Pulse noise suppression method based on gain control

Abstract

The invention discloses an impulse noise suppression method based on gain control, which comprises the following steps: S1, automatically monitoring an input original signal, and monitoring the change of the original signal in real time; s2, monitoring the intensity of the output signal through a feedback loop, and comparing the measured output signal characteristic with the original signal characteristic to obtain an error signal, namely a pulse noise signal; s3, processing the error signal by using a self-adaptive algorithm, providing an automatically adjusted gain parameter for a gain control module, and dynamically adjusting the gain; and S4, carrying out real-time monitoring and feedback on the output signal, and adjusting controller parameters to carry out real-time correction on the output signal. By setting a feedback loop and a self-adaptive algorithm, pulse noise is continuously corrected, and the quality and reliability of signal transmission are guaranteed to the maximum extent.

Description

Technical Field

[0001] The invention relates to the field of communications, and in particular to an impulse noise suppression method based on gain control. Background Art

[0002] In power line communication PLC, since data is transmitted through wires, it is susceptible to pulse noise from switching transients, power supply devices, electric dimmers, etc. In particular, asynchronous pulse noise, which is generated by the switching transients of the power line, has random and unpredictable characteristics, and the level value is usually very high.

[0003] In the prior art, the time-frequency peak filtering (TFPF) method is usually used to modulate the noisy signal into the instantaneous frequency of the analytical signal, and use the peak value of the Wigner-Ville distribution to estimate the instantaneous frequency, thereby restoring the effective signal. However, this method may produce certain deviations when processing nonlinear signals, especially at locations with strong nonlinearity such as the peaks and troughs of the signal. Summary of the invention

[0004] The present invention aims to provide a method for suppressing impulse noise based on gain control to solve the above-mentioned problem.

[0005] The present invention is achieved through the following technical solutions: A method for suppressing impulse noise based on gain control, comprising: S1: In the circuit system of PLC communication, the input original signal is automatically monitored and the changes of the original signal are monitored in real time; S2: Monitor the strength of the output signal through a feedback loop, compare the measured output signal characteristics with the original signal characteristics, and obtain an error signal; S3: Use an adaptive algorithm to process the error signal and provide an automatically adjusted gain parameter for the gain control module to achieve dynamic gain adjustment; S4: Monitor and feedback the output signal in real time, so that the gain control module can correct the output signal in real time.

[0006] The present invention provides a feedback loop and an adaptive algorithm to continuously correct the impulse noise, thereby ensuring the quality and reliability of signal transmission to the greatest extent.

[0007] Preferably, automatically monitoring the original signal includes the following: The amplitude of the original signal is measured using a root mean square (RMS) detector. First, the original signal passes through a square law device, such as a diode, to convert it into a square signal proportional to the amplitude of the original signal. Next, the square signal is averaged through a low-pass filter to obtain the average level over a period of time. This average level is the root mean square value of the original signal. Finally, a square root extractor is used to convert the average value into an RMS value.

[0008] The present invention mainly targets the problem of pulse noise in PLC signal transmission, where PLC signal transmission accuracy is high. Therefore, the present invention uses a root mean square (RMS) detector to measure the signal amplitude, which has the advantages of high accuracy, wide dynamic range, temperature stability, low power consumption, simplified design, and real-time response. These characteristics make the RMS detector an ideal power measurement tool for solving the technical problems of the present invention, especially in the accurate measurement of complex modulated signals, which can significantly reduce the system calibration requirements.

[0009] Preferably, the adaptive algorithm is specifically a radial basis function neural network algorithm, which specifically includes: S3.1: Extract the features of the error signal, including signal amplitude, phase, rise time, fall time, duration, and energy; S3.2: Use an unsupervised learning algorithm to determine the RBF (radial basis function) center point and use the center point as the neuron center of the network hidden layer; S3.3: Calculate the variance of each RBF center point based on the maximum distance between the center points; S3.4: Obtain the weights from the hidden layer to the output layer by the least squares method; S3.5: Dynamically adjust the signal gain by modifying the weight of the output layer; S3.6: The adjusted signal is output through the output layer, completing the gain adjustment process.

[0010] The present invention uses a radial basis function neural network algorithm to perform adaptive gain adjustment on impulse noise. Since RBF does not need to perform nonlinear mapping through weight connections, the learning speed of the impulse noise gain amplitude is accelerated, especially when applied to PLC real-time communication, the delay problem is well solved. In addition, the hidden layer output of the RBF network is linear with respect to the adjustable parameters, which means that the weight of the network can be directly obtained by solving a linear equation group, thereby accelerating the learning speed and avoiding the local minimum problem.

[0011] Preferably, the unsupervised learning algorithm is specifically a K-Means clustering algorithm, which specifically includes: Randomly select K data points as the centers of the initial clusters ; S3.4.1: For each data point , calculate its distance from all cluster centers and assign it to the nearest cluster center to form K clusters ; S3.4.2: Distance is calculated using Euclidean distance in, is the characteristic of the signal. Features, is the cluster center No. Features S3.4.3: Recalculate the cluster center of each cluster and the mean of all points in the cluster: ,in It is a cluster The number of data points in ; S3.4.4: Check whether the cluster center has changed. If not or the predetermined number of iterations has been reached, the algorithm ends; otherwise, return to step S3.4.2; S3.4.5: When the change of cluster center in consecutive iterations is less than the set threshold or reaches the maximum number of iterations, the algorithm terminates.

[0012] Since the determination of the RBF center point is the core of the radial basis function neural network algorithm, this application uses the K-Means clustering algorithm to determine the RBF center point. The K-Means algorithm can help the RBF network optimize the network structure and avoid overfitting while ensuring the accuracy requirements. In addition, K-Means clustering can identify representative points in the data as center points, reduce the number of nodes in the hidden layer of the RBF network, and simplify the network structure.

[0013] Preferably, the real-time correction of the output signal in S4 specifically includes: S4.1: Set the original input signal to The processed output signal is ; S4.2: Processing raw signals using recurrent neural networks With output signal The amplitude and phase differences between S4.3: The loss function is defined as follows: in, : The original signal can be a time series signal or any form of signal, where Indicates the signal samples.

[0014] : Output signal, which is the output signal after system processing (such as neural network), where Indicates the first samples.

[0015] : Original signal No. The amplitude of each sample, that is, the size or strength of the signal.

[0016] : Distorted signal No. The amplitude of the samples.

[0017] : Original signal No. The phase angle of each sample, usually expressed in radians.

[0018] : Distorted signal No. The phase angle of each sample.

[0019] : Number of samples, indicating the total number of signal samples considered in the loss function.

[0020] : Weight coefficient, used to balance the proportion of amplitude error in the total loss function. This parameter can be adjusted according to the sensitivity of the task to amplitude distortion.

[0021] : Weight coefficient, used to balance the proportion of phase error in the total loss function. This parameter can be adjusted according to the sensitivity of the task to phase distortion; : Total loss function, which is the weighted sum of the amplitude error and phase error between the original signal and the distorted signal. The smaller the value of this function, the more distorted the signal is. With the original signal The closer.

[0022] Since the output signal needs to be corrected in real time, the present invention uses a recurrent neural network, which can quickly respond to changes in the original signal and update the output signal in real time to adapt to the dynamically changing environment. Through the optimized loss function, it can stably converge to the global minimum, thereby ensuring the stability and accuracy of the output signal.

[0023] Preferably, in the circuit system of PLC communication, each instruction information for the switching power supply is first transmitted to the gain control module, and then the instruction information is sent down to the switching power supply.

[0024] The pulse noise generated by the switching power supply is unpredictable, so the present invention first transmits the switching instruction to the gain control module for prediction processing.

[0025] Preferably, when the switch power supply is a physical button and needs to be manually switched, the switch instruction controller further includes a proximity sensing module, wherein the proximity sensing module is mounted and fixed on the physical switch button.

[0026] Since the switching operation of the physical key cannot be prefabricated in advance, the present invention sets a sensing module on the physical key to increase the comprehensiveness and accuracy of the prediction process.

[0027] Preferably, the method further includes an impulse noise analysis step to perform source tracing analysis on each impulse noise.

[0028] Preferably, according to the result obtained in the pulse noise analysis step, for the pulse noise that cannot be eliminated but is relatively stable, the gain control module makes a prediction on the gain adjustment in advance based on the previous gain data, thereby reducing the amount of calculation and lowering the power consumption.

[0029] Preferably, when a signal indicating that the switching power supply is about to work is received, a gain prediction is made synchronously, thereby achieving a zero-delay gain for stable pulse noise.

[0030] Compared with the prior art, the present invention has the following advantages and beneficial effects: 1. Use a root mean square (RMS) detector to measure the amplitude of the original signal. First, the original signal passes through a square law device, such as a diode, to convert it into a square signal proportional to the amplitude of the original signal. Next, the square signal is averaged through a low-pass filter to obtain the average level over a period of time. This average level is the root mean square value of the original signal. Finally, a square root extractor is used to convert the average value into an RMS value.

[0031] 2. The present invention mainly targets the problem of pulse noise in PLC signal transmission, where PLC signal transmission accuracy is high. Therefore, the present invention uses a root mean square (RMS) detector to measure the signal amplitude, which has the advantages of high accuracy, wide dynamic range, temperature stability, low power consumption, simplified design, and real-time response. These characteristics make the RMS detector an ideal power measurement tool for solving the technical problems of the present invention, especially in the accurate measurement of complex modulated signals, which can significantly reduce the system calibration requirements.

[0032] 3. The present invention uses a radial basis function neural network algorithm to perform adaptive gain adjustment on pulse noise. Since RBF does not require nonlinear mapping through weight connections, the learning speed of the pulse noise gain amplitude is accelerated, especially when applied to PLC real-time communication, which solves the delay problem well. In addition, the hidden layer output of the RBF network is linear with respect to the adjustable parameters, which means that the weights of the network can be directly obtained by solving a set of linear equations, thereby accelerating the learning speed and avoiding local minima problems.

[0033] 4. Since the determination of the RBF center point is the core of the radial basis function neural network algorithm, this application uses the K-Means clustering algorithm to determine the RBF center point. The K-Means algorithm can help the RBF network optimize the network structure and avoid overfitting while ensuring the accuracy requirements. In addition, K-Means clustering can identify representative points in the data as center points, reduce the number of nodes in the hidden layer of the RBF network, and simplify the network structure. DETAILED DESCRIPTION

[0034] The technical solution of the present invention is further described in detail below, but the protection scope of the present invention is not limited to the following.

[0035] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the embodiments described are only part of the embodiments of the present invention, rather than all the embodiments.

[0036] Therefore, the following detailed description of the embodiments of the present invention is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention. It should be noted that relational terms such as the terms "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0037] Moreover, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a list of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or apparatus. In the absence of more restrictions, an element defined by the phrase "comprises a ..." does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0038] The features and performance of the present invention are further described in detail below in conjunction with the embodiments. Example

[0039] A method for suppressing impulse noise based on gain control, comprising: S1: In the circuit system of PLC communication, the input original signal is automatically monitored and the changes of the original signal are monitored in real time; S2: Monitor the strength of the output signal through a feedback loop, compare the measured output signal characteristics with the original signal characteristics, and obtain an error signal; S3: Use an adaptive algorithm to process the error signal and provide an automatically adjusted gain parameter for the gain control module to achieve dynamic gain adjustment; S4: Monitor and feedback the output signal in real time, so that the gain control module can correct the output signal in real time.

[0040] The present invention provides a feedback loop and an adaptive algorithm to continuously correct the impulse noise, thereby ensuring the quality and reliability of signal transmission to the greatest extent.

[0041] In this embodiment, automatically monitoring the original signal includes the following: The amplitude of the original signal is measured using a root mean square (RMS) detector. First, the original signal passes through a square law device, such as a diode, to convert it into a square signal proportional to the amplitude of the original signal. Next, the square signal is averaged through a low-pass filter to obtain the average level over a period of time. This average level is the root mean square value of the original signal. Finally, a square root extractor is used to convert the average value into an RMS value.

[0042] The present invention mainly targets the problem of pulse noise in PLC signal transmission, where PLC signal transmission accuracy is high. Therefore, the present invention uses a root mean square (RMS) detector to measure the signal amplitude, which has the advantages of high accuracy, wide dynamic range, temperature stability, low power consumption, simplified design, and real-time response. These characteristics make the RMS detector an ideal power measurement tool for solving the technical problems of the present invention, especially in the accurate measurement of complex modulated signals, which can significantly reduce the system calibration requirements.

[0043] In this embodiment, the adaptive algorithm is specifically a radial basis function neural network algorithm, which specifically includes: S3.1: Extract the features of the error signal, including signal amplitude, phase, rise time, fall time, duration, and energy; S3.2: Use an unsupervised learning algorithm to determine the RBF (radial basis function) center point and use the center point as the neuron center of the network hidden layer; S3.3: Calculate the variance of each RBF center point based on the maximum distance between the center points; S3.4: Obtain the weights from the hidden layer to the output layer by the least squares method; S3.5: Dynamically adjust the signal gain by modifying the weight of the output layer; S3.6: The adjusted signal is output through the output layer, completing the gain adjustment process.

[0044] The present invention uses a radial basis function neural network algorithm to perform adaptive gain adjustment on impulse noise. Since the RBF does not need to perform nonlinear mapping through weight connections, the learning speed of the impulse noise gain amplitude is accelerated, especially when a large number of extracted features are required, the delay time required for gain adjustment is greatly reduced. In addition, the hidden layer output of the RBF network is linear to the adjustable parameters, which means that the weight of the network can be directly obtained by solving a linear equation group, thereby accelerating the learning speed and avoiding the local minimum problem.

[0045] In this embodiment, the unsupervised learning algorithm is specifically a K-Means clustering algorithm, which specifically includes: Randomly select K data points as the centers of the initial clusters ; S3.4.1: For each data point , calculate its distance from all cluster centers and assign it to the nearest cluster center to form K clusters ; S3.4.2: Distance is calculated using Euclidean distance in, is the characteristic of the signal. Features, is the cluster center No. Features S3.4.3: Recalculate the cluster center of each cluster and the mean of all points in the cluster: ,in It is a cluster The number of data points in ; S3.4.4: Check whether the cluster center has changed. If not or the predetermined number of iterations has been reached, the algorithm ends; otherwise, return to step S3.4.2; S3.4.5: When the change of cluster center in consecutive iterations is less than the set threshold or reaches the maximum number of iterations, the algorithm terminates.

[0046] Since the determination of the RBF center point is the core of the radial basis function neural network algorithm, this application uses the K-Means clustering algorithm to determine the RBF center point. The K-Means algorithm can help the RBF network optimize the network structure and avoid overfitting while ensuring the accuracy requirements. In addition, K-Means clustering can identify representative points in the data as center points, reduce the number of nodes in the hidden layer of the RBF network, and simplify the network structure.

[0047] In this embodiment, the real-time correction of the output signal in S4 specifically includes: S4.1: Set the original input signal to The processed output signal is ; S4.2: Processing raw signals using recurrent neural networks With output signal The amplitude and phase differences between S4.3: The loss function is defined as follows: in, : The original signal can be a time series signal or any form of signal, where Indicates the signal samples.

[0048] : Output signal, which is the output signal after system processing (such as neural network), where Indicates the first samples.

[0049] : Original signal No. The amplitude of each sample, that is, the size or strength of the signal.

[0050] : Distorted signal No. The amplitude of the samples.

[0051] : Original signal No. The phase angle of each sample, usually expressed in radians.

[0052] : Distorted signal No. The phase angle of each sample.

[0053] : Number of samples, indicating the total number of signal samples considered in the loss function.

[0054] : Weight coefficient, used to balance the proportion of amplitude error in the total loss function. This parameter can be adjusted according to the sensitivity of the task to amplitude distortion.

[0055] : Weight coefficient, used to balance the proportion of phase error in the total loss function. This parameter can be adjusted according to the sensitivity of the task to phase distortion; : Total loss function, which is the weighted sum of the amplitude error and phase error between the original signal and the distorted signal. The smaller the value of this function, the more distorted the signal is. With the original signal The closer.

[0056] Since feedback correction needs to be timely, the present invention selects the two features that have the greatest impact on signal distortion, namely amplitude and phase, for targeted real-time correction. The present invention uses a recurrent neural network because the recurrent neural network can quickly respond to changes in the original signal and update the output signal in real time to adapt to the dynamically changing environment. Through the optimized loss function, it can stably converge to the global minimum, thereby ensuring the stability and accuracy of the output signal.

[0057] In this embodiment, in the circuit system of PLC communication, each instruction information for the switching power supply is first transmitted to the gain control module, and then the instruction information is sent to the switching power supply.

[0058] The pulse noise generated by the switching power supply is unpredictable, so the present invention first transmits the switching instruction to the gain control module for prediction processing.

[0059] In this embodiment, when the switch power supply is a physical button and needs to be manually switched, the switch instruction controller further includes a proximity sensing module, wherein the proximity sensing module is mounted and fixed on the physical switch button.

[0060] Since the switching operation of the physical key cannot be prefabricated in advance, the present invention sets a sensing module on the physical key to increase the comprehensiveness and accuracy of the prediction process.

[0061] In this embodiment, an impulse noise analysis step is also included to perform source tracing analysis on each impulse noise, wherein the following steps are specifically taken: Noise characteristic analysis: Analyze the characteristics of impulse noise. For example, analyzing the spectrum distribution of impulse noise can help identify whether it is natural noise or man-made noise.

[0062] Noise source location: Locate noise sources through measurement and analysis, and identify possible interference sources, such as high-voltage wires, electrical equipment, and other communication equipment.

[0063] Noise modeling: Building noise models, such as the Middleton A-type noise model or the symmetric alpha-stable distribution (SαS) noise model, can help simulate and understand the characteristics of noise.

[0064] Experimental verification: The effectiveness of the noise reduction measures taken is verified through experiments, such as using simulation experiments and measured data to verify the performance of the noise reduction algorithm.

[0065] Continuous monitoring: Establish a long-term monitoring mechanism to promptly detect and respond to new noise sources.

[0066] In this embodiment, according to the results obtained in the pulse noise analysis step, for pulse noise that cannot be eliminated but is relatively stable, the gain control module makes a prediction on the gain adjustment in advance based on previous gain data, thereby reducing the amount of calculation and lowering power consumption.

[0067] In this embodiment, when a signal indicating that the switching power supply is about to work is received, a gain prediction is made synchronously, thereby achieving a zero-delay gain for stable pulse noise.

[0068] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for suppressing impulse noise based on gain control, It is characterized in that include: S1: In the circuit system of PLC communication, the input original signal is automatically monitored and the changes of the original signal are monitored in real time; S2: Monitor the strength of the output signal through a feedback loop, compare the measured output signal characteristics with the original signal characteristics, and obtain an error signal; S3: Use an adaptive algorithm to process the error signal and provide an automatically adjusted gain parameter for the gain control module to achieve dynamic gain adjustment; S4: Monitor and feedback the output signal in real time, so that the gain control module can correct the output signal in real time.

2. The pulse noise suppression method based on gain control according to claim 1, It is characterized in that The automatic monitoring of raw signals includes the following: The amplitude of the original signal is measured using an RMS detector. First, the original signal passes through a square law device to convert it into a square signal proportional to the amplitude of the original signal. Next, the square signal is averaged through a low-pass filter to obtain the average level over a period of time. This average level is the RMS value of the original signal. Finally, a square root extractor is used to convert the average value into an RMS value.

3. The pulse noise suppression method based on gain control according to claim 1, It is characterized in that The adaptive algorithm is specifically a radial basis function neural network algorithm, which specifically includes: S3.1: Extract the features of the error signal, including signal amplitude, phase, rise time, fall time, duration, and energy; S3.2: Use an unsupervised learning algorithm to determine the RBF center point and use the center point as the neuron center of the network hidden layer; S3.3: Calculate the variance of each RBF center point based on the maximum distance between the center points; S3.4: Obtain the weights from the hidden layer to the output layer by the least squares method; S3.5: Dynamically adjust the signal gain by modifying the weight of the output layer; S3.6: The adjusted signal is output through the output layer, completing the gain adjustment process.

4. The method for suppressing impulse noise based on gain control according to claim 3, It is characterized in that The unsupervised learning algorithm is specifically the K-Means clustering algorithm, which specifically includes: Randomly select K data points as the centers of the initial clusters ; S3.4.1: For each signal feature set data point y, calculate its distance from all cluster centers and assign it to the nearest cluster center to form K clusters. ; S3.4.2: Distance is calculated using Euclidean distance in, is the characteristic of the signal. Features, is the cluster center No. Features S3.4.3: Recalculate the cluster center of each cluster and the mean of all points in the cluster: ,in It is a cluster The number of data points in ; S3.4.4: Check whether the cluster center has changed. If not or the predetermined number of iterations has been reached, the algorithm ends; otherwise, return to step S3.4.2; S3.4.5: When the change of cluster center in consecutive iterations is less than the set threshold or reaches the maximum number of iterations, the algorithm terminates.

5. The pulse noise suppression method based on gain control according to claim 1, It is characterized in that S4 performs real-time correction on the output signal, including: S4.1: Set the original input signal to The processed output signal is ; S4.2: Processing raw signals using recurrent neural networks With output signal The amplitude and phase differences between S4.3: The loss function is defined as follows: in, : The original signal, where Indicates the signal samples; : Output signal, where Indicates the first samples; : Original signal No. The amplitude of the samples; : Distorted signal No. The amplitude of the samples; : Original signal No. The phase angle of each sample; : Distorted signal No. The phase angle of each sample; : Number of samples, indicating the total number of signal samples considered in the loss function; : Weight coefficient, used to balance the proportion of amplitude error in the total loss function; : Weight coefficient, used to balance the proportion of phase error in the total loss function; : Total loss function.

6. The pulse noise suppression method based on gain control according to claim 5, It is characterized in that Each instruction information for the switching power supply is first transmitted to the gain control module, and then the instruction information is sent to the switching power supply.

7. The pulse noise suppression method based on gain control according to claim 6, It is characterized in that The switch command controller also includes a proximity sensing module, wherein the proximity sensing module is mounted and fixed on the physical switch button.

8. The method for suppressing impulse noise based on gain control according to claim 7, It is characterized in that It also includes an impulse noise analysis step to achieve source tracing analysis of each impulse noise.

9. The method for suppressing impulse noise based on gain control according to claim 8, It is characterized in that According to the results obtained in the pulse noise analysis step, for the pulse noise that cannot be eliminated but is relatively stable, the gain control module makes a prediction on the gain adjustment in advance based on the previous gain data.

10. The pulse noise suppression method based on gain control according to claim 9, It is characterized in that When receiving a signal that the switching power supply is about to work, a gain prediction is made synchronously.

Images