Intelligent sensing equipment electromagnetic interference suppression method and system based on AI identification

Abstract

The invention relates to the technical field of signal processing, and provides an intelligent sensing device electromagnetic interference suppression method and system based on AI identification, and the method comprises the steps: obtaining a multi-source electromagnetic signal set of a target intelligent sensing device in a preset monitoring period, and carrying out the electromagnetic interference feature extraction processing of the multi-source electromagnetic signal set, performing interference mode matching processing on the electromagnetic interference characteristic set based on a preset interference suppression model to generate an electromagnetic interference suppression strategy of the intelligent sensing equipment; the electromagnetic interference suppression strategy comprises a frequency band adjustment direction corresponding to an interference frequency band distribution characteristic, a signal gain compensation direction corresponding to an interference intensity fluctuation characteristic and an anti-interference filtering direction corresponding to an interference source correlation characteristic; and performing interference suppression operation on a signal acquisition module of the intelligent sensing equipment according to the electromagnetic interference suppression strategy so as to reduce the electromagnetic signal distortion rate of the intelligent sensing equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal processing, in particular to an intelligent sensing device electromagnetic interference suppression method and system based on AI identification. BACKGROUND

[0002] In today's technological development, intelligent sensing devices are widely used in various fields, and the electromagnetic environment during their operation is increasingly complex. Electromagnetic interference suppression has become a key link to ensure the normal operation of the device. Numerous technologies are committed to solving the problem of electromagnetic interference, but most of them only focus on single-dimensional electromagnetic signal features or simple interference suppression methods. Specifically, the existing technology does not comprehensively collect the electromagnetic signals of intelligent sensing devices, often only considers signals in a single working mode, and cannot reflect the real electromagnetic interference conditions of the device in diversified working scenarios. In terms of electromagnetic interference analysis, the existing technology lacks the ability to extract deep features from multiple key dimensions, making it difficult to accurately grasp the essential characteristics of interference. At the same time, the generation of interference suppression strategies lacks sufficient mining of the potential relationship between various interference characteristics, resulting in insufficient pertinence and effectiveness of the strategies, which cannot effectively reduce the electromagnetic signal distortion rate, and seriously affects the performance and reliability of intelligent sensing devices. SUMMARY

[0003] In order to improve the above problems, the present application provides an intelligent sensing device electromagnetic interference suppression method and system based on AI identification.

[0004] The first aspect of the embodiment of the present application provides an intelligent sensing device electromagnetic interference suppression method based on AI identification, applied to an intelligent sensing device electromagnetic interference suppression system, the method comprising: obtaining a multi-source electromagnetic signal set of a target intelligent sensing device in a preset monitoring period, the multi-source electromagnetic signal set comprising electromagnetic signal sequences collected by the intelligent sensing device in different working modes; performing electromagnetic interference feature extraction processing on the multi-source electromagnetic signal set to obtain an electromagnetic interference feature set of the intelligent sensing device, the electromagnetic interference feature set comprising interference frequency band distribution features, interference intensity fluctuation features and interference source correlation features of the electromagnetic signal sequences; performing interference mode matching processing on the electromagnetic interference feature set based on a preset interference suppression model to generate an electromagnetic interference suppression strategy of the intelligent sensing device, the electromagnetic interference suppression strategy comprising a frequency band adjustment direction corresponding to the interference frequency band distribution features, a signal gain compensation direction corresponding to the interference intensity fluctuation features, and an anti-interference filtering direction corresponding to the interference source correlation features; performing interference suppression operation on the signal acquisition module of the intelligent sensing device according to the electromagnetic interference suppression strategy to reduce the electromagnetic signal distortion rate of the intelligent sensing device.

[0005] In a second aspect, the application provides an intelligent sensing device electromagnetic interference suppression system, comprising: a processor and a memory and a bus connected to the processor; the processor and the memory complete mutual communication through the bus; the processor is used to call the computer program in the memory, so as to execute the AI identification based intelligent sensing device electromagnetic interference suppression method.

[0006] In a third aspect, the application provides a computer readable storage medium, which stores a program, and the program is executed by a processor to implement the AI identification based intelligent sensing device electromagnetic interference suppression method.

[0007] The AI identification based intelligent sensing device electromagnetic interference suppression method and system provided by the application first acquires a multi-source electromagnetic signal set of a target intelligent sensing device in different working modes within a preset monitoring period, comprehensively covers the electromagnetic signal conditions in various running states of the device, and provides rich and real data basis for subsequent analysis. Then, the multi-source electromagnetic signal set is subjected to electromagnetic interference feature extraction processing, and an electromagnetic interference feature set containing interference frequency band distribution features, interference intensity fluctuation features and interference source correlation features is accurately obtained, so that the electromagnetic interference characteristics are deeply analyzed from multiple key dimensions, and the grasp of the interference condition is more comprehensive and detailed. Based on the preset interference suppression model, interference mode matching processing is performed, the potential relationship between the electromagnetic interference features is fully mined, and a precise and targeted electromagnetic interference suppression strategy is generated. Finally, the signal acquisition module is subjected to interference suppression operation according to the strategy, the electromagnetic signal distortion rate of the intelligent sensing device can be effectively reduced, the accuracy and reliability of signal acquisition are greatly improved, the intelligent sensing device can be stably and efficiently operated in a complex electromagnetic environment, and the accurate perception and processing of various electromagnetic information by the intelligent sensing device are ensured. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0009] Figure 1 A flowchart of an AI identification based intelligent sensing device electromagnetic interference suppression method provided by the application.

[0010] Figure 2 A product module schematic diagram of an intelligent sensing device electromagnetic interference suppression system provided by the application. DETAILED DESCRIPTION

[0011] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood, and so that the full scope of the present disclosure is conveyed to those skilled in the art.

[0012] In order to better understand the above technical solutions, the technical solutions of the present application will be described in detail below with reference to the drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solutions of the present application, and are not limitations of the technical solutions of the present application. In the case of no conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0013] Please refer to Figure 1 A flowchart of an AI recognition-based intelligent sensing device electromagnetic interference suppression method provided by the embodiments of the present application. The method is applied to an intelligent sensing device electromagnetic interference suppression system. The specific content description of the method includes steps 110-140.

[0014] Step 110: Obtain a multi-source electromagnetic signal set of a target intelligent sensing device in a preset monitoring period, wherein the multi-source electromagnetic signal set includes electromagnetic signal sequences collected by the intelligent sensing device in different working modes.

[0015] In this embodiment, an intelligent sensing device in an industrial production workshop is taken as an example. There are various production devices working together in the workshop, and the intelligent sensing device is responsible for monitoring various electromagnetic data in the production environment to ensure the stability of the production process. The preset monitoring period is set to one day, and in this day, the intelligent sensing device will switch to multiple working modes according to different production processes. For example, in the device startup phase, the intelligent sensing device is in mode A, at this time it will collect electromagnetic signals generated when the device starts, forming an electromagnetic signal sequence S1; in the normal operation phase of the production device, the intelligent sensing device switches to mode B, and collects electromagnetic signals during normal operation, forming an electromagnetic signal sequence S2; in the production device maintenance and debugging phase, the intelligent sensing device is in mode C, and collects electromagnetic signals during the debugging process, forming an electromagnetic signal sequence S3. These electromagnetic signal sequences S1, S2, S3 collected in different working modes together constitute a multi-source electromagnetic signal set M. Among them, the electromagnetic signal sequence S1 contains high-frequency electromagnetic fluctuation information at the moment of device startup, S2 contains relatively stable electromagnetic signal characteristics during normal operation, and S3 contains electromagnetic signal changes in some special frequency bands during debugging.

[0016] Next, further analysis and preparation are performed on each electromagnetic signal sequence. For example, when acquiring electromagnetic signal sequence S1, it is sampled at a fixed sampling frequency f1, resulting in a series of discrete electromagnetic signal values. These values ​​are arranged in chronological order to form the complete sequence S1. Similarly, S2 is sampled at sampling frequency f2, and S3 is sampled at sampling frequency f3. Different sampling frequencies are set according to the changing characteristics of electromagnetic signals under different operating modes to ensure accurate capture of the characteristics of various electromagnetic signals.

[0017] For example, for an electromagnetic signal sequence S1, its signal value at a certain time t1 is v1, its signal value at time t2 is v2, and so on. These signal values ​​constitute a discrete sequence that changes over time. In this way, a comprehensive set of multi-source electromagnetic signals from the target intelligent sensing device under different operating modes within a preset monitoring period is obtained, providing a rich data foundation for subsequent electromagnetic interference feature extraction.

[0018] Step 120: Perform electromagnetic interference feature extraction processing on the multi-source electromagnetic signal set to obtain the electromagnetic interference feature set of the intelligent sensing device. The electromagnetic interference feature set includes the interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source association characteristics of the electromagnetic signal sequence.

[0019] In this embodiment, we continue to use intelligent sensing equipment in an industrial production workshop as an example. First, the electromagnetic signal sequence is divided into multiple signal segments, and each signal segment undergoes frequency domain transformation to obtain the corresponding spectral distribution vector. For example, for the electromagnetic signal sequence S1, it is divided according to a fixed time interval t, resulting in multiple signal segments F1, F2, F3, etc. Then, a Fourier transform algorithm is used to perform frequency domain transformation on each signal segment. Taking signal segment F1 as an example, through Fourier transform, it is converted from a series of signal values ​​in the time domain into a spectral distribution in the frequency domain. In the frequency domain, the signal strength of different frequency components is clearly displayed, forming a spectral distribution vector V1. Each element in the spectral distribution vector V1 corresponds to a specific frequency, and its value represents the strength of that frequency component in signal segment F1. For example, the element v11 in V1 corresponds to frequency f11, and its value represents the signal strength at frequency f11 in signal segment F1. The same method is applied to other signal segments F2, F3, etc., to obtain spectral distribution vectors V2, V3, etc., respectively.

[0020] Next, frequency band energy analysis is performed on the spectral distribution vector to determine the energy proportion characteristics of signal segments within each preset frequency band. The energy proportion characteristics of each signal segment are then time-series aggregated to generate the interference frequency band distribution characteristics of the electromagnetic signal sequence. The preset frequency bands are divided into multiple intervals, such as frequency band B1 (frequency range f1-f2), frequency band B2 (frequency range f3-f4), etc. For the spectral distribution vector V1, its energy proportion within frequency band B1 is calculated. By calculating the sum of squares of the signal strengths of all frequency components within frequency band B1 and comparing it with the total energy of the entire spectral distribution vector (the sum of squares of the signal strengths of all frequency components), the energy proportion p11 within frequency band B1 is obtained. Similarly, the energy proportions p12, p13, etc., within other preset frequency bands are calculated. The same calculation is performed on the spectral distribution vector V2 corresponding to signal segment F2 to obtain the energy proportions p21, p22, p23, etc., within each preset frequency band. The energy proportions of all signal segments within each preset frequency band are then aggregated in time sequence. For example, for frequency band B1, p11, p21, p31, etc., are arranged according to the time sequence of the signal segments to form a sequence showing the change of energy proportion within that frequency band over time. In this way, the interference frequency band distribution characteristic D1 of the electromagnetic signal sequence S1 is generated. The interference frequency band distribution characteristic D1 is a multi-dimensional feature set that includes the change of energy proportion over time within each preset frequency band, comprehensively reflecting the energy distribution characteristics of the electromagnetic signal sequence S1 in different frequency bands.

[0021] Then, signal strength differential processing is performed on the electromagnetic signal sequence to obtain the intensity difference sequence between adjacent signal segments. The interference intensity fluctuation characteristics are determined based on the fluctuation amplitude and frequency of the intensity difference sequence. Taking electromagnetic signal sequence S1 as an example, the intensity difference between adjacent signal segments is calculated. The average intensity of signal segment F1 is I1, and the average intensity of F2 is I2, so the intensity difference between them is ΔI1 = I2 - I1. The intensity differences between all adjacent signal segments are calculated sequentially, resulting in the intensity difference sequence ΔI = [ΔI1, ΔI2, ΔI3...]. The fluctuation amplitude is statistically analyzed through the intensity difference sequence ΔI. For example, the maximum and minimum values ​​in the sequence are identified, and the difference between them is calculated to obtain the fluctuation amplitude A. Simultaneously, the frequency of intensity changes in the intensity difference sequence is statistically analyzed, i.e., the fluctuation frequency F. The interference intensity fluctuation characteristic W1 is determined based on the fluctuation amplitude A and the fluctuation frequency F. The interference intensity fluctuation characteristic W1 describes the severity and frequency of intensity changes in the electromagnetic signal sequence, providing an important basis for judging the intensity change characteristics of electromagnetic interference.

[0022] Finally, a topology map of the environment in which the intelligent sensing device is located is obtained. Based on this topology map, a set of candidate interfering devices with electromagnetic coupling to the intelligent sensing device is identified. Correlation analysis is performed on the historical electromagnetic signals of each interfering device in the candidate set to generate interference source association characteristics. In an industrial production workshop, a topology map of the equipment deployment is obtained using the workshop's equipment layout plan and an electromagnetic environment monitoring system. The relative positional characteristics and connection medium characteristics between the intelligent sensing device and its neighboring devices are extracted from this topology map. For example, the relative positional characteristics between the intelligent sensing device and its neighboring device E1 include distance information d1 (the straight-line distance between the intelligent sensing device and device E1) and orientation information θ1 (the orientation angle of device E1 relative to the intelligent sensing device). The connection medium characteristics include a metal pipe as the transmission medium type and a medium attenuation coefficient of α1. Similarly, the relative positional characteristics and connection medium characteristics between the intelligent sensing device and other neighboring devices such as E2 and E3 are also obtained.

[0023] Step 121: Extract the relative position features and connection medium features between the intelligent sensing device and each neighboring device based on the device deployment topology map. The relative position features include distance information and orientation information, and the connection medium features include transmission medium type and medium attenuation coefficient.

[0024] In the aforementioned industrial production workshop scenario, the distance information d1 between the intelligent sensing device and device E1 is determined using measuring tools or workshop layout planning data. For example, the straight-line distance between the intelligent sensing device and device E1 is measured to be 5 meters using a laser rangefinder, i.e., d1 = 5 meters. The orientation information θ1 is determined using a compass or the workshop's coordinate positioning system. For example, if device E1 is located at a 45-degree angle northeast of the intelligent sensing device, then θ1 = 45 degrees. Regarding the characteristics of the connection medium, detection and analysis reveal that the transmission medium between the intelligent sensing device and device E1 is a metal pipe. This type of metal pipe experiences some attenuation during electromagnetic signal transmission. Through experimental measurement or relevant data research, the medium attenuation coefficient α1 is found to be 0.1 (the unit is determined based on actual conditions; in this embodiment, it is set to the attenuation coefficient per meter). Similarly, for the intelligent sensing device and device E2, the measured distance information d2 = 3 meters, the orientation information θ2 = 120 degrees, the transmission medium type is a plastic cable, and the medium attenuation coefficient α2 = 0.05. For the intelligent sensing device and device E3, the distance information d3 = 7 meters, the orientation information θ3 = 270 degrees, the transmission medium type is optical fiber, and the medium attenuation coefficient α3 = 0.01. In this way, the relative positional characteristics and connection medium characteristics between the intelligent sensing device and its neighboring devices are comprehensively and accurately extracted, providing a detailed data foundation for subsequent electromagnetic coupling strength prediction.

[0025] Step 122: Call the pre-trained electromagnetic coupling analysis model to perform coupling strength prediction processing on the relative position features and connection medium features to obtain the electromagnetic coupling strength between the intelligent sensing device and each neighboring device.

[0026] In industrial production workshops, pre-trained electromagnetic coupling analysis models are obtained through extensive training with a large amount of experimental and real-world scenario data. The model comprises an input layer, multiple hidden layers, and an output layer. The input layer receives the relative positional features (distance information d and orientation information θ) between the intelligent sensing device and its neighboring devices, as well as the characteristics of the connection medium (transmission medium type T and medium attenuation coefficient α). For example, for the intelligent sensing device and device E1, d1 = 5 meters, θ1 = 45 degrees, T1 = metal pipe, and α1 = 0.1 are input data passed to the model. Within the model, the input data first undergoes a series of calculations and feature extractions in the hidden layers. Neurons in the hidden layers process the input data using weights and activation functions. For example, a neuron receives input data x1, x2, x3, etc., and performs a weighted sum using weights w1, w2, w3, etc., followed by a transformation using an activation function such as the ReLU function (f(x) = max(0, x)) to obtain a new feature representation. After processing by multiple hidden layers, the data is passed to the output layer. Based on the previously learned patterns and features, the output layer outputs the electromagnetic coupling strength C1 between the intelligent sensing device and device E1. Similarly, for intelligent sensing devices E2, E3, etc., the corresponding relative position features and connection medium features are input into the model to obtain electromagnetic coupling strengths C2, C3, etc. These electromagnetic coupling strength values ​​reflect the strength of electromagnetic interaction between the intelligent sensing device and its neighboring devices, providing an important basis for identifying candidate interfering devices.

[0027] Step 123: Add the neighboring devices whose electromagnetic coupling strength is greater than the preset strength threshold to the candidate interference device set, and record the device type and operating frequency range of each candidate interference device.

[0028] The preset strength threshold is determined based on factors such as the actual electromagnetic environment of the industrial production workshop and the anti-interference capability of the intelligent sensing equipment. For example, the preset strength threshold is 0.5. The calculated electromagnetic coupling strengths C1, C2, and C3 are compared and judged. For example, if electromagnetic coupling strength C1 = 0.6, which is greater than the preset strength threshold of 0.5, then device E1 is added to the candidate interference device set. Simultaneously, device E1 is recorded as a production processing equipment, and its operating frequency range is f10-f20 (obtained from the equipment's technical parameter document or actual measurement). If electromagnetic coupling strength C2 = 0.3, which is less than the preset strength threshold of 0.5, then device E2 is not added to the candidate interference device set. If electromagnetic coupling strength C3 = 0.7, which is greater than the preset strength threshold of 0.5, then device E3 is added to the candidate interference device set, and its equipment type is recorded as a monitoring device, with an operating frequency range of f30-f40. In this way, a set of candidate interference devices was identified, which includes devices that may cause electromagnetic interference to smart sensing devices. Their device types and operating frequency ranges were recorded, providing clear targets for subsequent interference source correlation feature analysis.

[0029] Step 124: Obtain the historical electromagnetic signal sequence of each interference device in the candidate interference device set within the preset monitoring period.

[0030] In an industrial production workshop setting, for devices E1 and E3 in the candidate interference device set, their historical electromagnetic signal sequences within a preset monitoring period (one day) are obtained through the workshop's electromagnetic signal monitoring system or the devices' own recording functions. For example, device E1 collects electromagnetic signals at a sampling frequency f4 within the preset monitoring period, obtaining a historical electromagnetic signal sequence H1. Sequence H1 contains the electromagnetic signal values ​​of device E1 at different times, such as the signal value at time t1 being h11, the signal value at time t2 being h12, and so on. Similarly, device E3 collects signals at a sampling frequency f5, obtaining a historical electromagnetic signal sequence H3. The historical electromagnetic signal sequences H1 and H3 record the changes in the electromagnetic signals of the candidate interference devices within the preset monitoring period, providing data support for subsequent correlation analysis.

[0031] Step 125: Perform time-frequency alignment processing on the electromagnetic signal sequence of the intelligent sensing device and the historical electromagnetic signal sequence of each interfering device to obtain multiple aligned signal pairs.

[0032] Taking the electromagnetic signal sequence S1 from the intelligent sensing device and the historical electromagnetic signal sequence H1 from the candidate interference device E1 as examples, time-frequency alignment processing is performed. First, time alignment is performed. Since the sampling times of the two sequences may not be completely consistent, time interpolation or time synchronization algorithms are used to ensure that the two sequences have the same reference starting point and time interval. For example, if the sampling time interval of the electromagnetic signal sequence S1 is t1 and the sampling time interval of the historical electromagnetic signal sequence H1 is t2, a linear interpolation algorithm is used to adjust the time interval of H1 to be the same as t1 of S1, so that the two sequences correspond one-to-one in time. Next, frequency alignment is performed. For different electromagnetic signals, their frequency distributions may differ. Through spectrum analysis and frequency adjustment algorithms, the frequency range and frequency resolution of the two sequences are unified. For example, if the frequency range of S1 is f50-f60 and the frequency range of H1 is f55-f65, a frequency resampling algorithm is used to adjust the frequency range of H1 to be the same as f50-f60 of S1, and the frequency resolution of the two sequences is made the same. After time-frequency alignment processing, multiple aligned signal pairs are obtained. For example, at a certain time t, the signal value s1 in S1 and the aligned signal value h1 in H1 form an aligned signal pair (s1, h1). The same method is applied to the electromagnetic signal sequence S1 of the intelligent sensing device and the historical electromagnetic signal sequence H3 of the candidate interference device E3, as well as other possible combinations of electromagnetic signal sequences, to obtain multiple aligned signal pairs, providing a suitable data format for subsequent cross-correlation analysis.

[0033] Step 126: Perform cross-correlation analysis on each aligned signal pair to determine the correlation coefficient between the electromagnetic signal sequence of the intelligent sensing device and the historical electromagnetic signal sequence of each interfering device, and extract the interfering devices with a correlation coefficient greater than a preset correlation threshold as target interfering devices.

[0034] For each aligned signal pair, a cross-correlation analysis algorithm is used to calculate the correlation coefficient between the electromagnetic signal sequence of the intelligent sensing device and the historical electromagnetic signal sequences of each interfering device. Taking the aligned signal pair of the electromagnetic signal sequence S1 of the intelligent sensing device and the historical electromagnetic signal sequence H1 of the candidate interfering device E1 as an example, the cross-correlation analysis algorithm obtains the correlation coefficient by calculating the similarity of the two signal sequences under different time delays. For example, if the aligned signal pair is (s1, h1), the correlation value of s1 and h1 under the time delay τ is calculated. By calculating for all possible time delays and taking the average or maximum value, a correlation coefficient r1 is obtained. Similarly, for the aligned signal pair of the electromagnetic signal sequence S1 of the intelligent sensing device and the historical electromagnetic signal sequence H3 of the candidate interfering device E3, the correlation coefficient r3 is calculated. The preset correlation threshold is determined based on actual conditions and experience and is set to 0.6. The calculated correlation coefficients r1, r3, etc. are compared with the preset correlation threshold. If r1 = 0.7, which is greater than the preset correlation threshold of 0.6, then device E1 is considered the target interfering device. If r3 = 0.5, which is less than the preset correlation threshold of 0.6, then device E3 is not considered a target interference device. In this way, target interference devices with high correlation to the electromagnetic signal sequences of intelligent sensing devices are extracted; these target interference devices are more likely to be the source of electromagnetic interference.

[0035] Step 127: Generate the interference source association features based on the device type, operating frequency range, and electromagnetic coupling strength between the target interference device and the intelligent sensing device.

[0036] For a identified target jamming device E1, its device type is a manufacturing / processing equipment, its operating frequency range is f10-f20, and its electromagnetic coupling strength with the intelligent sensing device is C1. This information is integrated to generate interference source correlation features. Interference source correlation features are a multi-dimensional feature set that includes information such as the target jamming device's device type, operating frequency range, and electromagnetic coupling strength. For example, interference source correlation features can be represented as a vector V = [device type (manufacturing / processing equipment), operating frequency range (f10-f20), electromagnetic coupling strength (C1)]. In this way, interference source correlation features are generated that accurately reflect the correlation between the intelligent sensing device and the target jamming device, providing an important basis for subsequent interference suppression strategy formulation.

[0037] Step 130: Based on the preset interference suppression model, perform interference pattern matching processing on the electromagnetic interference feature set to generate the electromagnetic interference suppression strategy of the intelligent sensing device. The electromagnetic interference suppression strategy includes a frequency band adjustment direction corresponding to the interference frequency band distribution characteristics, a signal gain compensation direction corresponding to the interference intensity fluctuation characteristics, and an anti-interference filtering direction corresponding to the interference source association characteristics.

[0038] In an industrial production workshop setting, the interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source correlation characteristics are first fused to generate a comprehensive interference feature vector for the intelligent sensing device. Taking the interference frequency band distribution characteristics D1, interference intensity fluctuation characteristics W1, and interference source correlation characteristics V of the previously obtained electromagnetic signal sequence S1 as an example, these features are concatenated and fused, for example, the comprehensive interference feature vector I = [D1, W1, V]. This comprehensive interference feature vector I contains information on multiple aspects of electromagnetic interference, including frequency band distribution, intensity fluctuation, and interference source correlation.

[0039] Next, the interference pattern matching layer in the interference suppression model is invoked to perform pattern recognition processing on the comprehensive interference feature vector to determine the target interference mode of the electromagnetic interference currently affecting the intelligent sensing device. The target interference mode includes at least one of the following: frequency band conflict mode, signal overload mode, and external coupling mode. The interference pattern matching layer in the interference suppression model is a neural network layer trained on a large amount of data. It receives the comprehensive interference feature vector I as input. Inside the layer, neurons analyze and match the input features using learned weights and patterns. For example, neurons compare each feature in the comprehensive interference feature vector I with various pre-stored interference mode templates and calculate the similarity. If the interference frequency band distribution features in the comprehensive interference feature vector I show that some frequency bands have excessively high energy and overlap with the operating frequency bands of other devices, and the interference intensity fluctuation features show unstable signal strength, the interference pattern matching layer determines that the target interference mode of the electromagnetic interference currently affecting the intelligent sensing device is either a frequency band conflict mode or a signal overload mode.

[0040] Then, based on the target interference mode, an initial suppression strategy is obtained from a pre-defined suppression strategy library. This initial suppression strategy is then optimized based on the interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source association characteristics to generate an electromagnetic interference suppression strategy. The pre-defined suppression strategy library is a pre-built set of strategies containing initial strategies for different interference modes. For example, when the target interference mode is a frequency band conflict mode, a frequency band adjustment strategy corresponding to the interference frequency band distribution characteristics is obtained from the suppression strategy library. This frequency band adjustment strategy includes a priority avoidance frequency band set and a backup working frequency band set. For example, the priority avoidance frequency band set is B.priority=[f11-f12, f21-f22], and the backup working frequency band set is B.alternate=[f31-f32, f41-f42]. When the target interference mode is a signal overload mode, a gain compensation strategy corresponding to the interference intensity fluctuation characteristics is obtained from the suppression strategy library. This gain compensation strategy includes a gain adjustment curve and a maximum gain limit threshold. For example, the gain adjustment curve can be represented by the function G(x), and the maximum gain limit threshold is G.max. When the target interference mode is an external coupling mode, the filtering and suppression strategy corresponding to the characteristics associated with the interference source is obtained from the suppression strategy library. The filtering and suppression strategy includes adaptive filter type selection and filter parameter configuration rules.

[0041] Step 131: Perform feature fusion processing on the interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source correlation characteristics to generate a comprehensive interference feature vector for the intelligent sensing device.

[0042] In the industrial production workshop scenario of this embodiment, the feature fusion process is further described in detail. The interference frequency band distribution feature D1 is a multi-dimensional dataset that records the energy proportion of different preset frequency bands at different time points. For example, D1 can be represented as a two-dimensional array, where rows represent different preset frequency bands and columns represent different time points. For example, if there are n preset frequency bands and m time points, then D1[i][j] represents the energy proportion of the i-th preset frequency band at the j-th time point. The interference intensity fluctuation feature W1 includes two important parameters: fluctuation amplitude A and fluctuation frequency F. The interference source association feature V is represented in vector form and includes information such as the device type, operating frequency range, and electromagnetic coupling strength of the target interference device.

[0043] To perform feature fusion, the interference frequency band distribution feature D1 is first processed. A statistical method can be used, such as calculating the average energy percentage of each preset frequency band throughout the monitoring period, to obtain a one-dimensional array D1.avg with length n, where D1.avg[i] represents the average energy percentage of the i-th preset frequency band. Then, D1.avg, the fluctuation amplitude A and fluctuation frequency F in the interference intensity fluctuation feature W1, and the interference source association feature V are concatenated. For example, the comprehensive interference feature vector I = [D1.avg[0], D1.avg[1], ..., D1.avg[n], A, F, equipment type, operating frequency range, electromagnetic coupling strength]. In this way, different types of electromagnetic interference features are effectively fused together to form a comprehensive interference feature vector that fully reflects the electromagnetic interference situation, providing rich and accurate feature inputs for subsequent interference pattern matching.

[0044] Step 132: Call the interference pattern matching layer in the interference suppression model to perform pattern recognition processing on the integrated interference feature vector to determine the target interference mode of the electromagnetic interference currently experienced by the intelligent sensing device. The target interference mode includes at least one of frequency band conflict mode, signal overload mode and external coupling mode.

[0045] The interference pattern matching layer in the interference suppression model is built on a deep learning architecture. It consists of multiple convolutional layers, pooling layers, and fully connected layers. When the comprehensive interference feature vector I is input into the interference pattern matching layer, it first enters the convolutional layer. The convolutional kernels in the convolutional layer extract features from the comprehensive interference feature vector I. For example, the convolutional kernel can be viewed as a small weight matrix that slides on the comprehensive interference feature vector I, extracting local features through weighted summation. Different convolutional kernels can extract different types of features; for example, some convolutional kernels specifically extract features related to frequency band distribution, while others extract features related to interference intensity fluctuations.

[0046] After being processed by the convolutional layers, the data enters the pooling layers. The purpose of the pooling layers is to downsample the data, reducing the amount of data while retaining important features. For example, max pooling can be used to select the maximum value within a local region as the representative value for that region. Through pooling layers, the dimensionality of the data can be reduced, improving the computational efficiency and generalization ability of the model.

[0047] After multiple processing steps by convolutional and pooling layers, the data enters the fully connected layer. The fully connected layer comprehensively analyzes and matches the various features extracted earlier. Neurons in the fully connected layer are connected to all neurons in the preceding layers. The input features are weighted and summed using learned weights, and then transformed using an activation function. For example, a neuron in the fully connected layer receives multiple feature inputs x1, x2, ..., xk from the preceding layers, with weights w1, w2, ..., wk. After weighted summation, we get s = w1*x1 + w2*x2 + ... + wk*xk. This summation is then transformed using an activation function such as the sigmoid function f(s) = 1 / (1 + exp(-s)) to obtain the neuron's output.

[0048] The output of the fully connected layer is a probability distribution vector, with each element corresponding to an interference mode. For example, the output vector might be P = [p1, p2, p3], where p1 represents the probability of a band conflict mode, p2 represents the probability of a signal overload mode, and p3 represents the probability of an external coupling mode. By comparing these probability values, the interference mode with the highest probability is selected as the target interference mode for the electromagnetic interference currently experienced by the intelligent sensing device. If p1 is the largest, the target interference mode is a band conflict mode; if p2 is the largest, it is a signal overload mode; and if p3 is the largest, it is an external coupling mode. If multiple probability values ​​are similar and relatively large, the target interference mode may include multiple corresponding modes, such as simultaneously including both a band conflict mode and a signal overload mode.

[0049] Step 133: Obtain an initial suppression strategy from a preset suppression strategy library based on the target interference mode.

[0050] When the target interference mode is a frequency band conflict mode, a frequency band adjustment strategy corresponding to the interference frequency band distribution characteristics is obtained from the suppression strategy library. In an industrial production workshop scenario, for example, through analysis of the interference frequency band distribution characteristics D1, it is found that the intelligent sensing device has frequency band conflicts with other devices in frequency bands f11-f12 and f21-f22, resulting in electromagnetic interference. The priority avoidance frequency band set B.priority=[f11-f12, f21-f22] and the alternative working frequency band set B.alternate=[f31-f32, f41-f42] are obtained from the frequency band adjustment strategy library. These frequency bands are determined based on factors such as the working frequency band distribution of other devices in the workshop and the available frequency band range of the intelligent sensing device itself, which are pre-set in the suppression strategy library.

[0051] When the target interference mode is a signal overload mode, a gain compensation strategy corresponding to the interference intensity fluctuation characteristics is retrieved from the suppression strategy library. For example, analysis of the interference intensity fluctuation characteristic W1 reveals that the electromagnetic signal intensity fluctuation amplitude is large and there is a signal overload. The gain compensation strategies in the suppression strategy library include the gain adjustment curve G(x) and the maximum gain limit threshold G.max. The gain adjustment curve G(x) is determined based on the signal intensity fluctuation pattern and the amplifier characteristics of the intelligent sensing device, describing the amount of gain adjustment that should be given under different signal intensities. The maximum gain limit threshold G.max is set to prevent excessive signal gain from causing equipment damage or signal distortion.

[0052] When the target interference mode is external coupling mode, a filtering and suppression strategy corresponding to the characteristics associated with the interference source is obtained from the suppression strategy library. Taking the target interference device E1 as an example, its device type is a manufacturing and processing equipment, its operating frequency range is f10-f20, and its electromagnetic coupling strength with the intelligent sensing device is C1. The filtering and suppression strategies obtained from the suppression strategy library include adaptive filter type selection and filter parameter configuration rules. Based on information such as the operating frequency range and electromagnetic coupling strength of the target interference device, an appropriate adaptive filter type is selected. For example, if the interference generated by electromagnetic coupling is mainly concentrated in the high-frequency band, a high-pass filter can be selected. The filter parameter configuration rules are determined based on factors such as the electromagnetic coupling strength and the signal quality requirements of the intelligent sensing device, such as determining the filter's cutoff frequency and stopband attenuation.

[0053] Step 134: Based on the interference frequency band distribution characteristics, interference intensity fluctuation characteristics and interference source correlation characteristics, perform parameter optimization processing on the initial suppression strategy to generate the electromagnetic interference suppression strategy.

[0054] When the initial suppression strategy is a frequency band adjustment strategy, the priority avoidance frequency band set is sorted according to the energy proportion characteristics in the interference frequency band distribution characteristics, and the backup working frequency band set is expanded based on the available frequency band resources of the intelligent sensing device. In the industrial production workshop scenario, for the interference frequency band distribution characteristic D1, the energy proportion of each preset frequency band is further analyzed. For example, the total energy proportion of each preset frequency band in the entire monitoring period is calculated. For example, if there are n preset frequency bands, the total energy proportion of the i-th preset frequency band is Ei. The frequency bands in the priority avoidance frequency band set are sorted according to the value of Ei. If E1>E2, and frequency bands f11-f12 correspond to E1 and frequency bands f21-f22 correspond to E2, then frequency bands f11-f12 are placed at the front of the priority avoidance frequency band set so that frequency bands with higher energy proportions are avoided first when adjusting the frequency band.

[0055] Meanwhile, expand the set of alternate operating frequency bands based on the available frequency band resources of the intelligent sensing device. The available frequency band resources of the intelligent sensing device can be obtained through the device's technical parameter document or actual measurement. For example, the available frequency band range is F.available = [f51 - f52, f61 - f62, …]. Check the relationship between the set of alternate operating frequency bands B.alternate and the available frequency band resources F.available, and add the frequency bands in F.available that are not included in B.alternate and do not conflict with the operating frequency bands of other devices to B.alternate, thereby expanding the set of alternate operating frequency bands, providing more frequency band options for the intelligent sensing device, and better coping with the frequency band conflict problem.

[0056] When the initial suppression strategy is the gain compensation strategy, adjust the slope of the gain adjustment curve and the compensation delay time according to the fluctuation amplitude and fluctuation frequency in the interference intensity fluctuation characteristics, and update the maximum gain limit threshold based on the maximum load capacity of the intelligent sensing device. For the fluctuation amplitude A and fluctuation frequency F in the interference intensity fluctuation characteristics W1, analyze their influence on the signal gain. If the fluctuation amplitude A is large, it means that the signal intensity changes violently. To better compensate the signal intensity, it is necessary to appropriately increase the slope of the gain adjustment curve G(x) so that the gain can be adjusted more quickly when the signal intensity changes. For example, the original slope of the gain adjustment curve is k1, and according to the size of the fluctuation amplitude A, the slope is adjusted to k2 (k2 > k1).

[0057] Meanwhile, adjust the compensation delay time according to the fluctuation frequency F. If the fluctuation frequency F is high, it means that the signal intensity changes frequently. To avoid excessive or lagging gain adjustment, it is necessary to shorten the compensation delay time. For example, the original compensation delay time is t1, and it is adjusted to t2 (t2 < t1) according to the fluctuation frequency F. In addition, update the maximum gain limit threshold G.max based on the maximum load capacity of the intelligent sensing device. By detecting the hardware parameters and performance indicators of the intelligent sensing device, determine the signal intensity value S.max corresponding to its maximum load capacity. According to S.max and the current interference situation, reasonably adjust the maximum gain limit threshold G.max, for example, adjust G.max from the original value to G.max.new, to ensure that the signal will not exceed the maximum load capacity of the intelligent sensing device during gain compensation, and avoid device damage or signal distortion.

[0058] When the initial suppression strategy is a filtering suppression strategy, the target filter type is selected based on the device type and operating frequency range in the interference source correlation characteristics, and the cutoff frequency and stopband attenuation parameters of the target filter are configured based on the electromagnetic coupling strength and the correlation coefficient of historical electromagnetic signals. Taking the target interfering device E1 as an example, its device type is a manufacturing equipment, and its operating frequency range is f10-f20. Based on the device type and operating frequency range, and considering the electromagnetic interference characteristics and frequency range that the manufacturing equipment may generate, a suitable target filter type is selected. For example, if the interference is mainly concentrated in the f10-f20 frequency band, and the interfering signal has certain high-frequency characteristics, a band-stop filter can be selected as the target filter type to effectively suppress the interference signal in this frequency band.

[0059] Then, the cutoff frequency and stopband attenuation parameters of the target filter are configured based on the electromagnetic coupling strength C1 and the historical electromagnetic signal correlation coefficient r1. The electromagnetic coupling strength C1 reflects the strength of the electromagnetic interaction between the intelligent sensing device and the target interference device, while the historical electromagnetic signal correlation coefficient r1 reflects the similarity between their electromagnetic signals. If both C1 and r1 are large, it indicates severe interference, requiring more stringent filtering. For example, the cutoff frequency of the target filter is set to f.cutoff, so that it exactly covers the operating frequency range f10-f20 of the target interference device. Furthermore, the stopband attenuation parameters are appropriately configured according to the magnitude of the electromagnetic coupling strength and correlation coefficient, ensuring sufficient attenuation of the interference signal near the cutoff frequency. For instance, if the original stopband attenuation parameter is A1, it is adjusted to A2 (A2>A1) based on the values ​​of C1 and r1 to improve the filter's ability to suppress interference signals.

[0060] Step 140: Perform interference suppression operation on the signal acquisition module of the intelligent sensing device according to the electromagnetic interference suppression strategy to reduce the electromagnetic signal distortion rate of the intelligent sensing device.

[0061] When the electromagnetic interference suppression strategy includes frequency band adjustment, the control signal acquisition module switches to the target operating frequency band in the set of backup operating frequency bands and monitors the interference energy changes of the target operating frequency band in real time to trigger the frequency band switching decision. In an industrial production workshop scenario, for example, if the set of backup operating frequency bands determined by the electromagnetic interference suppression strategy is B.alternate=[f31-f32, f41-f42], the target operating frequency band is selected from B.alternate based on factors such as the current operating status of the intelligent sensing device and signal transmission quality, for example, f31-f32. The operating frequency band of the signal acquisition module is then switched to f31-f32 through the hardware circuitry or software settings of the control signal acquisition module.

[0062] After switching to the target operating frequency band, the interference energy changes within the target operating frequency band are monitored in real time. Multiple monitoring sub-bands are set within the target operating frequency band f31-f32, for example, dividing f31-f32 into two monitoring sub-bands: f31-f31.5 and f31.5-f32. The current interference energy value of each monitoring sub-band is periodically collected by adding a dedicated interference energy monitoring circuit to the signal acquisition module or by using software algorithms. For example, data is collected every time interval t; at time t1, the current interference energy value for monitoring sub-band f31-f31.5 is E1, and the current interference energy value for monitoring sub-band f31.5-f32 is E2.

[0063] When the current interference energy value of any monitored sub-band exceeds a preset energy threshold, a band switching request signal is generated, and the candidate switching band with the lowest interference energy value is selected from the set of backup working bands. For example, if the preset energy threshold is E.threshold, a band switching request signal is generated when E1 > E.threshold. Then, the interference energy values ​​of other bands in the set of backup working bands B.alternate are measured and compared. For example, the interference energy value of band f41-f42 is measured to be E3, and E3 is compared with the interference energy values ​​of other possible bands to determine f41-f42 as the candidate switching band with the lowest interference energy value.

[0064] Based on the signal transmission quality assessment results and handover time prediction results of the candidate switching frequency bands, the final switching frequency band is determined, and the signal acquisition module is controlled to switch to the final switching frequency band to complete the interference suppression operation. Signal transmission quality is assessed for the candidate switching frequency bands f41-f42 by transmitting and receiving test signals, measuring indicators such as bit error rate (BER) and signal-to-noise ratio (SNR) to evaluate signal transmission quality. For example, the assessed BER is denoted as BER, and the SNR as SNR. Simultaneously, the handover time required to switch the signal acquisition module to the f41-f42 frequency band is predicted. For example, through analysis of the hardware circuit switching time and software configuration time, the predicted handover time is T.switch. Considering both the signal transmission quality assessment results and the handover time prediction results, if the BER is low and the SNR is high, and the T.switch is within an acceptable range, then f41-f42 is determined as the final switching frequency band. By controlling the relevant hardware and software of the signal acquisition module, the signal acquisition module is switched to the f41-f42 frequency band to complete the interference suppression operation, thereby reducing the electromagnetic interference experienced by the intelligent sensing device in this frequency band and reducing the electromagnetic signal distortion rate.

[0065] When the electromagnetic interference suppression strategy includes signal gain compensation, the amplifier in the signal acquisition module is controlled for gain compensation according to the gain adjustment curve, and the compensated signal is limited based on the maximum gain limit threshold. In an industrial production workshop scenario, for example, the gain adjustment curve is G(x), which is a function of the signal strength x. When the signal acquisition module acquires an electromagnetic signal, it first measures the current signal strength value x0. Based on the gain adjustment curve G(x), the gain value G(x0) that should be given is calculated. For example, when x0=10, by querying the gain adjustment curve G(x), we get G(10)=2, which means that the signal strength needs to be amplified by 2 times. By controlling the amplifier circuit in the signal acquisition module, the acquired signal strength is compensated for gain according to G(x0).

[0066] Then, the compensated signal is amplitude-limited based on the maximum gain limit threshold G.max. For example, if the maximum gain limit threshold G.max = 20, the signal strength after gain compensation becomes y = x0 * G(x0) = 10 * 2 = 20. If y > G.max, the signal strength is limited to G.max, meaning the final output signal strength is 20, preventing excessive signal strength from damaging the equipment or causing signal distortion. In this way, reasonable gain compensation control is applied to the amplifier in the signal acquisition module, and the compensated signal is amplitude-limited, effectively improving signal quality and reducing electromagnetic signal distortion caused by signal overload.

[0067] When the electromagnetic interference suppression strategy includes anti-interference filtering, a target filter is inserted into the preprocessing circuit of the signal acquisition module, and the electromagnetic signal is filtered in real time according to the cutoff frequency and stopband attenuation parameters. In an industrial production workshop scenario, for example, a band-stop filter is selected with a cutoff frequency of f.cutoff and a stopband attenuation parameter of A. The target filter is inserted into the preprocessing circuit of the signal acquisition module, which is located before the signal acquisition module performs further processing on the electromagnetic signal.

[0068] When an electromagnetic signal enters the signal acquisition module, it first passes through a target filter. The target filter filters the electromagnetic signal based on its cutoff frequency, f.cutoff. For electromagnetic signals with frequencies near the cutoff frequency f.cutoff, attenuation is achieved according to the stopband attenuation parameter A. For example, for an interference signal with a frequency of f.cutoff and an original strength of I0, after passing through the target filter, the signal strength is attenuated to I1 = I0 * 10^(-A / 20) (according to the filter's attenuation formula). This real-time filtering effectively suppresses electromagnetic interference signals generated by the target interference device, reducing the electromagnetic signal distortion rate of the intelligent sensing device. During real-time filtering, the characteristics of the electromagnetic signal after filtering are continuously monitored, such as the signal's spectral distribution and intensity changes. If significant interference is still found in the filtered signal or if signal loss in certain key frequency bands is excessive, the parameter settings of the target filter will be re-evaluated based on the actual situation.

[0069] For example, if significant residual interference signals are detected within a specific frequency band, the cutoff frequency (f.cutoff) may be adjusted to more accurately cover the interference band, or the stopband attenuation parameter (A) may be further increased to enhance the suppression capability of interference signals in that frequency band. Simultaneously, the filter parameters will be adjusted comprehensively, taking into account the sensitivity of the intelligent sensing device to signals in different frequency bands and the signal requirements of subsequent processing modules, to ensure that while effectively suppressing interference, the integrity and accuracy of the useful signal are preserved to the greatest extent possible.

[0070] Throughout the interference suppression operation, the monitoring and adjustments of each of the above steps are continuously repeated. For frequency band adjustments, the interference energy changes and signal transmission quality of the newly switched frequency band are continuously monitored. Once the interference energy exceeds the threshold or the signal transmission quality deteriorates, the frequency band switching decision is triggered again, and a more suitable frequency band is selected from the set of backup working frequency bands for switching.

[0071] For signal gain compensation, the system tracks signal strength fluctuations in real time and dynamically adjusts the gain adjustment curve and maximum gain limit threshold based on changes in interference strength fluctuation characteristics. For example, if the interference strength fluctuation characteristics show a change in the frequency of signal strength fluctuations, the response speed of gain compensation will be adjusted accordingly to better adapt to rapid changes in signal strength, ensuring that the signal remains within a suitable strength range and reducing distortion.

[0072] For anti-interference filtering, various parameters of the filtered signal are continuously analyzed. Based on changes in the correlation characteristics of interference sources and dynamic changes in the electromagnetic environment, the type, cutoff frequency, and stopband attenuation parameters of the target filter are adjusted in a timely manner. For example, if a new interference source is detected or the operating frequency band of an existing interference source changes, a more suitable filter type is re-evaluated and selected, and the cutoff frequency and stopband attenuation parameters are adjusted to ensure effective filtering and suppression of new interference situations.

[0073] Through continuous monitoring, analysis, and adjustment, the application of electromagnetic interference suppression strategies in the signal acquisition module of intelligent sensing devices is constantly optimized, thereby continuously reducing the electromagnetic signal distortion rate of intelligent sensing devices and ensuring that intelligent sensing devices can stably and accurately acquire electromagnetic signals, providing reliable data support for subsequent industrial production monitoring and control operations based on these signals.

[0074] During actual operation, relevant data for each interference suppression operation will be recorded, including interference frequency band distribution characteristics, interference intensity fluctuation characteristics, interference source correlation characteristics, suppression strategies adopted, and the final electromagnetic signal distortion rate. This data will be stored in the local storage module of the intelligent sensing device or uploaded to a remote server for further analysis.

[0075] By analyzing a large amount of historical data, we can summarize the patterns and characteristics of electromagnetic interference (EMI) in different industrial production scenarios. For example, we can discover that certain production equipment is prone to generating EMI of specific frequencies and intensities during specific operating phases, as well as the changing trends of EMI across different seasons and time periods. Based on these analytical results, we can optimize and adjust the interference suppression strategies for intelligent sensing devices in advance.

[0076] For example, if historical data predicts a high incidence of frequency conflicts during peak production periods, the set of backup operating frequency bands can be expanded in advance, or the frequency band switching algorithm can be optimized to make it faster and more accurate. For signal overload issues, based on the patterns of interference intensity fluctuations in historical data, gain adjustment curves and maximum gain limit thresholds can be adjusted in advance to better cope with upcoming signal strength changes. For anti-interference filtering, appropriate types of filters can be reserved in advance based on the changing trends of interference sources, and filter parameter configuration rules can be optimized to ensure a rapid and effective filtering response when new interference sources appear.

[0077] Furthermore, this historical data can be used to evaluate the long-term effectiveness of interference suppression strategies. By comparing the changes in electromagnetic signal distortion rate after adopting the same or similar suppression strategies over different time periods, it can be analyzed whether the strategies need further improvement. If it is found that the effectiveness of a certain suppression strategy gradually declines after a period of time, it may be due to changes in the electromagnetic environment or changes in the performance of the intelligent sensing device itself. In this case, the strategy will be optimized and adjusted in a targeted manner based on the data analysis results.

[0078] Meanwhile, this data also provides important information for the overall electromagnetic environment management of industrial production workshops. By comprehensively analyzing the electromagnetic interference data collected by all intelligent sensing devices, the distribution of electromagnetic interference throughout the workshop can be understood, and the main sources and propagation paths of electromagnetic interference can be identified. This allows for more comprehensive measures to improve the electromagnetic environment, such as adjusting equipment layout, replacing equipment with better electromagnetic compatibility, and optimizing transmission lines, fundamentally reducing the impact of electromagnetic interference on intelligent sensing devices and further improving the stability and reliability of the entire industrial production system.

[0079] In the routine maintenance and upgrades of intelligent sensing devices, this historical data and the analysis results based on it are also of great significance. Maintenance personnel can understand the performance of intelligent sensing devices under different operating conditions based on historical data, and discover potential faults in advance. For example, if it is found that the electromagnetic signal distortion rate of a certain intelligent sensing device continues to rise over a period of time, and various interference suppression strategies are ineffective, it may indicate that there is a fault in the signal acquisition module or other components inside the device, and maintenance personnel can promptly check and repair it.

[0080] For upgrading intelligent sensing devices, R&D personnel can leverage historical data and analysis results to understand users' needs and challenges regarding interference suppression functionality during actual use. For example, if existing interference suppression strategies prove ineffective in certain complex electromagnetic environments, R&D personnel can specifically improve interference suppression models and algorithms, or optimize hardware design, to enhance the anti-interference capabilities and performance of intelligent sensing devices. By continuously summarizing experience, optimizing strategies, and upgrading equipment, intelligent sensing devices can operate stably and continuously in the electromagnetic environment of industrial production workshops, providing strong support for the efficient and safe operation of industrial production.

[0081] While equipment types, production processes, and electromagnetic environments may differ across industrial production workshops, the aforementioned data-driven analysis and optimization methods allow for the tailoring of the most suitable electromagnetic interference suppression solutions for each workshop. For instance, in electronics manufacturing workshops, the dense equipment and high precision requirements for electromagnetic signals necessitate more refined frequency band adjustment strategies and higher-performance anti-interference filters. In chemical production workshops, the presence of high-power equipment necessitates a focus on signal overload issues, requiring more flexible gain compensation strategies.

[0082] For newly built industrial production workshops, these historical data and analysis results can be referenced during the equipment selection and layout planning stages. By understanding in advance the potential electromagnetic interference from different equipment and the corresponding suppression strategies, equipment with better electromagnetic compatibility can be selected, and equipment layout and signal transmission lines can be rationally planned to reduce the generation of electromagnetic interference from the source and lower the risk of interference to intelligent sensing devices.

[0083] This embodiment first acquires a set of multi-source electromagnetic signals from the target intelligent sensing device under different operating modes within a preset monitoring period. This comprehensively covers the electromagnetic signal situation under various operating conditions of the device, providing a rich and realistic data foundation for subsequent analysis. Next, electromagnetic interference feature extraction processing is performed on the multi-source electromagnetic signal set to accurately obtain an electromagnetic interference feature set including interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source correlation characteristics. This allows for in-depth analysis of electromagnetic interference characteristics from multiple key dimensions, resulting in a more comprehensive and detailed understanding of the interference situation. Interference mode matching processing is performed based on a preset interference suppression model, which can fully explore the potential connections between electromagnetic interference features, thereby generating a precise and targeted electromagnetic interference suppression strategy. Finally, interference suppression operations are performed on the signal acquisition module according to this strategy, effectively reducing the electromagnetic signal distortion rate of the intelligent sensing device, greatly improving the accuracy and reliability of signal acquisition, ensuring the stable and efficient operation of the intelligent sensing device in complex electromagnetic environments, and guaranteeing its accurate perception and processing of various types of electromagnetic information.

[0084] Please refer to the following: Figure 2 This application also provides an electromagnetic interference suppression system 100 for intelligent sensing devices, including a processor 111, a memory 112, and a bus 113 connected to the processor 111. The processor 111 and the memory 112 communicate with each other via the bus 113. The processor 111 is used to call program instructions in the memory 112 to execute the aforementioned AI-based intelligent sensing device electromagnetic interference suppression method.

[0085] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, product, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, product, or system that includes that element.

[0086] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0087] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for suppressing electromagnetic interference in intelligent sensing devices based on AI recognition, characterized in that, include: Acquire a set of multi-source electromagnetic signals from the target intelligent sensing device within a preset monitoring period, wherein the set of multi-source electromagnetic signals includes electromagnetic signal sequences collected by the intelligent sensing device under different working modes; Electromagnetic interference feature extraction processing is performed on the multi-source electromagnetic signal set to obtain the electromagnetic interference feature set of the intelligent sensing device. The electromagnetic interference feature set includes the interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source association characteristics of the electromagnetic signal sequence. Based on a preset interference suppression model, interference pattern matching processing is performed on the electromagnetic interference feature set to generate an electromagnetic interference suppression strategy for the intelligent sensing device. The electromagnetic interference suppression strategy includes a frequency band adjustment direction corresponding to the interference frequency band distribution characteristics, a signal gain compensation direction corresponding to the interference intensity fluctuation characteristics, and an anti-interference filtering direction corresponding to the interference source association characteristics. The electromagnetic interference suppression strategy is applied to the signal acquisition module of the intelligent sensing device to reduce the electromagnetic signal distortion rate of the intelligent sensing device.

2. The method according to claim 1, characterized in that, Electromagnetic interference feature extraction processing is performed on the multi-source electromagnetic signal set to obtain the electromagnetic interference feature set of the intelligent sensing device, including: The electromagnetic signal sequence is divided into multiple signal segments, and frequency domain transformation is performed on each signal segment to obtain the spectral distribution vector corresponding to the signal segment. The frequency band energy analysis processing is performed on the spectrum distribution vector to determine the energy proportion characteristics of the signal segment in each preset frequency band, and the energy proportion characteristics of each signal segment are time-series aggregated to generate the interference frequency band distribution characteristics of the electromagnetic signal sequence. The electromagnetic signal sequence is subjected to signal strength differential processing to obtain the intensity difference sequence between adjacent signal segments in the electromagnetic signal sequence. The interference intensity fluctuation characteristics are determined based on the fluctuation amplitude and fluctuation frequency of the intensity difference sequence. Obtain the device deployment topology map of the environment where the intelligent sensing device is located, determine the set of candidate interference devices that have electromagnetic coupling relationship with the intelligent sensing device based on the device deployment topology map, perform correlation analysis on the historical electromagnetic signals of each interference device in the candidate interference device set, and generate the interference source association features.

3. The method according to claim 2, characterized in that, Based on the device deployment topology, a set of candidate interference devices that have electromagnetic coupling relationships with the intelligent sensing device is determined, including: Based on the device deployment topology map, the relative position features and connection medium features between the intelligent sensing device and each neighboring device are extracted. The relative position features include distance information and orientation information, and the connection medium features include transmission medium type and medium attenuation coefficient. The pre-trained electromagnetic coupling analysis model is invoked to perform coupling strength prediction processing on the relative position features and connection medium features to obtain the electromagnetic coupling strength between the intelligent sensing device and each neighboring device. Add neighboring devices whose electromagnetic coupling strength is greater than a preset strength threshold to the candidate interference device set, and record the device type and operating frequency range of each candidate interference device.

4. The method according to claim 3, characterized in that, Correlation analysis is performed on the historical electromagnetic signals of each interfering device in the candidate interfering device set to generate the interference source association features, including: Obtain the historical electromagnetic signal sequence of each interference device in the candidate interference device set within the preset monitoring period; The electromagnetic signal sequence of the intelligent sensing device is time-frequency aligned with the historical electromagnetic signal sequence of each interfering device to obtain multiple aligned signal pairs. Cross-correlation analysis is performed on each aligned signal pair to determine the correlation coefficient between the electromagnetic signal sequence of the intelligent sensing device and the historical electromagnetic signal sequence of each interfering device, and interfering devices with a correlation coefficient greater than a preset correlation threshold are extracted as target interfering devices. The interference source association features are generated based on the device type, operating frequency range, and electromagnetic coupling strength between the target interference device and the intelligent sensing device.

5. The method according to claim 1, characterized in that, Based on a preset interference suppression model, interference pattern matching processing is performed on the electromagnetic interference feature set to generate an electromagnetic interference suppression strategy for the intelligent sensing device, including: The interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source correlation characteristics are fused to generate a comprehensive interference feature vector for the intelligent sensing device. The interference pattern matching layer in the interference suppression model is invoked to perform pattern recognition processing on the comprehensive interference feature vector to determine the target interference mode of the electromagnetic interference currently experienced by the intelligent sensing device. The target interference mode includes at least one of frequency band conflict mode, signal overload mode and external coupling mode. An initial suppression strategy is obtained by matching the target interference mode from a preset suppression strategy library, and the parameters of the initial suppression strategy are optimized based on the interference frequency band distribution characteristics, interference intensity fluctuation characteristics and interference source association characteristics to generate the electromagnetic interference suppression strategy.

6. The method according to claim 5, characterized in that, An initial suppression strategy is obtained by matching the target interference pattern from a preset suppression strategy library, including: When the target interference mode is a frequency band conflict mode, a frequency band adjustment strategy corresponding to the distribution characteristics of the interference frequency band is obtained from the suppression strategy library. The frequency band adjustment strategy includes a set of priority avoidance frequency bands and a set of backup working frequency bands. When the target interference mode is a signal overload mode, a gain compensation strategy corresponding to the interference intensity fluctuation characteristics is obtained from the suppression strategy library. The gain compensation strategy includes a gain adjustment curve and a maximum gain limit threshold. When the target interference mode is an external coupling mode, a filtering and suppression strategy corresponding to the interference source associated features is obtained from the suppression strategy library. The filtering and suppression strategy includes adaptive filter type selection and filter parameter configuration rules.

7. The method according to claim 6, characterized in that, Based on the interference frequency band distribution characteristics, interference intensity fluctuation characteristics, and interference source correlation characteristics, the initial suppression strategy is optimized to generate the electromagnetic interference suppression strategy, including: When the initial suppression strategy is a frequency band adjustment strategy, the priority avoidance frequency band set is sorted according to the energy proportion characteristics in the interference frequency band distribution characteristics, and the backup working frequency band set is expanded based on the available frequency band resources of the intelligent sensing device. When the initial suppression strategy is a gain compensation strategy, the slope of the gain adjustment curve and the compensation delay time are adjusted according to the fluctuation amplitude and fluctuation frequency in the interference intensity fluctuation characteristics, and the maximum gain limit threshold is updated based on the maximum carrying capacity of the smart sensing device. When the initial suppression strategy is a filtering suppression strategy, the target filter type is selected according to the device type and operating frequency range in the interference source association characteristics, and the cutoff frequency and stopband attenuation parameters of the target filter are configured based on the electromagnetic coupling strength and the correlation coefficient of historical electromagnetic signals.

8. The method according to claim 1, characterized in that, According to the electromagnetic interference suppression strategy, the signal acquisition module of the intelligent sensing device performs interference suppression operation, including: When the electromagnetic interference suppression strategy includes frequency band adjustment direction, the signal acquisition module is controlled to switch to the target operating frequency band in the set of backup operating frequency bands, and the interference energy change of the target operating frequency band is monitored in real time to activate the frequency band switching decision. When the electromagnetic interference suppression strategy includes a signal gain compensation direction, the amplifier of the signal acquisition module is subjected to gain compensation control according to the gain adjustment curve, and the compensated signal is subjected to amplitude limiting processing based on the maximum gain limit threshold. When the electromagnetic interference suppression strategy includes an anti-interference filtering direction, the target filter is inserted into the preprocessing circuit of the signal acquisition module, and the electromagnetic signal is filtered in real time according to the cutoff frequency and stopband attenuation parameters. The real-time monitoring of interference energy changes in the target operating frequency band to activate frequency band switching decisions includes: Multiple monitoring sub-bands are set within the target operating frequency band, and the current interference energy value of each monitoring sub-band is periodically collected; When the current interference energy value of any monitoring sub-band exceeds the preset energy threshold, a band switching request signal is generated, and the candidate switching band with the lowest interference energy value is selected from the set of backup working bands. The final switching frequency band is determined based on the signal transmission quality assessment results and switching time prediction results of the candidate switching frequency bands, and the signal acquisition module is controlled to switch to the final switching frequency band to complete the interference suppression operation.

9. An electromagnetic interference suppression system for intelligent sensing devices, characterized in that, It includes a processor, a memory, and a bus connected to the processor; the processor and the memory communicate with each other via the bus. The processor is used to call the computer program in the memory to execute the electromagnetic interference suppression method for AI-based intelligent sensing devices according to any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, It stores a program that, when executed by a processor, implements the electromagnetic interference suppression method for intelligent sensing devices based on AI recognition as described in any one of claims 1-8.

Images