A multi-source heterogeneous noise treatment method and device thereof

Abstract

The application discloses a multi-source heterogeneous noise treatment method and device, which separates input noise into a first noise and a first power line signal, analyzes noise characteristics, and performs noise hazard evaluation on the power line signal, compares the noise hazard evaluation with a hazard threshold to determine a hazard level, inputs the noise characteristics into a noise coefficient space for mapping to obtain a noise sparse feature vector, and performs sparse approximation on an existing noise characteristic sparse vector to determine a noise type; parameters stored in a noise collaborative treatment autonomous decision module for different noise hazard levels and categories are trained to obtain a preliminary treatment result, the preliminary treatment result is further adjusted according to existing uncertainty factors to obtain a final treatment scheme, the treatment scheme is sent to a noise treatment system, and corresponding noise treatment is performed. Through the above method, collaborative treatment autonomous decision is made based on noise information analysis and evaluation and historical treatment, so that the noise treatment system performs corresponding noise treatment according to the collaborative treatment autonomous decision.

Description

Technical Field

[0001] This invention relates to the field of multi-source heterogeneous noise, and in particular to a method and apparatus for controlling multi-source heterogeneous noise. Background Technology

[0002] With the development of science and technology, power line communication has become an indispensable and important technological advancement. Power line communication refers to a communication method that uses power grids to transmit various information such as data, voice, and images. It has advantages such as wide coverage and low cost, making it an important communication technology in low-voltage power supply areas. However, technological development has driven the development of new power systems, but it has also brought many unavoidable challenges: the large-scale integration of power electronic equipment into low-voltage power supply areas leads to a complex power line communication environment; and the different specifications and sampling frequencies of electronic equipment result in power line communication being subject to various types of power noise interference. Currently, existing power line noise control technologies struggle to effectively filter out multi-source heterogeneous noise.

[0003] Existing power line carrier multi-source heterogeneous noise testing methods often classify multi-source noise based on a single aspect such as information or energy, without quantifying the degree of harm of the noise. This results in poor accuracy in judging the noise category, making it impossible to accurately manage the noise and thus compromising the performance of power line communication. Summary of the Invention

[0004] This invention provides a method and apparatus for multi-source heterogeneous noise control, which can make autonomous decisions for collaborative multi-source noise control, accurately control noise, and ensure the performance of power line communication.

[0005] To address the aforementioned technical problems, this invention provides a method for controlling multi-source heterogeneous noise, comprising: Acquire power line signals in a power system and perform noise separation on the power line signals to obtain a first noise and a first power line signal after noise separation; Based on the signal-to-noise ratio of the first power line signal, preset power service value parameters, preset link importance parameters, noise filterability parameters, and a first noise characteristic, a hazard assessment is performed on the power line signal to obtain a hazard assessment result; wherein, the noise filterability parameters and the first noise characteristic are obtained by analyzing the first noise. Based on the preset sparse transformation matrix, the first noise feature is subjected to noise coefficient space mapping to obtain a sparse vector of noise features. The noise feature sparse vector is sparsely approximated with an existing noise feature sparse vector to determine the noise type of the first noise; The noise feature sparse vector, the hazard assessment result, and the noise type are input into a preset autonomous decision analysis mapping network, and the weight vector between the input neurons and the output neurons is adjusted to obtain the autonomous decision analysis result; wherein, the autonomous decision analysis result includes several noise control methods ranked according to the control priority; Based on the historical governance data corresponding to the first noise type, the autonomous decision analysis results are optimized to generate a first noise governance scheme; The first noise control plan is sent to the noise control system so that the noise control system can control and adjust the power lines according to the first noise control plan.

[0006] This invention proposes a multi-source heterogeneous noise control method. It separates input noise into a first noise source and a first power line signal after noise separation. Noise characteristics are analyzed, and the power line signal is assessed for noise hazard, compared with a hazard threshold to determine the hazard level. The noise characteristics are then mapped into a noise coefficient space to obtain a sparse noise feature vector. This vector is sparsely approximated with existing sparse noise feature vectors to determine the noise type. The method trains the system with parameters stored in a collaborative noise control autonomous decision-making module, considering different noise hazard levels, categories, and noise hazard levels, to obtain preliminary control results. Further adjustments are made based on existing uncertainties to obtain a final control scheme. This scheme is then sent to the noise control system for corresponding noise control measures. This invention achieves collaborative noise control autonomous decision-making based on noise information analysis and assessment and historical control practices, enabling the noise control system to implement corresponding noise control measures and ensuring power line communication performance.

[0007] As a preferred example, the step of conducting a hazard assessment of the power line signal based on the signal-to-noise ratio of the first power line signal, preset power service value parameters, preset link importance parameters, noise filterability parameters, and first noise characteristics, and obtaining a hazard assessment result, specifically involves: Substitute the signal-to-noise ratio of the first power line signal, the preset power service value parameter, the preset link importance parameter, the noise filterability parameter, and the first noise feature into the hazard assessment calculation formula to calculate the noise hazard score of the first power line signal. When the noise hazard score is less than a preset first threshold, the hazard assessment result is determined to be of low hazard level; when the noise hazard score is greater than or equal to the preset first threshold and less than a preset second threshold, the hazard assessment result is determined to be of medium hazard level; when the noise hazard score is greater than or equal to the preset second threshold, the hazard assessment result is determined to be of high hazard level.

[0008] This preferred example assesses the hazard of the power line signal and sets a threshold to classify the noise of the power signal into three levels: low hazard, medium hazard, and high hazard. The filterability of the noise is analyzed based on its different characteristics. By combining these characteristics with the power spectrum and energy spectrum of the noise to determine the hazard level, the higher the importance of the power line link where the noise is located, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the filterability of the noise, the greater the hazard to the power line signal and the more severe the noise hazard. Furthermore, the noise hazard level is divided into three levels based on two preset values ​​for further classification in subsequent noise assessment studies.

[0009] As a preferred example, the hazard assessment calculation formula is as follows: in, The weighting parameter for the power spectral density of the noise. The weighting parameter for the energy spectral density; The power line signal-to-noise ratio; The parameters representing the value of power services carried by power line signals; This is a parameter related to the importance of the link. This is a parameter for noise filterability. , This indicates that the noise can be completely filtered out. This indicates that the noise cannot be filtered out; The weights for the link signal-to-noise ratio. Weights for power business value parameters The weights of the link importance parameter, The weights for the power spectral density of the noise. The weights are for the energy spectral density.

[0010] This preferred example demonstrates, through setting a hazard assessment calculation formula, that the higher the value of the power service and the current importance of the power line link, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the noise filterability, the greater the hazard of the noise to the power line signal, and the more severe the noise hazard.

[0011] As a preferred example, the step of performing noise coefficient space mapping on the first noise feature according to a preset sparse transformation matrix to obtain a sparse vector of noise features specifically involves: The noise features are stored in a known matrix, the matrix is ​​mapped to the corresponding noise coefficient space, and the sparse mapping formula is used to solve for the noise feature sparse vector corresponding to the noise features.

[0012] This preferred example stores the noise features in the matrix of the sparse transform approximation module. In, and the matrix The sparse vector representation of each noise feature is obtained by mapping to the corresponding noise coefficient space, so that the sparse vector corresponding to each noise feature can be approximated and judged with the existing noise samples in the noise dictionary module.

[0013] As a preferred example, the sparse mapping formula is specifically as follows: Where C is a known matrix, A is a coefficient matrix, and N is the number of noise features based on the analysis. The first noise The The coefficients of the noise characteristics, where X is the first noise. The total eigenvalues, The first noise The A sparse vector of noise features.

[0014] This preferred example sets a sparse mapping formula to map noise features into sparse vectors, so that the sparse vectors corresponding to each noise feature can be approximated and judged with the existing noise samples in the noise dictionary module.

[0015] As a preferred example, the step of sparsely approximating the noise feature sparse vector with an existing noise feature sparse vector to determine the noise type of the first noise specifically involves: Obtain the existing sparse vectors of noise features corresponding to each noise type from the preset matrix; wherein, the preset matrix contains several types of noise feature vectors; The sparse approximation formula is used to solve the sparse approximation results between the noise feature sparse vector and each existing noise feature sparse vector. The noise type that minimizes the sparse approximation result is taken as the noise type of the first noise.

[0016] This preferred example approximates the noise feature with the existing noise samples in the noise dictionary module by comparing the sparse vector corresponding to each noise feature. The smaller the approximation result, the closer the noise is to the sample noise. A preset value is set to divide the noise judgment into two levels. When the approximation result is lower than the preset value, the noise can be judged to be of the same type as the sample noise. The type judgment is performed on unknown noise so that autonomous decisions can be made on the noise in the future.

[0017] As a preferred example, the present invention further includes the sparse approximation formula as follows: in, This represents the sparse approximation result between the noise feature sparse vector and the existing noise feature sparse vector. For existing noise The Each noise feature vector Given the existing noise base matrix, It is a 2-norm. This represents the mean square error.

[0018] This preferred example uses a sparse approximation formula to approximate the sparse vector corresponding to each noise feature with the existing noise samples in the noise dictionary module. If the approximation result... Less than the sparsity threshold At that time, the current input noise is determined. Noise characteristics and noise samples If the noise characteristics are of the same type, it can be determined that... and It is the same type of noise.

[0019] As a preferred example, adjusting the weight vector between the input neuron and the output neuron to obtain the autonomous decision analysis result specifically involves: The output neuron with the largest difference between the input vector and the weight vector is defined as the winning neuron, and the other output neurons between the input vector and the weight vector are defined as other neurons. Adjust the neighborhood function between the winning neuron and other neurons to regulate the weight vector between the input and output neurons, so that the distance between the input vector and the weight vector gradually converges; The governance result corresponding to the output neuron with the smallest distance between the input vector and the weight vector is taken as the autonomous decision analysis result.

[0020] This preferred example updates the weight vector by adjusting the neighborhood function of the winning neuron and other neurons. Through repeated adjustments, the neighborhood function gradually decreases each time, which also makes the autonomous decision-making result for the noise gradually stabilize, so that the noise control scheme approaches the autonomous decision-making result.

[0021] As a preferred example, the results of the autonomous decision analysis include a set of governance methods; The set of governance methods includes four methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. These methods are defined as a set in order of their governance priorities. ,in .

[0022] This preferred example sets four governance methods as a set of governance methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. The set is defined in the order of the above methods, which is sorted from smallest to largest according to the significance of the governance methods. By implementing the method of switching to the next method when the conditions are not met, the cost of the governance scheme can be reduced.

[0023] As a preferred example, the autonomous decision-making analysis results are optimized based on the historical governance data corresponding to the first noise type, specifically as follows: Based on existing historical data on noise type management, calculate the uncertainty of management for the first noise type. The first noise type is set as a preset category, and the current noise control results are adjusted according to the historical control methods of the preset category noise to determine the uncertainty of the preset category noise control. When the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold.

[0024] This preferred example calculates the governance uncertainty of the noise and existing noise based on historical governance characteristics. It starts with the method that has the least significant effect, setting a preset uncertainty threshold. When the uncertainty value calculated by this method exceeds the threshold, a better method is used. The methods, from least to most effective, are increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. When hierarchical copy multiplexing is used, if the uncertainty of this governance method still exceeds the uncertainty threshold, more transmission paths are added, and this process is repeated until the uncertainty of the governance method is less than the uncertainty threshold. By adjusting the governance scheme, a suitable governance solution for the noise is optimized.

[0025] As a preferred example, when the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold, specifically: When adopting a scheme to improve signal transmission power, the uncertainty of noise control for the preset category is calculated. If it exceeds the preset third threshold, the anti-interference coding mechanism scheme is changed by changing the formula. When using the switching anti-interference coding mechanism to manage noise, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path management scheme is changed by changing the formula. When adopting the switching transmission path management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the hierarchical copy multiple transmission management scheme is changed by changing the formula. When using a graded copy multiplex transmission management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path is increased by changing the formula until the uncertainty of noise management for the preset category is lower than the preset third threshold.

[0026] This preferred example compares the uncertainty of noise control of each method with a threshold, and selects the method with noise control uncertainty below the threshold as the noise control scheme for the power line. This allows the noise control scheme to be updated and compared with the data, and finally selects a more stable control scheme to achieve the best noise control and ensure the communication performance of the power line.

[0027] As a preferred example, the replacement formula is as follows: in, For hierarchical copy multiplexing The number of multiple transmission paths is initially set to 2.

[0028] This preferred example compares the uncertainties of governance solutions, and when governance uncertainty... Exceeding the uncertainty threshold of the method If the current method is already a hierarchical copy multiplexing method, then switch to a better method. If the problem persists, this method will continue to be used, but the number of transmission paths will be increased to further improve transmission performance, reduce noise impact, and enhance efficiency. The effectiveness of governance.

[0029] This invention provides a multi-source heterogeneous noise control device, including a noise separation module, a noise analysis module, a hazard assessment module, a noise dictionary module, a sparse transformation approximation module, a collaborative governance autonomous decision-making module, and a noise control system; The noise separation module is used to acquire power line signals in the power system and perform noise separation on the power line signals to obtain first noise and noise-separated first power line signals; The noise analysis module is used to analyze the first noise to obtain the noise filterability parameter and the first noise characteristics; The hazard assessment module is used to assess the hazard of the power line signal based on the signal-to-noise ratio of the first power line signal, preset power service value parameters, preset link importance parameters, noise filterability parameters, and first noise characteristics, and to obtain the hazard assessment result. The noise dictionary module includes a broadband noise sub-dictionary and a narrowband noise sub-dictionary, storing the spectral bandwidth range, periodic distribution characteristics, power spectral density characteristics, and noise mathematical modeling information of various broadband power line communication noises and narrowband power line communication noises, including colored background noise, burst impulse noise, periodic impulse noise, and so on. The sparse transformation approximation module is used to perform noise coefficient space mapping on the first noise feature according to the preset sparse transformation matrix to obtain a noise feature sparse vector; and to perform sparse approximation between the noise feature sparse vector and an existing noise feature sparse vector to determine the noise type of the first noise. The collaborative governance autonomous decision-making module is used to input the noise feature sparse vector, the hazard assessment result, and the noise type into a preset autonomous decision analysis mapping network, and adjust the weight vector between the input neurons and the output neurons to obtain the autonomous decision analysis result; wherein, the autonomous decision analysis result includes several noise governance methods sorted according to governance priority; based on the historical governance data corresponding to the first noise type, the autonomous decision analysis result is optimized to generate a first noise governance scheme; The noise control system is used to receive a first noise control plan and to control and adjust the power lines according to the first noise control plan.

[0030] This invention proposes a multi-source heterogeneous noise testing device. A noise separation module separates input noise into broadband and narrowband noise, which are then analyzed by a noise analysis module. A hazard assessment module evaluates the noise hazard level and compares it with a hazard threshold to determine the hazard level. The noise is further input to a sparse transformation approximation module, where it undergoes a sparse transformation approximation with the existing sparse vectors of noise features in a noise dictionary module. This mapping of features to the corresponding noise coefficient space determines the noise type. A collaborative governance autonomous decision-making module formulates a governance plan based on the analyzed information and the determined level, which is then implemented by a noise control system. Through the interaction of these modules, this invention achieves accurate noise control, ensuring the performance of power line communication.

[0031] As a preferred example, the hazard assessment module includes a first unit and a second unit; The first unit is used to input the signal-to-noise ratio of the first power line signal, the preset power service value parameter, the preset link importance parameter, the noise filterability parameter, and the first noise feature into the hazard assessment calculation formula to calculate the noise hazard score of the first power line signal. The second unit is used to determine the hazard assessment result as low hazard level when the noise hazard score is less than a preset first threshold; to determine the hazard assessment result as medium hazard level when the noise hazard score is greater than or equal to the preset first threshold and less than a preset second threshold; and to determine the hazard assessment result as high hazard level when the noise hazard score is greater than or equal to the preset second threshold.

[0032] This preferred example assesses the hazard of the power line signal and sets a threshold to classify the noise of the power signal into three levels: low hazard, medium hazard, and high hazard. The filterability of the noise is analyzed based on its different characteristics. By combining these characteristics with the power spectrum and energy spectrum of the noise to determine the hazard level, the higher the importance of the power line link where the noise is located, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the filterability of the noise, the greater the hazard to the power line signal and the more severe the noise hazard. Furthermore, the noise hazard level is divided into three levels based on two preset values ​​for further classification in subsequent noise assessment studies.

[0033] As a preferred example, the hazard assessment calculation formula is as follows: in, The weighting parameter for the power spectral density of the noise. The weighting parameter for the energy spectral density; The power line signal-to-noise ratio; The parameters representing the value of power services carried by power line signals; This is a parameter related to the importance of the link. This is a parameter for noise filterability. , This indicates that the noise can be completely filtered out. This indicates that the noise cannot be filtered out; The weights for the link signal-to-noise ratio. Weights for power business value parameters The weights of the link importance parameter, The weights for the power spectral density of the noise. The weights are for the energy spectral density.

[0034] This preferred example demonstrates, through setting a hazard assessment calculation formula, that the higher the value of the power service and the current importance of the power line link, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the noise filterability, the greater the hazard of the noise to the power line signal, and the more severe the noise hazard.

[0035] As a preferred example, the sparse transformation approximation module includes a third module; The third module is used to store the noise features in a known matrix, map the matrix to the corresponding noise coefficient space, and solve the sparse mapping formula to obtain the noise feature sparse vector corresponding to the noise features.

[0036] This preferred example stores the noise features in the matrix of the sparse transform approximation module. In, and the matrix The sparse vector representation of each noise feature is obtained by mapping to the corresponding noise coefficient space, so that the sparse vector corresponding to each noise feature can be approximated and judged with the existing noise samples in the noise dictionary module.

[0037] As a preferred example, the sparse mapping formula is specifically as follows: Where C is a known matrix, A is a coefficient matrix, and N is the number of noise features based on the analysis. The first noise The The coefficients of the noise characteristics, where X is the first noise. The total eigenvalues, The first noise The A sparse vector of noise features.

[0038] This preferred example sets a sparse mapping formula to map noise features into sparse vectors, so that the sparse vectors corresponding to each noise feature can be approximated and judged with the existing noise samples in the noise dictionary module.

[0039] As a preferred example, the sparse transformation approximation module further includes a fourth unit; The fourth unit is used to obtain the existing sparse vectors of noise features corresponding to each noise type from the preset matrix; wherein, the preset matrix contains several noise feature vectors. The sparse approximation formula is used to solve the sparse approximation results between the noise feature sparse vector and each existing noise feature sparse vector. The noise type that minimizes the sparse approximation result is taken as the noise type of the first noise.

[0040] This preferred example approximates the noise feature with the existing noise samples in the noise dictionary module by comparing the sparse vector corresponding to each noise feature. The smaller the approximation result, the closer the noise is to the sample noise. A preset value is set to divide the noise judgment into two levels. When the approximation result is lower than the preset value, the noise can be judged to be of the same type as the sample noise. The type judgment is performed on unknown noise so that autonomous decisions can be made on the noise in the future.

[0041] As a preferred example, the sparse approximation formula is specifically as follows: in, This represents the sparse approximation result between the noise feature sparse vector and the existing noise feature sparse vector. For existing noise The Each noise feature vector Given the existing noise base matrix, It is a 2-norm. This represents the mean square error.

[0042] This preferred example uses a sparse approximation formula to approximate the sparse vector corresponding to each noise feature with the existing noise samples in the noise dictionary module. If the approximation result... Less than the sparsity threshold At that time, the current input noise is determined. Noise characteristics and noise samples If the noise characteristics are of the same type, it can be determined that... and It is the same type of noise.

[0043] As a preferred example, the collaborative governance autonomous decision-making module includes a fifth unit, a sixth unit, and a seventh unit; The fifth unit is used to define the output neuron with the largest value between the input vector and the weight vector as the winning neuron, and to define other output neurons between the input vector and the weight vector as other neurons. The sixth unit is used to adjust the neighborhood function between the winning neuron and other neurons, so as to adjust the weight vector between the input neuron and the output neuron, so that the distance between the input vector and the weight vector gradually converges. The seventh unit is used to take the governance result corresponding to the output neuron with the smallest distance between the input vector and the weight vector as the autonomous decision analysis result.

[0044] This preferred example updates the weight vector by adjusting the neighborhood function of the winning neuron and other neurons. Through repeated adjustments, the neighborhood function gradually decreases each time, which also makes the autonomous decision-making result for the noise gradually stabilize, so that the noise control scheme approaches the autonomous decision-making result.

[0045] As a preferred example, the results of the autonomous decision analysis include a set of governance methods; The set of governance methods includes four methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. These methods are defined as a set in order of their governance priorities. ,in .

[0046] This preferred example sets four governance methods as a set of governance methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. The set is defined in the order of the above methods, which is sorted from smallest to largest according to the significance of the governance methods. By implementing the method of switching to the next method when the conditions are not met, the cost of the governance scheme can be reduced.

[0047] As a preferred example, the collaborative governance autonomous decision-making module further includes an eighth unit, a ninth unit, and a tenth unit; The eighth unit is used to calculate the uncertainty of the first noise type based on existing historical governance data of noise type governance. The first noise type is set as a preset category, and the current noise control results are adjusted according to the historical control methods of the preset category noise to determine the uncertainty of the preset category noise control. The tenth unit is used to change the treatment method until the uncertainty of the noise treatment of the preset category is lower than the preset third threshold when the uncertainty of the noise treatment of the preset category exceeds the preset third threshold.

[0048] This preferred example calculates the governance uncertainty of the noise and existing noise based on historical governance characteristics. It starts with the method that has the least significant effect, setting a preset uncertainty threshold. When the uncertainty value calculated by this method exceeds the threshold, a better method is used. The methods, from least to most effective, are increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. When hierarchical copy multiplexing is used, if the uncertainty of this governance method still exceeds the uncertainty threshold, more transmission paths are added, and this process is repeated until the uncertainty of the governance method is less than the uncertainty threshold. By adjusting the governance scheme, a suitable governance solution for the noise is optimized.

[0049] As a preferred example, when the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold, specifically: When adopting a scheme to improve signal transmission power, the uncertainty of noise control for the preset category is calculated. If it exceeds the preset third threshold, the anti-interference coding mechanism scheme is changed by changing the formula. When using the switching anti-interference coding mechanism to manage noise, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path management scheme is changed by changing the formula. When adopting the switching transmission path management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the hierarchical copy multiple transmission management scheme is changed by changing the formula. When using a graded copy multiplex transmission management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path is increased by changing the formula until the uncertainty of noise management for the preset category is lower than the preset third threshold.

[0050] This preferred example compares the uncertainty of noise control of each method with a threshold, and selects the method with noise control uncertainty below the threshold as the noise control scheme for the power line. This allows the noise control scheme to be updated and compared with the data, and finally selects a more stable control scheme to achieve the best noise control and ensure the communication performance of the power line.

[0051] As a preferred example, the replacement formula is as follows: in, For hierarchical copy multiplexing The number of multiple transmission paths is initially set to 2.

[0052] This preferred example compares the uncertainties of governance solutions, and when governance uncertainty... Exceeding the uncertainty threshold of the method If the current method is already a hierarchical copy multiplexing method, then switch to a better method. If the problem persists, this method will continue to be used, but the number of transmission paths will be increased to further improve transmission performance, reduce noise impact, and enhance efficiency. The effectiveness of governance. Attached Figure Description

[0053] Figure 1 This is a flowchart of a multi-source heterogeneous noise control method according to a certain embodiment of the present invention; Figure 2 This is a diagram of a multi-source heterogeneous noise control method apparatus according to a certain embodiment of the present invention; Figure 3 This is a flowchart of a multi-source heterogeneous noise control method according to another embodiment of the present invention; Figure 4 This is a diagram of a multi-source heterogeneous noise control device according to one embodiment of the present invention. Detailed Implementation

[0054] The technical solutions of this invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0055] The present invention provides a multi-source heterogeneous noise control method that can make autonomous decisions for collaborative multi-source noise control, make accurate noise control measures, and ensure power line communication performance.

[0056] Please see Figure 1 In one embodiment of the present invention, a method is provided. Figure 1 The flowchart shown represents a method for controlling multi-source heterogeneous noise. This method includes steps S1 to S7. The specific steps are as follows: S1. Acquire the power line signal in the power system and perform noise separation on the power line signal to obtain a first noise and a first power line signal after noise separation; Please see Figures 1 to 3 In this embodiment of the invention, adaptive filtering technology is employed. Based on the different mean values, correlation functions, and distribution functions of the broadband and narrowband signals, the filter coefficients are adaptively adjusted to separate the noise into a first noise component and a first power line signal. Specifically, the first noise component comprises a narrowband noise component and a broadband noise component. Noise analysis is performed by inputting the separated narrowband noise into a narrowband noise analysis module for analysis. Fourier transform and other methods are used to analyze the characteristics of the narrowband noise, such as single-frequency signal, narrow-band distribution, spectral peaks, power spectrum, and energy spectrum. Similarly, the separated broadband noise is input into a broadband noise analysis module, and Fourier transform and other methods are used to analyze the broadband noise's wideband pulse distribution, frequency domain distribution, power spectrum, and energy spectrum.

[0057] S2. Based on the signal-to-noise ratio of the first power line signal, preset power service value parameters, preset link importance parameters, noise filterability parameters, and first noise characteristics, a hazard assessment is performed on the power line signal to obtain a hazard assessment result; wherein, the noise filterability parameters and the first noise characteristics are obtained by analyzing the first noise. Please see Figures 1 to 3 In this embodiment of the invention, the power line signal after noise separation is input into a multi-source heterogeneous noise testing hazard assessment module to monitor the power line signal-to-noise ratio (SNR), analyze the value of the power service and the importance of the link based on the service type, and analyze the noise filterability based on different noise characteristics. The obtained power line signal-to-noise ratio, the value of the power service, the importance of the link, the noise filterability, and the noise power spectrum and energy spectrum obtained from the noise analysis are input into the multi-source heterogeneous noise testing hazard assessment module for hazard scoring.

[0058] The hazard assessment of the power line signal is performed based on the signal-to-noise ratio of the first power line signal, preset power service value parameters, preset link importance parameters, noise filterability parameters, and first noise characteristics to obtain a hazard assessment result, specifically as follows: Substitute the signal-to-noise ratio of the first power line signal, the preset power service value parameter, the preset link importance parameter, the noise filterability parameter, and the first noise feature into the hazard assessment calculation formula to calculate the noise hazard score of the first power line signal. When the noise hazard score is less than a preset first threshold, the hazard assessment result is determined to be of low hazard level; when the noise hazard score is greater than or equal to the preset first threshold and less than a preset second threshold, the hazard assessment result is determined to be of medium hazard level; when the noise hazard score is greater than or equal to the preset second threshold, the hazard assessment result is determined to be of high hazard level.

[0059] Please see Figures 1 to 3 In this embodiment of the invention, the power line signal is subjected to a hazard assessment, and a judgment threshold is set to classify the noise of the power signal into three levels: low hazard, medium hazard, and high hazard. The filterability of the noise is analyzed based on different noise characteristics. The hazard score is obtained by combining the above-obtained characteristics with the power spectrum, energy spectrum, and other characteristics of the noise. This yields the results showing that the higher the importance of the power line link where the noise is located, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the filterability of the noise, the greater the hazard to the power line signal and the more severe the noise hazard. The noise assessment and hazard level judgment can serve as the mapping network input in subsequent collaborative noise control, thereby enabling autonomous decision-making in noise control.

[0060] The hazard assessment calculation formula is as follows: in, The weighting parameter for the power spectral density of the noise. The weighting parameter for the energy spectral density; The power line signal-to-noise ratio; The parameters representing the value of power services carried by power line signals; This is a parameter related to the importance of the link. This is a parameter for noise filterability. , This indicates that the noise can be completely filtered out. This indicates that the noise cannot be filtered out; The weights for the link signal-to-noise ratio. Weights for power business value parameters The weights of the link importance parameter, The weights for the power spectral density of the noise. The weights are for the energy spectral density.

[0061] Please see Figures 1 to 3In this embodiment of the invention, by setting a hazard assessment calculation formula, it is shown that the higher the value of the power service and the importance of the current power line link, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the noise filterability, the greater the hazard of the noise to the power line signal and the more serious the noise hazard.

[0062] S3. Based on the preset sparse transformation matrix, perform noise coefficient space mapping on the first noise feature to obtain a sparse vector of noise features. Please see Figures 1 to 3 In this embodiment of the invention, the noise features extracted by the wide and narrow band noise analysis modules are stored in the matrix of the sparse transformation approximation module, and the matrix is ​​mapped to the corresponding noise coefficient space to obtain the sparse vector representation of each noise feature of the input noise.

[0063] The step of performing noise coefficient space mapping on the first noise feature according to the preset sparse transformation matrix to obtain the noise feature sparse vector is as follows: The noise features are stored in a known matrix, the matrix is ​​mapped to the corresponding noise coefficient space, and the sparse mapping formula is used to solve for the noise feature sparse vector corresponding to the noise features.

[0064] Please see Figures 1 to 3 In this embodiment of the invention, the noise features extracted by the wide and narrow band noise analysis modules are stored in the matrix of the sparse transformation approximation module, and the matrix is ​​mapped to the corresponding noise coefficient space to obtain the sparse vector representation of each noise feature of the input noise, so as to approximate the sparse vector corresponding to each noise feature with the existing noise samples in the noise dictionary module.

[0065] The sparse mapping formula is specifically as follows: Where C is a known matrix, A is a coefficient matrix, and N is the number of noise features based on the analysis. The first noise The The coefficients of the noise characteristics, where X is the first noise. The total eigenvalues, The first noise The A sparse vector of noise features.

[0066] Please see Figures 1 to 3 In this embodiment of the invention, by setting a sparse mapping formula, noise features are mapped to sparse vectors of noise features, so that the sparse vectors corresponding to each noise feature can be approximated and judged with the existing noise samples in the noise dictionary module.

[0067] S4. Perform sparse approximation between the noise feature sparse vector and the existing noise feature sparse vector to determine the noise type of the first noise; Please see Figures 1 to 3 In this embodiment of the invention, the mapping result is input into the noise dictionary module, and a sparse approximation judgment is performed with the existing noise samples to determine the noise type.

[0068] The step of sparsely approximating the noise feature sparse vector with an existing noise feature sparse vector to determine the noise type of the first noise specifically involves: Obtain the existing sparse vectors of noise features corresponding to each noise type from the preset matrix; wherein, the preset matrix contains several types of noise feature vectors; The sparse approximation formula is used to solve the sparse approximation results between the noise feature sparse vector and each existing noise feature sparse vector. The noise type that minimizes the sparse approximation result is taken as the noise type of the first noise.

[0069] Please see Figures 1 to 3 In this embodiment of the invention, the sparse vector corresponding to each noise feature is approximated by existing noise samples in the noise dictionary module. The smaller the approximation result, the closer the noise is to the sample noise. A preset value is set to divide the noise judgment into two levels. When the approximation result is lower than the preset value, the noise can be judged to be of the same type as the sample noise. Type judgment is performed on unknown noise to enable subsequent autonomous decision-making regarding the noise. Specifically, the sparse approximation result between the sparse vector of the noise feature of the first noise and the feature vector of the noise feature of the noise sample is defined as... , The smaller the value, the closer the first noise is to the noise sample.

[0070] The sparse approximation formula is specifically as follows: in, This represents the sparse approximation result between the noise feature sparse vector and the existing noise feature sparse vector. For existing noise The Each noise feature vector Given the existing noise base matrix, It is a 2-norm. This represents the mean square error.

[0071] Please see Figures 1 to 3 In this embodiment of the invention, the sparse vector corresponding to each noise feature is approximated by the existing noise samples in the noise dictionary module using a sparse approximation formula. If the approximation result is... Less than the sparsity threshold At that time, the current input noise is determined. Noise characteristics and noise samples If the noise characteristics are of the same type, it can be determined that... and It is the same type of noise.

[0072] S5. Input the noise feature sparse vector, the hazard assessment result, and the noise type into a preset autonomous decision analysis mapping network, and adjust the weight vector between the input neurons and the output neurons to obtain the autonomous decision analysis result; wherein, the autonomous decision analysis result includes several noise control methods ranked according to the control priority; Please see Figures 1 to 3 In this embodiment of the invention, the parameters stored in the noise collaborative governance autonomous decision-making module, such as different noise hazard levels and categories, are trained to obtain preliminary governance results.

[0073] The adjustment of the weight vector between the input and output neurons to obtain autonomous decision analysis results specifically involves: The output neuron with the largest difference between the input vector and the weight vector is defined as the winning neuron, and the other output neurons between the input vector and the weight vector are defined as other neurons. Adjust the neighborhood function between the winning neuron and other neurons to regulate the weight vector between the input and output neurons, so that the distance between the input vector and the weight vector gradually converges; The governance result corresponding to the output neuron with the smallest distance between the input vector and the weight vector is taken as the autonomous decision analysis result.

[0074] In this embodiment of the invention, the mapping network suitable for autonomous decision-making in noise control includes two layers: an input layer and an output layer. By adjusting the weight vectors between the input and output neurons, the distance between the input vector and the weight vector gradually converges. The control result corresponding to the output neuron with the smallest distance is the input noise. The results of the governance.

[0075] Define the number of neurons in the input and output layers of the mapping network as follows: and ( The input vector of the input layer is Then the distance between the input vector and the weight vector is: in, Corresponding input noise The The parameters are sent to the output layer. The weight vectors of each neuron are initially assigned small random numbers; the neuron with the largest distance between its input vector and weight vector is defined as the initial winning neuron. The weight vector is updated by adjusting the neighborhood function between the winning neuron and other neurons. The formula for updating the weight vector is as follows: in, and The input layer vectors at training times t and t+1 are respectively. With the output layer The weights of each neuron; For the winning neuron With neurons The neighborhood function gradually decreases with the increase of training times, and the autonomous decision-making results for input noise gradually stabilize.

[0076] Repeated training is performed, and the neighborhood function is set. Less than a given threshold At that point, the calculation terminates and the results of the autonomous decision-making analysis for noise control are obtained. The results of the autonomous decision-making analysis include a set of governance methods; The set of governance methods includes four methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. These methods are defined as a set in order of their governance priorities. ,in .

[0077] In this embodiment of the invention, four governance methods are set as a set of governance methods: increasing signal transmission power, switching anti-interference coding mechanism, switching transmission path, and hierarchical copy multiplexing. The above order defines the set, where the order is sorted from smallest to largest according to the significance of the governance method's effect. Then, the cost of the governance scheme can be reduced by implementing a method of switching to the next method when the condition is not met.

[0078] S6. Based on the historical governance data corresponding to the first noise type, optimize the autonomous decision analysis results and generate a first noise governance scheme; Please see Figures 1 to 3 In this embodiment of the invention, further adjustments to historical governance data are made based on existing uncertainties to obtain the final governance solution.

[0079] Based on the historical governance data corresponding to the first noise type, the autonomous decision analysis results are optimized, specifically as follows: Based on existing historical data on noise type management, calculate the uncertainty of management for the first noise type. The first noise type is set as a preset category, and the current noise control results are adjusted according to the historical control methods of the preset category noise to determine the uncertainty of the preset category noise control. When the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold.

[0080] Please see Figures 1 to 3 In this embodiment of the invention, the uncertainty of noise control and existing noise control is calculated based on historical control characteristics. The calculation begins with the method that has the least significant effect, with a preset uncertainty threshold. When the uncertainty value calculated by this method exceeds the uncertainty threshold, a better method is used. The methods, from least to most effective, are increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. When hierarchical copy multiplexing is used, if the uncertainty of this control method is still greater than the uncertainty threshold, more transmission paths are added, and this process is repeated until the uncertainty of the control method is less than the uncertainty threshold. By adjusting the control scheme, a suitable control scheme for the noise is optimized.

[0081] Governance decisions based on uncertainty perception, assuming the first noise Category Based on the noise collaborative governance autonomous decision-making module's stored data on... The current noise control results are adjusted based on factors such as historical post-control gain, control cost, and post-control transmission performance fluctuations of the noise type. The calculations are then performed relative to the input noise. same category Uncertainty in the governance of noise This corresponds to the fourth of the four governance methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. Each one can be represented as: in, For method right Average gain after historical noise mitigation For method right The average historical cost of noise remediation For method right The average value of transmission performance fluctuation after historical noise mitigation. In other words, the smaller the post-mutilation gain, the higher the mitigation cost, and the greater the transmission performance fluctuation after mitigation, the greater the uncertainty of the method's mitigation results for this type of noise.

[0082] When the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold, specifically: When adopting a scheme to improve signal transmission power, the uncertainty of noise control for the preset category is calculated. If it exceeds the preset third threshold, the anti-interference coding mechanism scheme is changed by changing the formula. When using the switching anti-interference coding mechanism to manage noise, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path management scheme is changed by changing the formula. When adopting the switching transmission path management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the hierarchical copy multiple transmission management scheme is changed by changing the formula. When using a graded copy multiplex transmission management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path is increased by changing the formula until the uncertainty of noise management for the preset category is lower than the preset third threshold.

[0083] Please see Figures 1 to 3 In this embodiment of the invention, by comparing the uncertainty of noise control of each method with a threshold, the method with noise control uncertainty lower than the threshold is selected as the noise control scheme for the power line. This allows the noise control scheme to be updated and compared with the data, and finally a more stable control scheme is selected to achieve the best noise control and ensure the communication performance of the power line.

[0084] The specific replacement formula is as follows: in, For hierarchical copy multiplexing The number of multiple transmission paths is initially set to 2.

[0085] In this embodiment of the invention, by comparing the uncertainty of the governance scheme, when the governance uncertainty... Exceeding the uncertainty threshold of the method If the current method is already a hierarchical copy multiplexing method, then switch to a better method. If the problem persists, this method will continue to be used, but the number of transmission paths will be increased to further improve transmission performance, reduce noise impact, and enhance efficiency. The effectiveness of governance.

[0086] S7. Send the first noise control scheme to the noise control system so that the noise control system can control and adjust the power line according to the first noise control scheme.

[0087] Please see Figures 1 to 3In this embodiment of the invention, the noise control scheme is sent to the noise control system and corresponding noise control is performed. This realizes the collaborative governance autonomous decision-making based on noise information analysis and evaluation and historical governance, enabling the noise control system to make corresponding noise control based on the collaborative governance autonomous decision-making, thus ensuring the power line communication performance.

[0088] Please see Figure 4 The present invention provides a multi-source heterogeneous noise control device. In one embodiment of the present invention, a multi-source heterogeneous noise control device diagram is also provided, including a noise separation module 1, a noise analysis module 2, a hazard assessment module 3, a noise dictionary module 4, a sparse transformation approximation module 5, a collaborative governance autonomous decision-making module 6, and a noise control system 7. The noise separation module 1 is used to acquire power line signals in the power system and perform noise separation on the power line signals to obtain first noise and noise-separated first power line signals; The noise analysis module 2 is used to analyze the first noise to obtain the noise filterability parameter and the first noise characteristics; The hazard assessment module 3 is used to assess the hazard of the power line signal based on the signal-to-noise ratio of the first power line signal, preset power service value parameters, preset link importance parameters, noise filterability parameters, and first noise characteristics, and to obtain the hazard assessment result. The noise dictionary module 4 includes a broadband noise sub-dictionary and a narrowband noise sub-dictionary, storing information such as the spectral bandwidth range, periodic distribution characteristics, power spectral density characteristics, and noise mathematical modeling of various broadband power line communication noises and narrowband power line communication noises, including colored background noise, burst impulse noise, and periodic impulse noise. The sparse transformation approximation module 5 is used to perform noise coefficient space mapping on the first noise feature according to the preset sparse transformation matrix to obtain a noise feature sparse vector; and to perform sparse approximation between the noise feature sparse vector and an existing noise feature sparse vector to determine the noise type of the first noise. The collaborative governance autonomous decision-making module 6 is used to input the noise feature sparse vector, the hazard assessment result, and the noise type into a preset autonomous decision analysis mapping network, and adjust the weight vector between the input neurons and the output neurons to obtain the autonomous decision analysis result; wherein, the autonomous decision analysis result includes several noise governance methods sorted according to governance priority; based on the historical governance data corresponding to the first noise type, the autonomous decision analysis result is optimized to generate a first noise governance scheme; The noise control system 7 is used to receive a first noise control plan and control and adjust the power lines according to the first noise control plan.

[0089] Please see Figure 4 In this embodiment of the invention, the input noise is separated into broadband noise and narrowband noise by a noise separation module, which are then analyzed by a noise analysis module. A hazard assessment module evaluates the noise hazard and compares it with a hazard threshold to determine the hazard level. The noise is then further input to a sparse transform approximation module, where it is approximated by a sparse transform with the existing sparse vectors of noise features in a noise dictionary module. This mapping of features to the corresponding noise coefficient space determines the noise type. A collaborative governance autonomous decision-making module formulates a governance plan based on the analyzed information and the determined hazard level, which is then implemented by the noise governance system. Through the interaction of these modules, this invention achieves accurate noise governance, ensuring the performance of power line communication.

[0090] The hazard assessment module 3 includes a first unit and a second unit; The first unit is used to input the signal-to-noise ratio of the first power line signal, the preset power service value parameter, the preset link importance parameter, the noise filterability parameter, and the first noise feature into the hazard assessment calculation formula to calculate the noise hazard score of the first power line signal. The second unit is used to determine the hazard assessment result as low hazard level when the noise hazard score is less than a preset first threshold; to determine the hazard assessment result as medium hazard level when the noise hazard score is greater than or equal to the preset first threshold and less than a preset second threshold; and to determine the hazard assessment result as high hazard level when the noise hazard score is greater than or equal to the preset second threshold.

[0091] Please see Figure 4 In this embodiment of the invention, the module assesses the hazard of the power line signal and sets a judgment threshold to classify the noise of the power signal into three levels: low hazard, medium hazard, and high hazard. It analyzes the filterability of the noise based on different noise characteristics. By combining the obtained characteristics with the power spectrum, energy spectrum, and other features of the noise to determine the hazard level, the higher the importance of the power line link where the noise is located, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the filterability of the noise, the greater the hazard to the power line signal and the more severe the noise hazard. Furthermore, the noise hazard level is divided into three levels based on two preset values ​​for further classification in subsequent noise assessment studies.

[0092] The hazard assessment calculation formula is as follows: in, The weighting parameter for the power spectral density of the noise. The weighting parameter for the energy spectral density; The power line signal-to-noise ratio; The parameters representing the value of power services carried by power line signals; This is a parameter related to the importance of the link. This is a parameter for noise filterability. , This indicates that the noise can be completely filtered out. This indicates that the noise cannot be filtered out; The weights for the link signal-to-noise ratio. Weights for power business value parameters The weights of the link importance parameter, The weights for the power spectral density of the noise. The weights are for the energy spectral density.

[0093] Please see Figure 4 In this embodiment of the invention, by setting a hazard assessment calculation formula, it is shown that the higher the value of the power service and the importance of the current power line link, the lower the signal-to-noise ratio of the power line signal, the higher the power and energy of the noise, and the worse the noise filterability, the greater the hazard of the noise to the power line signal and the more serious the noise hazard.

[0094] The sparse transformation approximation module 5 includes a third module; The third module is used to store the noise features in a known matrix, map the matrix to the corresponding noise coefficient space, and solve the sparse mapping formula to obtain the noise feature sparse vector corresponding to the noise features.

[0095] Please see Figure 4 In this embodiment of the invention, the module stores the noise features in the matrix of the sparse transform approximation module. In, and the matrix The sparse vector representation of each noise feature is obtained by mapping to the corresponding noise coefficient space, so that the sparse vector corresponding to each noise feature can be approximated and judged with the existing noise samples in the noise dictionary module.

[0096] The sparse mapping formula is specifically as follows: Where C is a known matrix, A is a coefficient matrix, and N is the number of noise features based on the analysis. The first noise The The coefficients of the noise characteristics, where X is the first noise. The total eigenvalues, The first noise The A sparse vector of noise features.

[0097] Please see Figure 4In this embodiment of the invention, by setting a sparse mapping formula, noise features are mapped to sparse vectors of noise features, so that the sparse vectors corresponding to each noise feature can be approximated and judged with the existing noise samples in the noise dictionary module.

[0098] The sparse transformation approximation module 5 also includes a fourth unit; The fourth unit is used to obtain the existing sparse vectors of noise features corresponding to each noise type from the preset matrix; wherein, the preset matrix contains several noise feature vectors. The sparse approximation formula is used to solve the sparse approximation results between the noise feature sparse vector and each existing noise feature sparse vector. The noise type that minimizes the sparse approximation result is taken as the noise type of the first noise.

[0099] Please see Figure 4 In this embodiment of the invention, the module approximates the sparse vector corresponding to each noise feature with the existing noise samples in the noise dictionary module. The smaller the approximation result, the closer the noise is to the sample noise. A preset value is set to divide the noise judgment into two levels. When the approximation result is lower than the preset value, it can be judged that the noise is of the same type as the sample noise. The module performs type judgment on unknown noise in order to make autonomous decisions on noise in the future.

[0100] The sparse approximation formula is specifically as follows: in, This represents the sparse approximation result between the noise feature sparse vector and the existing noise feature sparse vector. For existing noise The Each noise feature vector Given the existing noise base matrix, It is a 2-norm. This represents the mean square error.

[0101] Please see Figure 4 In this embodiment of the invention, the sparse vector corresponding to each noise feature is approximated by the existing noise samples in the noise dictionary module using a sparse approximation formula. If the approximation result is... Less than the sparsity threshold At that time, the current input noise is determined. Noise characteristics and noise samples If the noise characteristics are of the same type, it can be determined that... and It is the same type of noise.

[0102] The collaborative governance autonomous decision-making module 6 includes a fifth unit, a sixth unit, and a seventh unit; The fifth unit is used to define the output neuron with the largest value between the input vector and the weight vector as the winning neuron, and to define other output neurons between the input vector and the weight vector as other neurons. The sixth unit is used to adjust the neighborhood function between the winning neuron and other neurons, so as to adjust the weight vector between the input neuron and the output neuron, so that the distance between the input vector and the weight vector gradually converges. The seventh unit is used to take the governance result corresponding to the output neuron with the smallest distance between the input vector and the weight vector as the autonomous decision analysis result.

[0103] Please see Figure 4 In this embodiment of the invention, the module updates the weight vector by adjusting the neighborhood function of the winning neuron and other neurons. Through repeated adjustments, the neighborhood function gradually decreases each time, and the autonomous decision-making result for the noise gradually stabilizes, so that the noise control scheme approaches the autonomous decision-making result.

[0104] The results of the autonomous decision-making analysis include a set of governance methods; The set of governance methods includes four methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. These methods are defined as a set in order of their governance priorities. ,in .

[0105] Please see Figure 4 In this embodiment of the invention, four governance methods are set as a set of governance methods: increasing signal transmission power, switching anti-interference coding mechanism, switching transmission path, and hierarchical copy multiplexing. The above order defines the set, where the order is sorted from smallest to largest according to the significance of the governance method's effect. Thus, the cost of the governance scheme can be reduced by implementing a method of switching to the next method when the condition is not met.

[0106] The collaborative governance autonomous decision-making module 6 also includes an eighth unit, a ninth unit, and a tenth unit; The eighth unit is used to calculate the uncertainty of the first noise type based on existing historical governance data of noise type governance. The first noise type is set as a preset category, and the current noise control results are adjusted according to the historical control methods of the preset category noise to determine the uncertainty of the preset category noise control. The tenth unit is used to change the treatment method until the uncertainty of the noise treatment of the preset category is lower than the preset third threshold when the uncertainty of the noise treatment of the preset category exceeds the preset third threshold.

[0107] Please see Figure 4 In this embodiment of the invention, the module calculates the governance uncertainty of the noise and existing noise based on historical governance characteristics. It starts by using the method with the lowest effectiveness, setting a preset uncertainty threshold. When the uncertainty value calculated by this method exceeds the threshold, a better method is used. The effectiveness of these methods, from lowest to highest, includes increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. When hierarchical copy multiplexing is used, if the uncertainty of this governance method still exceeds the uncertainty threshold, more transmission paths are added, and this process is repeated until the uncertainty of the governance method is less than the uncertainty threshold. By adjusting the governance scheme, a suitable governance solution for the noise is optimized.

[0108] When the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold, specifically: When adopting a scheme to improve signal transmission power, the uncertainty of noise control for the preset category is calculated. If it exceeds the preset third threshold, the anti-interference coding mechanism scheme is changed by changing the formula. When using the switching anti-interference coding mechanism to manage noise, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path management scheme is changed by changing the formula. When adopting the switching transmission path management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the hierarchical copy multiple transmission management scheme is changed by changing the formula. When using a graded copy multiplex transmission management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path is increased by changing the formula until the uncertainty of noise management for the preset category is lower than the preset third threshold.

[0109] Please see Figure 4 In this embodiment of the invention, by comparing the uncertainty of noise control of each method with a threshold, the method with noise control uncertainty lower than the threshold is selected as the noise control scheme for the power line. This allows the noise control scheme to be updated and compared with the data, and finally a more stable control scheme is selected to achieve the best noise control and ensure the communication performance of the power line.

[0110] The specific replacement formula is as follows: in, For hierarchical copy multiplexing The number of multiple transmission paths is initially set to 2.

[0111] Please see Figure 4 In this embodiment of the invention, by comparing the uncertainties of the governance schemes, when the governance uncertainty is... Exceeding the uncertainty threshold of the method If the current method is already a hierarchical copy multiplexing method, then switch to a better method. If the problem persists, this method will continue to be used, but the number of transmission paths will be increased to further improve transmission performance, reduce noise impact, and enhance efficiency. The effectiveness of governance.

[0112] For a more detailed explanation of the working principles and processes of this system, please refer to the embodiments described above.

[0113] This invention provides a multi-source heterogeneous noise control method. It separates input noise into a first noise source and a first power line signal after noise separation. Noise characteristics are analyzed, and the power line signal is assessed for noise hazard, compared with a hazard threshold to determine the hazard level. The noise characteristics are then mapped into a noise coefficient space to obtain a sparse noise feature vector. This vector is sparsely approximated with existing sparse noise feature vectors to determine the noise type. The method trains the system with parameters stored in a noise collaborative governance autonomous decision-making module, considering different noise hazard levels, categories, and other factors, to obtain preliminary governance results. Further adjustments are made based on existing uncertainties to obtain a final governance scheme. This scheme is then sent to the noise control system for corresponding noise control measures. This invention achieves collaborative governance autonomous decision-making based on noise information analysis and assessment, and historical governance data. This enables the noise control system to implement corresponding noise control measures based on these autonomous decisions, ensuring the performance of power line communication.

[0114] Simultaneously, this invention also proposes a multi-source heterogeneous noise testing device. A noise separation module separates the input noise into broadband and narrowband noise, which are then analyzed by a noise analysis module. A hazard assessment module evaluates the noise hazard and compares it with a hazard threshold to determine the hazard level. The noise is further input to a sparse transformation approximation module, where it undergoes a sparse transformation approximation with the existing sparse vectors of noise features in a noise dictionary module. This mapping of features to the corresponding noise coefficient space determines the noise type. A collaborative governance autonomous decision-making module formulates a governance plan based on the analyzed information and the determined level, which is then implemented by the noise governance system. Through the interaction of these modules, this invention achieves accurate noise governance, ensuring the performance of power line communication.

[0115] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for controlling multi-source heterogeneous noise, characterized in that, include: Acquire power line signals in a power system and perform noise separation on the power line signals to obtain a first noise and a first power line signal after noise separation; Substitute the signal-to-noise ratio of the first power line signal, the preset power service value parameter, the preset link importance parameter, the noise filterability parameter, and the first noise feature into the hazard assessment calculation formula to calculate the noise hazard score of the first power line signal. The hazard assessment calculation formula is as follows: in, The weighting parameter for the power spectral density of the noise. The weighting parameter for the energy spectral density; The power line signal-to-noise ratio; The parameters representing the value of power services carried by power line signals; This is a parameter related to the importance of the link. This is a parameter for noise filterability. , This indicates that the noise can be completely filtered out. This indicates that the noise cannot be filtered out; The weights for the link signal-to-noise ratio. Weights for power business value parameters The weights for the link importance parameter, The weights for the power spectral density of the noise. The weights for the energy spectral density; When the noise hazard score is less than a preset first threshold, the hazard assessment result is determined to be of low hazard level; when the noise hazard score is greater than or equal to the preset first threshold and less than a preset second threshold, the hazard assessment result is determined to be of medium hazard level; when the noise hazard score is greater than or equal to the preset second threshold, the hazard assessment result is determined to be of high hazard level; wherein, the noise filterability parameter and the first noise feature are obtained by analysis based on the first noise; Based on the preset sparse transformation matrix, the first noise feature is subjected to noise coefficient space mapping to obtain a sparse vector of noise features. The noise feature sparse vector is sparsely approximated with an existing noise feature sparse vector to determine the noise type of the first noise; The noise feature sparse vector, the hazard assessment result, and the noise type are input into a preset autonomous decision analysis mapping network, and the weight vector between the input neurons and the output neurons is adjusted to obtain the autonomous decision analysis result; wherein, the autonomous decision analysis result includes several noise control methods ranked according to the control priority; Based on the historical governance data corresponding to the first noise type, the autonomous decision analysis results are optimized to generate a first noise governance scheme; The first noise control scheme is sent to the noise control system so that the noise control system can control and adjust the power lines according to the first noise control scheme.

2. The method for controlling multi-source heterogeneous noise according to claim 1, characterized in that, The step of mapping the first noise feature to a noise coefficient space according to a preset sparse transformation matrix to obtain a sparse vector of noise features is as follows: The noise features are stored in a known matrix, the matrix is ​​mapped to the corresponding noise coefficient space, and the sparse mapping formula is used to solve for the noise feature sparse vector corresponding to the noise features.

3. The method for controlling multi-source heterogeneous noise according to claim 2, characterized in that, The sparse mapping formula is specifically as follows: Where C is a known matrix, A is a coefficient matrix, and N is the number of noise features based on the analysis. The first noise The The coefficients of the noise characteristics, where X is the first noise. The total eigenvalues, The first noise The A sparse vector of noise features.

4. The method for controlling multi-source heterogeneous noise according to claim 1, characterized in that, The step of sparsely approximating the noise feature sparse vector with an existing noise feature sparse vector to determine the noise type of the first noise specifically involves: Obtain the existing sparse vectors of noise features corresponding to each noise type from the preset matrix; wherein, the preset matrix contains several types of noise feature vectors; The sparse approximation formula is used to solve the sparse approximation results between the noise feature sparse vector and each existing noise feature sparse vector. The noise type that minimizes the sparse approximation result is taken as the noise type of the first noise.

5. The method for controlling multi-source heterogeneous noise according to claim 4, characterized in that, The sparse approximation formula is specifically as follows: in, This represents the sparse approximation result between the noise feature sparse vector and the existing noise feature sparse vector. For existing noise The Each noise feature vector Given the existing noise base matrix, It is a norm 2. Mean square error, The first noise The A sparse vector of noise features, where N is the number of noise features after analysis.

6. The method for controlling multi-source heterogeneous noise according to claim 1, characterized in that, The adjustment of the weight vector between the input and output neurons to obtain autonomous decision analysis results specifically involves: The output neuron with the largest difference between the input vector and the weight vector is defined as the winning neuron, and the other output neurons between the input vector and the weight vector are defined as other neurons. Adjust the neighborhood function between the winning neuron and other neurons to regulate the weight vector between the input and output neurons, so that the distance between the input vector and the weight vector gradually converges; The governance result corresponding to the output neuron with the smallest distance between the input vector and the weight vector is taken as the autonomous decision analysis result.

7. The method for controlling multi-source heterogeneous noise according to claim 6, characterized in that, The results of the autonomous decision-making analysis include a set of governance methods; The set of governance methods includes four methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. These methods are defined as a set in order of their governance priorities. ,in .

8. A method for controlling multi-source heterogeneous noise according to claim 1, characterized in that, Based on the historical governance data corresponding to the first noise type, the autonomous decision analysis results are optimized, specifically as follows: Based on existing historical data on noise type management, calculate the uncertainty of management for the first noise type. The first noise type is set as a preset category, and the current noise control results are adjusted according to the historical control methods of the preset category noise to determine the uncertainty of the preset category noise control. When the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold.

9. A method for controlling multi-source heterogeneous noise according to claim 8, characterized in that, When the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold, specifically: When adopting a scheme to improve signal transmission power, the uncertainty of noise control for the preset category is calculated. If it exceeds the preset third threshold, the anti-interference coding mechanism scheme is changed by changing the formula. When using the switching anti-interference coding mechanism to manage noise, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path management scheme is changed by changing the formula. When adopting the switching transmission path management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the hierarchical copy multiple transmission management scheme is changed by changing the formula. When using a graded copy multiplex transmission management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path is increased by changing the formula until the uncertainty of noise management for the preset category is lower than the preset third threshold.

10. A method for controlling multi-source heterogeneous noise according to claim 9, characterized in that, The specific replacement formula is as follows: in, For hierarchical copy multiplexing The number of multi-path transmission paths is initially set to 2; l is the governance mode number, l=1,2,3,4; L is the total number of governance modes; The governance uncertainty of adopting the l-th governance approach; Let be the uncertainty threshold for the l-th governance method.

11. A multi-source heterogeneous noise control device, characterized in that, It includes a noise separation module, a noise analysis module, a hazard assessment module, a noise dictionary module, a sparse transformation approximation module, a collaborative governance autonomous decision-making module, and a noise control system; The noise separation module is used to acquire power line signals in the power system and perform noise separation on the power line signals to obtain first noise and noise-separated first power line signals; The noise analysis module is used to analyze the first noise to obtain the noise filterability parameter and the first noise characteristics; The hazard assessment module includes a first unit and a second unit; The first unit is used to input the signal-to-noise ratio of the first power line signal, the preset power service value parameter, the preset link importance parameter, the noise filterability parameter, and the first noise feature into the hazard assessment calculation formula to calculate the noise hazard score of the first power line signal. The hazard assessment calculation formula is as follows: in, The weighting parameter for the power spectral density of the noise. The weighting parameter for the energy spectral density; The power line signal-to-noise ratio; The parameters representing the value of power services carried by power line signals; This is a parameter related to the importance of the link. This is a parameter for noise filterability. , This indicates that the noise can be completely filtered out. This indicates that the noise cannot be filtered out; The weights for the link signal-to-noise ratio. Weights for power business value parameters The weights for the link importance parameter, The weights for the power spectral density of the noise. The weights for the energy spectral density; The second unit is used to determine the hazard assessment result as low hazard level when the noise hazard score is less than a preset first threshold; to determine the hazard assessment result as medium hazard level when the noise hazard score is greater than or equal to the preset first threshold and less than a preset second threshold; and to determine the hazard assessment result as high hazard level when the noise hazard score is greater than or equal to the preset second threshold. The noise dictionary module includes a broadband noise sub-dictionary and a narrowband noise sub-dictionary, storing the spectral bandwidth range, periodic distribution characteristics, power spectral density characteristics, and noise mathematical modeling information of various broadband power line communication noises and narrowband power line communication noises, including colored background noise, burst impulse noise, periodic impulse noise, and so on. The sparse transformation approximation module is used to perform noise coefficient space mapping on the first noise feature according to the preset sparse transformation matrix to obtain a noise feature sparse vector; and to perform sparse approximation between the noise feature sparse vector and an existing noise feature sparse vector to determine the noise type of the first noise. The collaborative governance autonomous decision-making module is used to input the noise feature sparse vector, the hazard assessment result, and the noise type into a preset autonomous decision analysis mapping network, and adjust the weight vector between the input neurons and the output neurons to obtain the autonomous decision analysis result; wherein, the autonomous decision analysis result includes several noise governance methods sorted according to governance priority; based on the historical governance data corresponding to the first noise type, the autonomous decision analysis result is optimized to generate a first noise governance scheme; The noise control system is used to receive the first noise control plan and control and adjust the power lines according to the first noise control plan.

12. The multi-source heterogeneous noise control device according to claim 11, characterized in that, The sparse transformation approximation module includes a third module; The third module is used to store the noise features in a known matrix, map the matrix to the corresponding noise coefficient space, and solve the sparse mapping formula to obtain the noise feature sparse vector corresponding to the noise features.

13. The multi-source heterogeneous noise control device according to claim 12, characterized in that, The sparse mapping formula is specifically as follows: Where C is a known matrix, A is a coefficient matrix, and N is the number of noise features based on the analysis. The first noise The The coefficients of the noise characteristics, where X is the first noise. The total eigenvalues, The first noise The A sparse vector of noise features.

14. The multi-source heterogeneous noise control device according to claim 11, characterized in that, The sparse transformation approximation module also includes a fourth unit; The fourth unit is used to obtain the existing sparse vectors of noise features corresponding to each noise type from the preset matrix; wherein, the preset matrix contains several noise feature vectors. The sparse approximation formula is used to solve the sparse approximation results between the noise feature sparse vector and each existing noise feature sparse vector. The noise type that minimizes the sparse approximation result is taken as the noise type of the first noise.

15. The multi-source heterogeneous noise control device according to claim 14, characterized in that, The sparse approximation formula is specifically as follows: in, This represents the sparse approximation result between the noise feature sparse vector and the existing noise feature sparse vector. For existing noise The Each noise feature vector Given the existing noise base matrix, It is a norm 2. Mean square error, The first noise The A sparse vector of noise features, where N is the number of noise features after analysis.

16. The multi-source heterogeneous noise control device according to claim 11, characterized in that, The collaborative governance autonomous decision-making module includes a fifth unit, a sixth unit, and a seventh unit; The fifth unit is used to define the output neuron with the largest value between the input vector and the weight vector as the winning neuron, and to define other output neurons between the input vector and the weight vector as other neurons. The sixth unit is used to adjust the neighborhood function between the winning neuron and other neurons, so as to adjust the weight vector between the input neuron and the output neuron, so that the distance between the input vector and the weight vector gradually converges. The seventh unit is used to take the governance result corresponding to the output neuron with the smallest distance between the input vector and the weight vector as the autonomous decision analysis result.

17. The multi-source heterogeneous noise control device according to claim 16, characterized in that, The results of the autonomous decision-making analysis include a set of governance methods; The set of governance methods includes four methods: increasing signal transmission power, switching anti-interference coding mechanisms, switching transmission paths, and hierarchical copy multiplexing. These methods are defined as a set in order of their governance priorities. ,in .

18. The multi-source heterogeneous noise control device according to claim 11, characterized in that, The collaborative governance autonomous decision-making module also includes an eighth unit, a ninth unit, and a tenth unit; The eighth unit is used to calculate the uncertainty of the first noise type based on existing historical governance data of noise type governance. The first noise type is set as a preset category, and the current noise control results are adjusted according to the historical control methods of the preset category noise to determine the uncertainty of the preset category noise control. The tenth unit is used to change the treatment method until the uncertainty of the noise treatment of the preset category is lower than the preset third threshold when the uncertainty of the noise treatment of the preset category exceeds the preset third threshold.

19. The multi-source heterogeneous noise control device according to claim 18, characterized in that, When the uncertainty of noise control for the preset category exceeds a preset third threshold, the control method is changed until the uncertainty of noise control for the preset category is lower than the preset third threshold, specifically: When adopting a scheme to improve signal transmission power, the uncertainty of noise control for the preset category is calculated. If it exceeds the preset third threshold, the anti-interference coding mechanism scheme is changed by changing the formula. When using the switching anti-interference coding mechanism to manage noise, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path management scheme is changed by changing the formula. When adopting the switching transmission path management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the hierarchical copy multiple transmission management scheme is changed by changing the formula. When using a graded copy multiplex transmission management scheme, the uncertainty of noise management for the preset category is calculated. If it exceeds the preset third threshold, the transmission path is increased by changing the formula until the uncertainty of noise management for the preset category is lower than the preset third threshold.

20. The multi-source heterogeneous noise control device according to claim 19, characterized in that, The specific replacement formula is as follows: in, For hierarchical copy multiplexing The number of multi-path transmission paths is initially set to 2; l is the governance mode number, l=1,2,3,4; L is the total number of governance modes; The governance uncertainty of adopting the l-th governance approach; Let be the uncertainty threshold for the l-th governance method.

Images