Harmonic metering method and system for electric energy meter based on adaptive filtering
By using multi-scale time series decomposition and an improved adaptive filtering algorithm, the filtering parameters are dynamically adjusted to adapt to changes in signal state, thus solving the problem of metering accuracy in complex harmonic environments for traditional energy meters and achieving high-precision and stable harmonic metering.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- JIANGYIN ZHONGHE POWER METER
- Filing Date
- 2026-04-23
- Publication Date
- 2026-06-26
AI Technical Summary
Traditional electricity meters are not very accurate in complex harmonic environments, especially when the load changes rapidly or the operating conditions are complex. The fixed shape parameters of existing adaptive filtering algorithms lead to distortion of filtering results in the face of sudden interference or strong noise.
By employing multi-scale time series decomposition and an improved adaptive filtering algorithm, the filtering parameters are dynamically adjusted to adapt to changes in signal state by constructing an inverse correlation between shape parameters and adjustment factors, and combining indices such as kurtosis, harmonic abruptness, and standard deviation of short-time residual sequences and long-time residual sequences, thereby achieving effective suppression of sudden disturbances and noise.
It improves the metering accuracy and robustness of electricity meters under complex non-stationary signals, reduces the impact of noise interference on metering results, and enhances the adaptability and reliability of the system.
Smart Images
Figure CN122084970B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electrical variable measurement technology. More specifically, this invention relates to a harmonic measurement method and system for electricity meters based on adaptive filtering. Background Technology
[0002] With the integration of numerous nonlinear loads into power systems, such as frequency converters, power electronic devices, and distributed power sources, harmonic problems in the power grid are becoming increasingly prominent. Harmonics not only degrade power quality but also significantly impact the accuracy of power metering devices. Against this backdrop, accurate measurement of harmonic energy, while maintaining basic active power metering, has gradually become an important research direction in the field of power metering. Traditional power meters are mostly designed based on the assumption of ideal sine waves. When complex harmonic components and random disturbances exist in the actual power grid, it can easily lead to increased deviations in metering results. Especially under conditions of rapid load changes or complex operating conditions, harmonic signals exhibit obvious non-stationarity and nonlinear characteristics, thus placing higher demands on the stability and accuracy of metering algorithms.
[0003] To improve the accuracy of harmonic signal extraction and analysis, adaptive filtering algorithms are increasingly being introduced into existing technologies to preprocess current or voltage signals, thereby suppressing noise interference and enhancing the distinguishability of harmonic components. Adaptive filtering algorithms can dynamically adjust filtering parameters based on the statistical characteristics of the input signal, thus improving adaptability to complex signal environments to a certain extent.
[0004] However, in practical applications, most adaptive filtering methods, especially the generalized maximum correlation entropy adaptive filtering algorithm, typically use fixed shape parameters to control the sensitivity and robustness of the error distribution. Once set, these parameters remain unchanged throughout the filtering process, lacking the ability to respond to dynamic changes in the signal. In complex power grid environments, harmonic signals often exhibit strong abruptness and significant time-varying distribution characteristics. When the shape parameter in the filtering algorithm is fixed, it can maintain high estimation accuracy during stable signal phases. However, under sudden interference or strong noise environments, it is prone to insufficient or excessive suppression of outliers, leading to distorted filtering results and consequently, inaccurate harmonic component extraction and power metering. Summary of the Invention
[0005] To address the problems of low accuracy in harmonic component extraction and power metering mentioned in the background art, the present invention provides solutions in the following aspects.
[0006] In a first aspect, the present invention provides a harmonic measurement method for electricity meters based on adaptive filtering, comprising: acquiring parameter data of the electricity meter at the current time node and its historical multiple time nodes to construct a target parameter sequence; decomposing the target parameter sequence using a time series decomposition algorithm to obtain a target residual sequence; decomposing the target residual sequence into a short-time residual sequence and a long-time residual sequence; performing adaptive filtering processing on the target parameter sequence using an improved adaptive filtering algorithm to obtain harmonic components, and completing the harmonic measurement analysis of the electricity meter based on the harmonic components; wherein, the improved adaptive filtering algorithm includes a shape parameter, the shape parameter being inversely correlated with a adjustment factor; the adjustment factor being positively correlated with the kurtosis value and harmonic abrupt change of the short-time residual sequence; the harmonic abrupt change being positively correlated with the standard deviation of the short-time residual sequence and the autocorrelation coefficient between the parameter data at the current time node and the corresponding parameter data after a lag set time node, and inversely correlated with the standard deviation of the long-time residual sequence.
[0007] The above technical solution introduces a multi-scale time series decomposition mechanism to effectively separate the short-term fluctuation characteristics and long-term trends in the original parameter data. Based on this, a comprehensive adjustment mechanism reflecting the intensity and distribution characteristics of sudden disturbances is constructed. Furthermore, the adjustment results are dynamically correlated with key control parameters in the adaptive filtering process, enabling the filtering capability to be adjusted in real time according to changes in signal state. Its overall advantage lies in achieving deep coupling between signal feature extraction and filtering parameter optimization: it can maintain high detail preservation capability and measurement accuracy during signal stability, while enhancing the suppression capability of abnormal components under sudden interference or complex noise environments. This effectively avoids the problem of difficulty in balancing accuracy and robustness caused by fixed parameters in traditional methods.
[0008] Furthermore, the shape parameters for: , , These are the upper limit and lower limit values of the shape parameter, respectively. For the natural constant An exponential function with base 0. for Adjustment factors for time nodes.
[0009] The aforementioned technical solution constructs a nonlinear mapping relationship that continuously changes with the adjustment result, enabling key control parameters in the filtering process to smoothly and adaptively adjust within a preset range. Its advantages are: when the signal is in a stable state, this parameter remains at a high level, thereby enhancing the ability to retain detailed information and improving processing accuracy; while when sudden disturbances or abnormal fluctuations occur in the signal, this parameter rapidly decreases as the adjustment result increases, significantly improving the suppression of outliers and strong noise. Simultaneously, the exponential change avoids instability caused by abrupt parameter changes, achieving a smooth transition from a stable state to a disturbed state. Based on this dynamic adjustment mechanism, it effectively solves the problem of insufficient adaptability caused by fixed parameters in traditional methods, balancing accuracy and robustness under different operating conditions, improving the ability to process complex non-stationary signals, and thus enhancing the overall filtering effect and the reliability and stability of subsequent harmonic analysis and measurement results.
[0010] Furthermore, Adjustment factor at time point for: , To find the maximum value function, The kurtosis value of the short-time residual sequence. To preset the kurtosis threshold, for Harmonic abrupt change at time points.
[0011] The above technical solution integrates statistical characteristics reflecting the sharpness of signal distribution with indicators describing the change in instantaneous fluctuation intensity, and introduces threshold constraints and nonlinear normalization mechanisms. This ensures that the adjustment result is effectively activated only when the signal deviates significantly from the normal distribution. Simultaneously, it gradually strengthens when the change intensity is large and tends to stabilize in the high-intensity range, thereby achieving selective response and amplitude suppression to abnormal disturbances. Its beneficial effects are that it can avoid unnecessary adjustment caused by slight fluctuations or noise, improving system stability, and can quickly improve response capability when sudden shocks or significant anomalies occur, enhancing adaptability to complex non-stationary signals. At the same time, the smooth change mechanism avoids drastic fluctuations during the adjustment process, thereby improving the robustness and reliability of the overall processing process, and providing a more stable and discriminative adjustment basis for subsequent filtering and harmonic measurement.
[0012] Furthermore, Harmonic abrupt change at time node for: , , These are the standard deviations of the short-time residual sequence and the long-time residual sequence, respectively. for The autocorrelation coefficient between the time node parameter data and the corresponding parameter data after the lag setting time node.
[0013] The aforementioned technical solution comprehensively characterizes sudden changes in signals by comparing fluctuation characteristics at different time scales and combining the correlation information between the current state and historical states. Its beneficial effect lies in its ability to simultaneously reflect the intensity of instantaneous disturbances and the overall change background, so that the judgment of abnormal fluctuations no longer depends on single-scale features, thereby significantly improving the ability to identify sudden impacts, intermittent disturbances, and continuous changes. At the same time, by introducing time correlation constraints, it can effectively distinguish between random noise and continuous structural changes, reduce the occurrence of misjudgments and omissions, make the detection results more stable and reliable, and thus enhance the adaptability to non-stationary signals under complex working conditions. It also provides a more accurate and discriminative basis for subsequent adjustment and filtering processes, thereby improving the accuracy and robustness of overall harmonic analysis and measurement.
[0014] Furthermore, the metering analysis specifically involves: performing frequency domain decomposition on the harmonic components to obtain the amplitude of each order harmonic, obtaining the corresponding effective value based on the amplitude, and accumulating the effective value within a set time interval to obtain the harmonic energy value of the energy meter.
[0015] The aforementioned technical solution effectively separates the frequency components of complex current or voltage signals by performing frequency domain decomposition on the harmonic signal, accurately obtains their amplitude information, and then calculates the corresponding effective values to reflect the contribution of each harmonic in actual energy transmission. By accumulating the effective values within a set time interval, it can not only accurately quantify instantaneous and short-term harmonic energy but also reflect the energy accumulation characteristics during long-term operation, thereby significantly improving the accuracy and stability of harmonic metering. This method can effectively reduce the impact of noise interference and signal fluctuations on metering results, making the energy identification of each harmonic under complex non-stationary operating conditions more reliable. It provides refined and reliable data support for power quality assessment, harmonic mitigation, and equipment condition monitoring, thereby improving the robustness, adaptability, and engineering application value of the power metering system as a whole.
[0016] Furthermore, the time series decomposition algorithm is the STL time series decomposition algorithm.
[0017] Furthermore, the parameter data is current signal data or voltage signal data.
[0018] Furthermore, it also includes: denoising the parameter data.
[0019] Furthermore, the adaptive filtering algorithm is a generalized maximum correlation entropy adaptive filtering algorithm.
[0020] In a second aspect, the present invention provides an energy meter harmonic metering system based on adaptive filtering, comprising a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the energy meter harmonic metering method based on adaptive filtering described above is implemented.
[0021] The beneficial effects of this invention are as follows:
[0022] This invention constructs a harmonic signal processing framework based on multi-scale time series decomposition, effectively separating short-term fluctuations from long-term trends. It introduces a dynamic adjustment mechanism that comprehensively reflects the intensity of instantaneous fluctuations, the sharpness of their distribution, and the autocorrelation characteristics of the signal. This allows key control parameters in the adaptive filtering process to be adjusted in real time according to the signal state, maintaining high-precision detail retention under stable operating conditions and enhancing the suppression of anomalous components when sudden disturbances or abnormal fluctuations occur, achieving high adaptability to complex non-stationary signals. Simultaneously, through frequency domain decomposition and RMS accumulation metering analysis methods, the energy contribution of each harmonic order can be accurately extracted, significantly improving the harmonic metering accuracy and stability of electricity meters under dynamic and non-ideal grid conditions. This reduces the interference of noise and transient fluctuations on metering results, thereby enhancing the overall reliability, robustness, and engineering practicality of the system. It provides accurate and reliable technical means for power quality assessment, harmonic mitigation, and grid equipment condition monitoring. Attached Figure Description
[0023] Figure 1 This is a flowchart illustrating an adaptive filtering-based harmonic metering method for electricity meters according to an embodiment of the present invention;
[0024] Figure 2 This is a schematic diagram illustrating the shape parameter adjustment process of the adaptive filtering algorithm in the harmonic metering method for energy meters based on adaptive filtering according to an embodiment of the present invention.
[0025] Figure 3 This is a schematic block diagram illustrating the structure of an energy meter harmonic metering system based on adaptive filtering according to an embodiment of the present invention. Detailed Implementation
[0026] An example of a harmonic metering method for electricity meters based on adaptive filtering.
[0027] like Figure 1 As shown in the flowchart of the harmonic metering method for electricity meters based on adaptive filtering according to an embodiment of the present invention, the method includes the following steps:
[0028] S1: Obtain the parameter data of the electricity meter at the current time node and its historical multiple time nodes to construct the target parameter sequence.
[0029] In a preferred embodiment, the parameter data is specifically selected as current signal data or voltage signal data reflecting the system's operating state. By collecting and constructing the original time-series data formed by the changes in current or voltage over time, the dynamic electrical characteristics of the equipment during operation can be comprehensively characterized. Furthermore, to address the issue that the parameter data is easily affected by environmental noise, electromagnetic interference, and sensor errors during acquisition, denoising processing is performed on the parameter data. Specifically, wavelet threshold denoising methods or Kalman filtering methods can be used to smooth and suppress noise in the original signal, thereby effectively reducing the impact of high-frequency random noise and abnormal impulse interference on data quality, significantly improving the signal-to-noise ratio and stability of the data used for subsequent analysis, and thus providing a more reliable data foundation for subsequent time-series feature extraction and anomaly identification.
[0030] After denoising, the process further includes multi-scale structural analysis of the target parameter sequence using a time series decomposition algorithm to obtain the target residual sequence. Specifically, the STL time series decomposition algorithm is preferably used to decompose the target parameter sequence, representing the original time series as the sum of a trend term, a periodic term, and a residual term. By effectively separating trend changes and periodic fluctuations, the target residual sequence that can characterize non-stationary disturbances and anomalous fluctuations is extracted. Furthermore, to more precisely characterize the disturbance behavior at different time scales, the target residual sequence is decomposed again, dividing it into short-time residual sequences and long-time residual sequences. The short-time residual sequence mainly reflects sudden disturbances, transient shocks, and high-frequency fluctuations, while the long-time residual sequence is used to characterize slowly evolving system offsets, cumulative errors, and low-frequency drift characteristics.
[0031] Through the above-mentioned hierarchical decomposition process, not only can the multi-scale features in complex time-series signals be effectively separated, but the system's ability to perceive different types of anomalies can also be significantly enhanced: on the one hand, the introduction of short-time residual sequences helps to improve the detection sensitivity of instantaneous anomalies and abrupt events, and avoids anomalies being masked by trend terms or periodic terms; on the other hand, the extraction of long-time residual sequences helps to identify long-term latent faults or progressive performance degradation problems, thereby achieving early warning of equipment operating status.
[0032] S2: The target parameter sequence is adaptively filtered using an improved adaptive filtering algorithm to obtain harmonic components.
[0033] like Figure 2 The figure shows the curve of the shape parameter adjustment process of the adaptive filtering algorithm of the harmonic metering method based on adaptive filtering in an embodiment of the present invention.
[0034] In a preferred embodiment, the adaptive filtering algorithm is a generalized maximum correlation entropy adaptive filtering algorithm, wherein the improved adaptive filtering algorithm includes a shape parameter, the shape parameter... for: , , These are the upper limit and lower limit values of the shape parameter, respectively. For the natural constant An exponential function with base 0. for Adjustment factors for time nodes.
[0035] By establishing a functional mapping relationship between the shape parameter and the adjustment factor in the filter, the shape parameter can adaptively adjust according to changes in the signal state. When the system is in a stable or slightly fluctuating state, the adjustment amount is small, and the corresponding parameter remains at a high level, thereby enhancing the ability to preserve detailed information and improving estimation accuracy. When there are significant abrupt changes or abnormal disturbances in the signal, the adjustment amount increases, and the parameter is rapidly reduced through an exponential nonlinear mapping, thereby enhancing the filter's ability to suppress outliers and impulse noise, while avoiding the discontinuity problems caused by parameter abrupt changes, achieving smooth transition and dynamic response. Based on this data feature-driven adaptive adjustment mechanism, robustness to strong noise, impulse interference, and non-Gaussian noise can be effectively improved without excessively weakening signal details. This allows the filtering process to balance accuracy and stability under different operating conditions, thereby significantly improving the reliability of overall signal processing and state analysis.
[0036] Adjustment factor at time point for: , To find the maximum value function, The kurtosis value of the short-time residual sequence. To preset the kurtosis threshold, for Harmonic abrupt change at time points.
[0037] Using the sharpness of the signal distribution pattern as a trigger criterion, and comparing it with a preset standard, the adjustment mechanism is activated only when there is a significant deviation from the normal distribution, thus achieving a selective response to abnormal fluctuations. This allows the adjustment result to gradually improve during periods of heightened abnormality and stabilize in high-intensity ranges, avoiding the instability caused by excessive amplification. By integrating the determination of whether an anomaly exists with the measurement of its degree, the system remains sensitive to sudden shocks while maintaining stability under noise or minor disturbances. Based on this method, the ability to identify peak-shaped anomalies and heavy-tailed distribution characteristics can be effectively improved, the interference of invalid fluctuations on the adjustment process can be reduced, and the robustness and adaptability under complex operating conditions can be enhanced, thereby improving the accuracy and reliability of the overall analysis and control process.
[0038] Harmonic abrupt change at time node for: , , These are the standard deviations of the short-time residual sequence and the long-time residual sequence, respectively. for The autocorrelation coefficient between the time node parameter data and the corresponding parameter data after the lag setting time node.
[0039] By comparing and analyzing the fluctuation characteristics of signals at different time scales, this method comprehensively characterizes the degree of rapid change within a short time range and the overall stability within a long time range. Simultaneously, it incorporates correlation information between the current moment and historical moments to dynamically correct abrupt changes, thus constructing a comprehensive evaluation mechanism that considers both the intensity of instantaneous disturbances and the background of long-term changes. On the one hand, it utilizes the fluctuation characteristics at short time scales to highlight the sensitivity to sudden changes; on the other hand, it uses the change characteristics at long time scales to constrain the overall trend, and further reflects the consistency between current changes and historical states through time correlation. This allows the judgment of abnormal fluctuations to no longer rely on a single scale or a single feature, but rather to form a multi-dimensional fusion of judgment criteria. Based on the above method, it can effectively improve the ability to identify sudden anomalies, intermittent disturbances, and gradual changes in harmonic signals, avoiding misjudgments or omissions caused by single-scale analysis, while enhancing the adaptability to non-stationary signals under complex operating conditions, thereby improving the accuracy and stability of overall monitoring and early warning results.
[0040] S3: Perform harmonic metering analysis based on the harmonic components.
[0041] In a preferred embodiment, the metrological analysis process includes: firstly, performing frequency domain decomposition on the harmonic components. This can be achieved using Fast Fourier Transform or other spectral analysis methods to map the time-domain signal to the frequency domain, thereby separating the harmonic components corresponding to different frequency levels and obtaining the amplitude information of each harmonic. This step enables effective decoupling of different frequency components in complex electrical signals, clearly characterizing the features of each harmonic, and avoiding interference from the superposition of different frequency components, thus significantly improving the resolution and accuracy of harmonic identification.
[0042] After obtaining the amplitude of each harmonic, the corresponding effective values are calculated based on the amplitude information to reflect the equivalent energy contribution of each harmonic in the actual power transmission process. The introduction of effective values eliminates the influence of alternating positive and negative signals, making the energy characterization more stable and physically meaningful, thus providing a reliable basis for subsequent energy metering. Furthermore, within a set time interval, the effective values corresponding to each harmonic are accumulated to obtain the harmonic energy value of the electricity meter within that time range. Through integral accumulation processing over the time dimension, not only can the energy changes of harmonics in short-term transient processes be reflected, but also their energy contribution level during longer-term operation can be characterized, thereby achieving a multi-scale quantitative assessment of the harmonic impact.
[0043] This invention achieves high-precision extraction and quantification analysis of harmonic signals from electricity meters by organically combining multi-timescale signal decomposition, dynamic adjustment mechanisms, and improved adaptive filtering algorithms. By separating short-term fluctuations from long-term trends and constructing adjustment factors based on fluctuation sharpness, sudden changes, and time correlation, the filtering parameters can adaptively adjust with the signal state. This maintains high fidelity in detail during stable signal phases, enhances the suppression of anomalous components during sudden disturbances or abnormal fluctuations, and avoids the instability caused by abrupt changes in filtering parameters. Combining frequency domain decomposition and RMS value accumulation calculation, this scheme can accurately quantify the energy of each harmonic order, significantly improving metering accuracy and robustness under complex non-stationary conditions. Overall, this scheme balances signal processing accuracy, anomaly suppression capability, and system adaptability, effectively reducing the impact of noise and interference on harmonic metering results, improving the reliability, stability, and engineering application value of electricity meters in actual power systems, and providing a feasible technical path for intelligent harmonic metering through systematic implementation.
[0044] Example of a harmonic metering system for electricity meters based on adaptive filtering:
[0045] like Figure 3As shown in the diagram, the structural block diagram of the harmonic metering system for an energy meter based on adaptive filtering according to an embodiment of the present invention includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the harmonic metering method for an energy meter based on adaptive filtering according to the present invention.
[0046] The energy meter harmonic metering system based on adaptive filtering also includes other components well known to those skilled in the art, such as communication interfaces. Their settings and functions are known in the art and will not be described in detail here.
[0047] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions stored or otherwise maintained by such a computer-readable medium.
[0048] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise explicitly specified.
[0049] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.
Claims
1. A harmonic metering method for electricity meters based on adaptive filtering, characterized in that, include: Obtain parameter data of the electricity meter at the current time point and at multiple historical time points to construct a target parameter sequence; The target parameter sequence is decomposed using a time series decomposition algorithm to obtain the target residual sequence; the target residual sequence is then decomposed into a short-time residual sequence and a long-time residual sequence. An improved adaptive filtering algorithm is used to adaptively filter the target parameter sequence to obtain harmonic components, and the metering analysis of harmonics in the energy meter is completed based on the harmonic components; the adaptive filtering algorithm is the generalized maximum correlation entropy adaptive filtering algorithm. The improved adaptive filtering algorithm includes a shape parameter, which is inversely correlated with the adjustment factor; the adjustment factor is positively correlated with the kurtosis and harmonic abruptness of the short-time residual sequence; the shape parameter... for: , , These are the upper limit and lower limit values of the shape parameter, respectively. For the natural constant An exponential function with base 0. for Adjustment factors at specific time points; Adjustment factor at time point for: , To find the maximum value function, The kurtosis value of the short-time residual sequence. To preset the kurtosis threshold, for Harmonic abrupt change at time points; Harmonic abrupt change at time node for: , , These are the standard deviations of the short-time residual sequence and the long-time residual sequence, respectively. for The autocorrelation coefficient between the time node parameter data and the corresponding parameter data after the lag setting time node; The harmonic abrupt change degree is positively correlated with the standard deviation of the short-time residual sequence and the autocorrelation coefficient between the current time node parameter data and the corresponding parameter data after the set time node lag, and negatively correlated with the standard deviation of the long-time residual sequence.
2. The harmonic metering method for energy meters based on adaptive filtering according to claim 1, characterized in that, The metering analysis specifically involves: performing frequency domain decomposition on the harmonic components to obtain the amplitude of each order harmonic, obtaining the corresponding effective value based on the amplitude, and accumulating the effective value within a set time interval to obtain the harmonic energy value of the energy meter.
3. The harmonic metering method for energy meters based on adaptive filtering according to claim 1, characterized in that, The time series decomposition algorithm is the STL time series decomposition algorithm.
4. The harmonic metering method for energy meters based on adaptive filtering according to claim 1, characterized in that, The parameter data is either current signal data or voltage signal data.
5. The harmonic metering method for energy meters based on adaptive filtering according to claim 1, characterized in that, Also includes: The parameter data is then denoised.
6. A harmonic metering system for electricity meters based on adaptive filtering, characterized in that, It includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the harmonic metering method for energy meters based on adaptive filtering as described in any one of claims 1 to 5 is implemented.