequency mode based on the intramodal parameters to obtain decoupled modes; The time-frequency feature extraction module is used to perform time-shift compensation on the disturbances in the decoupled mode based on the time compensation coefficient and the frequency scale adjustment coefficient to obtain steady-state features; The mode weight update module is used to adaptively calculate the modal energy of each interference mode in the steady-state characteristics and dynamically adjust the contribution ratio of each interference mode to obtain a multimode mask, thereby realizing the adaptive separation of the power line carrier signal.

9. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the adaptive separation method for power line carrier signals as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the adaptive separation method for power line carrier signals as described in any one of claims 1 to 7.

Images