Electronic equipment intelligent control method and system based on Internet of Things

Abstract

The invention discloses an electronic equipment intelligent control method and system based on the Internet of Things, and relates to the technical field of Internet of Things data analysis, and the method comprises the steps: analyzing the contribution degree of each data category to a control target, and considering the contribution or association conditions of different data categories corresponding to different control targets; and a reliable basis is provided for subsequent data fusion and equipment intelligent control. Kalman filtering parameters are set according to the contribution degree of the data category to the control target, and the Kalman filtering parameters are jointly set according to the attribute of the data category and the contribution degree. A fuzzy set and a membership function are defined on the basis of the fusion features, the type of the membership function is selected according to the features of the fusion features, and the pertinence of fuzzy reasoning is improved. And fuzzy reasoning is performed according to the fuzzy control rule to generate the control instruction, so that the adaptability and accuracy of intelligent control of the equipment are improved, normal, automatic and safe operation of intelligent equipment is ensured, and the working efficiency of the equipment is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things data analysis, and particularly to an intelligent control method and system for electronic devices based on the Internet of Things. Background Art

[0002] Through technologies such as sensors, RFID, and network communication, the Internet of Things has achieved ubiquitous connection between things and between things and people, providing a basis for the intelligent control of electronic devices. This solution uses Internet of Things technology to connect electronic devices to the Internet to achieve remote monitoring, intelligent adjustment, and automatic control. By integrating multiple sensors and intelligent algorithms, the system can real-time sense environmental changes and automatically adjust the operating state of the device according to preset rules or user instructions, thereby improving the device usage efficiency and user experience. This solution not only promotes the intelligent upgrade of electronic devices but also provides strong support for building application scenarios such as smart life and smart factories.

[0003] In the prior art, there are many types of data involved in the intelligent control of electronic devices in the Internet of Things, and the correlation between data and control objectives is relatively complex, resulting in poor adaptability and accuracy of device intelligent control and unable to ensure the automation level of electronic device intelligent control.

[0004] Therefore, how to improve the adaptability and accuracy of device intelligent control is a technical problem to be solved at present. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem of poor adaptability and accuracy of device intelligent control in the prior art, and propose an intelligent control method for electronic devices based on the Internet of Things, which includes: Collect all data categories and control objectives of the target electronic device in the Internet of Things, and analyze the contribution degree of each data category to the control objective; Set Kalman filter parameters according to the contribution degree of the data category to the control objective, and perform data fusion on all data categories to obtain a fusion feature; Define a fuzzy set and a membership function based on the fusion feature, and establish a fuzzy control rule for the target electronic device; Real-time collect all data of the target electronic device in the Internet of Things, perform fuzzy reasoning according to the fuzzy control rule to generate a control instruction, and perform intelligent control on the target electronic device according to the control instruction.

[0006] In some embodiments of the present application, after collecting all data categories and control objectives of the target electronic device in the Internet of Things, the method further includes: Classify the control objectives of the target electronic device, determine the matching relationship between all data categories and each type of control objective for some data categories, and establish the category matching relationship between data and control objectives; Collect the effect parameters describing the effect situation of each type of control objective, and add the corresponding effect parameters on the basis of the category matching relationship between data and control objectives to form the category matching relationship between data-control objective-effect parameters.

[0007] In some embodiments of the present application, analyze the contribution degree of each data category to the control objective, including, On the basis of the category matching relationship between data-control objective-effect parameters, draw a scatter plot between each data category and each effect parameter category, connect the points on the scatter plot for smoothing, determine the shape of the line on the scatter plot, and thus determine the linear relationship and non-linear relationship; For the linear relationship, describe the contribution degree between each data category and each effect parameter category through the correlation coefficient; For the non-linear relationship, describe the contribution degree between each data category and each effect parameter category through the SHAP value; Integrate the contribution degrees according to the effect parameter categories to obtain the contribution degree of each data category; Among them, the contribution degree between each data category and each effect parameter category is the contribution degree of the data category to the effect parameter.

[0008] In some embodiments of the present application, set the Kalman filter parameters according to the contribution degree of the data category to the control objective, including, Calculate the measurement noise variance and standard deviation under each data category, and determine the process noise variance under each data category according to the standard deviation; Construct the measurement noise covariance matrix and the process noise covariance matrix according to the measurement noise variance and the process noise variance respectively; Allocate element values to each data category according to the contribution degree of the data category to the control objective, and construct the observation matrix; Determine the Kalman gain based on the measurement noise covariance matrix, the process noise covariance matrix and the observation matrix; Among them, the Kalman filter parameters include the preset initial error covariance matrix, the measurement noise covariance matrix, the process noise covariance matrix, the observation matrix and the Kalman gain.

[0009] In some embodiments of the present application, perform data fusion on all data categories to obtain the fusion features, including, Under the category matching relationship between data and control objectives, perform the prediction step, update step and fusion feature step of the corresponding data category according to the Kalman filter parameters to obtain the fusion features under each group of category matching relationships between data and control objectives.

[0010] In some embodiments of the present application, on the basis of the fusion features, a fuzzy set and a membership function are defined, including: Defining the fuzzy set of the fusion features according to the category of the control target; Constructing the distribution map of each fusion feature, connecting each point on the distribution map, performing smoothing processing to obtain the fusion feature curve, and performing a normality test on the fusion feature curve according to the normality test method to obtain the test statistic; Statistical skewness and kurtosis on the fusion feature curve, and obtaining a normality index based on the test statistic, skewness, and kurtosis; Selecting the type of membership function according to the normality index, and constructing the membership functions of different fuzzy sets accordingly. The types of membership functions include Gaussian membership functions and other membership functions. Among them, the other membership functions include triangular membership functions and trapezoidal membership functions.

[0011] In some embodiments of the present application, a fuzzy control rule for the target electronic device is established, including: Establishing the fuzzy control rules for each control target of the target electronic device according to multiple fuzzy sets and multiple membership functions.

[0012] In some embodiments of the present application, fuzzy inference is performed according to the fuzzy control rules to generate a control instruction, including: Drawing the change curve of each type of data based on all the data collected during a period of time according to the facts, analyzing the change curve to determine the confidence level of each type of data, and adding the confidence level to the fuzzy inference to assist in generating the control instruction.

[0013] Correspondingly, the present application also provides an intelligent control system for an electronic device based on the Internet of Things, including: A first module for collecting all data categories and control targets of the target electronic device of the Internet of Things, and analyzing the contribution degree of each data category to the control target; A second module for setting Kalman filter parameters according to the contribution degree of the data category to the control target, and performing data fusion on all data categories to obtain fusion features; A third module for defining a fuzzy set and a membership function on the basis of the fusion features, and establishing a fuzzy control rule for the target electronic device; A fourth module for real-time collecting all data of the target electronic device of the Internet of Things, performing fuzzy inference according to the fuzzy control rules to generate a control instruction, and performing intelligent control on the target electronic device according to the control instruction.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. Analyze the contribution degree of each data category to the control target, consider the contribution or association of different data categories corresponding to different control targets, and provide a reliable basis for subsequent data fusion and intelligent device control. Set Kalman filter parameters according to the contribution degree of data categories to the control target, or jointly set Kalman filter parameters according to the attributes and contribution degrees of data categories, reasonably reduce noise, and help improve the real-time performance, robustness and scalability of data analysis and processing.

[0015] 2. Define fuzzy sets and membership functions based on the fused features, select the type of membership function according to the characteristics of the fused features, and improve the pertinence of fuzzy inference. Conduct fuzzy inference according to the fuzzy control rules to generate control instructions, and consider the confidence level of data to assist in determining control instructions. Improve the adaptability and accuracy of intelligent device control, ensure the normal and automatic safe operation of device intelligence, and improve the working efficiency of the device. Description of the Drawings

[0016] Figure 1 It is a schematic flow chart of the intelligent control method for electronic devices based on the Internet of Things proposed by the present invention; Figure 2 It is a schematic structural diagram of the intelligent control system for electronic devices based on the Internet of Things proposed by the present invention. Detailed Embodiments

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0018] Refer to Figure 1 , the intelligent control method for electronic devices based on the Internet of Things includes the following steps. Step S101, collect all data categories and control targets of the target electronic device of the Internet of Things, and analyze the contribution degree of each data category to the control target.

[0019] In this embodiment, the target electronic device of the Internet of Things may be an electronic device related to the Internet of Things such as a smart home control device or an industrial automation device. Different target electronic devices may involve a large number of data categories, and these data can be collected by sensors. The control objectives may include the start-stop conditions of control instructions, alarm monitoring, device protection, and specific effect adjustment (energy consumption optimization, comfort adjustment), etc. The data includes environmental parameters, device operation parameters, user parameters, etc. Different control objectives correspond to different data categories. For example, under smart home, when the control objective is the start-stop of certain control instructions, environmental data trigger thresholds are required (such as temperature > 30°C, humidity compliant, automatically turn on the air conditioner), and user location information is used to assist in judging the usage scenario. When the control objective is energy consumption optimization, it is necessary to comprehensively judge by combining environmental data and device operation data. Therefore, different control objectives correspond to different data categories.

[0020] In some embodiments of the present application, after collecting all data categories and control objectives of the target electronic device of the Internet of Things, the method further includes, classifying the control objectives of the target electronic device, determining the matching relationship between all data categories and some data categories of each type of control objective, and establishing a category matching relationship between data and control objectives; collecting the effect parameters describing the effect situation of each type of control objective, and adding the corresponding effect parameters on the basis of the category matching relationship between data and control objectives to form a category matching relationship between data-control objective-effect parameters.

[0021] In this embodiment, the effect parameters describing the effect situation of each type of control objective include the accuracy of the start instruction, the accuracy of the warning, the change in energy consumption, etc. The effect parameters of the start instruction include the start accuracy rate, response delay, false trigger rate, etc. The effect parameters of energy efficiency optimization include the energy efficiency ratio improvement rate, unit energy consumption cost, etc. The effect parameters of device protection include the failure rate reduction rate, overheat protection success rate, maintenance cost, downtime, etc.

[0022] In some embodiments of the present application, analyzing the contribution degree of each data category to the control objective includes, Based on the category matching relationship between data-control objective-effect parameters, draw a scatter plot between each data category and each effect parameter category, connect the points on the scatter plot for smoothing processing, determine the shape of the line on the scatter plot, and thus determine the linear relationship and non-linear relationship; For the linear relationship, describe the contribution degree between each data category and each effect parameter category through the correlation coefficient; For the non-linear relationship, describe the contribution degree between each data category and each effect parameter category through the SHAP value; Integrate the contribution degrees according to the effect parameter category to obtain the contribution degree of each data category; Among them, the contribution degree between each data category and each effect parameter category is the contribution degree of the data category to the effect parameter.

[0023] In this embodiment, in order to analyze the contribution or correlation of each type of data to the control target, each control target is refined, all effect parameters that can describe the control target are collected, and the strength of the correlation between the data and the effect parameters is calculated to indirectly reflect the contribution of each type of data to the control target. According to the shape of the scatter plot between each data category and each effect parameter category, linear and non-linear relationships are identified. For linear relationships, the correlation relationship of the linear relationship is described by the correlation coefficient (Pearson correlation coefficient). For non-linear relationships, the correlation relationship of the non-linear relationship is described by the SHAP value. A suitable machine learning model (such as random forest, gradient boosting tree, neural network, etc.) is selected for training. Ensure that the model can capture the non-linear relationships in the data, and use the SHAP (SHapley Additive exPlanations) method to calculate the contribution degree of each data category to each effect parameter category. The SHAP value is based on the Shapley value in game theory and can fairly distribute the contribution of each feature to the prediction result. The specific calculation steps include: a. For each data sample, calculate the prediction difference when a certain data category is included and not included. b. Perform a weighted sum on all possible feature combinations to obtain the SHAP value of this data category to the prediction result. c. Repeat the above steps to calculate the SHAP value of each data category to each effect parameter category.

[0024] In this embodiment, the linear and non-linear relationships (multiple effect parameters) of each data category are respectively converted into contribution degrees, and after integration, the contribution degree of the data category (contribution degree of data categories) is determined. The calculation formula is as follows: ; Among them, is the contribution degree of the th data category, , are the conversion coefficients of the linear and non-linear relationships respectively, , are the numbers of the linear and non-linear relationships of the th data category, that is, the numbers of their respective effect parameters, , are the combination weights of the contribution degrees of the th linear relationship and the th non-linear relationship respectively, , are the The contribution degrees of the th linear relationship and the th nonlinear relationship under each data category, 、 are respectively and the maximum values in, 、 are respectively the first constant and the second constant of the th data category, 、 respectively represent and the correction of the maximum value in to the average value. The first constant and the second constant exist to balance the magnitude of the correction function.

[0025] Step S102, set the Kalman filter parameters according to the contribution degree of the data category to the control target, and perform data fusion on all data categories to obtain the fusion feature.

[0026] In this embodiment, the Kalman filter is based on the state space model, which consists of a state equation and an observation equation. The state equation describes the dynamic change of the system state, while the observation equation describes the relationship between the sensor measurement value and the system state. In multi-class data fusion, the state vector can include data from different sensors, such as environmental data (temperature, humidity, etc.) and device state data (energy consumption, operating mode, etc.). The Kalman filter is a recursive estimation algorithm that uses the previous estimate value and the current observation value to update the estimate of the system state. Through this recursive method, the Kalman filter can dynamically combine multi-class data to achieve real-time estimation of the system state. The Kalman filter processes the noise in the system model and sensor measurements through the process noise covariance matrix and the measurement noise covariance matrix. The process noise covariance matrix describes the uncertainty in the dynamic change of the system state, while the measurement noise covariance matrix describes the uncertainty in the sensor measurements.

[0027] It can be understood that the Kalman filter is a recursive algorithm that can process newly arrived observation data in real time and update the estimate of the system state. This makes it very useful in scenarios that require real-time decision-making and control. The Kalman filter can handle the noise and uncertainty in the system model and sensor measurements. By adjusting the weights of the noise covariance matrix, the Kalman filter can adaptively cope with different noise levels. The Kalman filter can be easily extended to handle the case of multi-class data. Just incorporate data of different categories into the state vector and the observation matrix and adjust the noise covariance matrix accordingly.

[0028] In some embodiments of the present application, setting the Kalman filter parameters according to the contribution degree of the data category to the control target includes, Calculate the measurement noise variance and standard deviation for each data category, and determine the process noise variance for each data category according to the standard deviation; Construct a measurement noise covariance matrix and a process noise covariance matrix respectively according to the measurement noise variance and the process noise variance; Allocate element values for each data category according to the contribution degree of the data category to the control target, and construct an observation matrix; Determine the Kalman gain based on the measurement noise covariance matrix, the process noise covariance matrix and the observation matrix; Among them, the Kalman filter parameters include a preset initial error covariance matrix, a measurement noise covariance matrix, a process noise covariance matrix, an observation matrix and a Kalman gain.

[0029] In this embodiment, the process noise covariance matrix (Q): Describes the noise caused by external factors (such as random disturbances) during the system state transition process. It reflects the uncertainty of the system model and is an important basis for the covariance matrix update in the prediction step of the Kalman filter.

[0030] The measurement noise covariance matrix (R): Describes the noise caused by factors such as sensor errors during the observation process. It reflects the uncertainty of the observed value and is a key parameter for calculating the Kalman gain in the update step of the Kalman filter. Observation matrix (H): Describes the relationship between the observed value and the state variable. It maps the system state to the observation space, enabling the Kalman filter to use the observed value to update the state estimate. The noise variance (measurement noise variance) can be calculated by analyzing the historical observation data of each data category, and the diagonal elements of the R matrix can be set accordingly. The stability of each data category can be evaluated by calculating indicators such as the time series variance, change rate, and standard deviation of the data. A process noise variance is assigned to each data category according to the stability. These process noise variances are combined into a Q matrix, which has a similar form to the R matrix and is usually also a diagonal matrix. The elements of the H matrix can be set according to the contribution degree of the data category to the control target. Data categories with higher contribution degrees can be given larger element values of the observation matrix, so as to give higher weights in the Kalman filter.

[0031] In this embodiment, the Kalman gain is a key parameter, which determines the weight of the observed value for updating the state estimate. The calculation of the Kalman gain depends on the measurement noise covariance matrix (R), the process noise covariance matrix (Q) and the observation matrix (H). Kalman gain (K): Calculated in the update step and used to balance the weights between the predicted value and the observed value. It is one of the core parameters of the Kalman filter and determines the update degree of the state estimate.

[0032] In some embodiments of the present application, all data categories are fused to obtain fusion features, including Under the category matching relationship of data-control objectives, perform the prediction step, update step, and fusion feature step for the corresponding data categories according to the Kalman filter parameters to obtain the fusion features under the category matching relationship of each group of data-control objectives.

[0033] In this embodiment, the specific steps for implementing multi-class data fusion using the Kalman filter are given: Initialization: Set the initial values of the state vector and the covariance matrix (initial error covariance matrix). These initial values are usually determined based on prior knowledge or historical data.

[0034] Prediction step: Use the system model (state equation) to predict the state vector and covariance matrix at the next moment. This step does not consider the observation value at the current moment and only makes predictions based on the estimated value at the previous moment and the system model.

[0035] Update step: Combine the observation value at the current moment and the observation matrix to calculate the Kalman gain. The Kalman gain is used to balance the weights between the predicted value and the observation value to obtain a more accurate estimated value. Use the Kalman gain to update the state vector and covariance matrix to obtain the optimal estimated value at the current moment.

[0036] Output fusion features: Extract the key features after fusion, such as the comprehensive energy efficiency index, user comfort index, etc. These features can be calculated based on the fused state vector and are used for subsequent decision-making and control.

[0037] It should be noted that in the prediction step, the update of the prediction error covariance matrix (P) takes into account the process noise covariance matrix (Q), that is, the noise caused by external factors (such as random perturbations) during the system state transition. In the update step, combine the observation value at the current moment and the observation matrix, calculate the Kalman gain, and use the Kalman gain to update the state vector and covariance matrix. The updated covariance matrix reflects the uncertainty of the state estimate value after fusing the observation values. This update process also takes into account the measurement noise covariance matrix (R), that is, the noise caused by factors such as sensor errors during the observation process.

[0038] Step S103, define a fuzzy set and a membership function based on the fusion features, and establish the fuzzy control rules for the target electronic device.

[0039] In this embodiment, the control objectives of control instructions can be divided into fuzzy sets such as performance, energy conservation, and comfort. The control instructions for parameter optimization can be divided into fuzzy sets such as environmental comfort level (high, medium, low), equipment load (high, medium, low), and user activity level (high, medium, low). Fuzzy sets are an extension of classical sets, allowing elements to belong to a set to a certain degree (membership degree), rather than simply "belonging" or "not belonging". Fuzzy sets are the basis of fuzzy reasoning. In fuzzy reasoning, rule matching and reasoning are performed based on fuzzy sets and membership functions to obtain fuzzy conclusions, which are more suitable for the intelligent control of electronic devices in the Internet of Things environment. The membership function is the core of a fuzzy set, which defines the degree to which an element belongs to the fuzzy set. The value range of the membership function is usually [0, 1], where 0 means the element does not belong to the set at all, 1 means the element belongs to the set completely, and values between 0 and 1 indicate the degree to which the element belongs to the set.

[0040] In some embodiments of the present application, based on the fused features, fuzzy sets and membership functions are defined, including defining the fuzzy sets of the fused features according to the categories of the control objectives; constructing a distribution map for each fused feature, connecting the points on the distribution map, and performing smoothing processing to obtain a fused feature curve, and performing a normality test on the fused feature curve according to the normality test method to obtain a test statistic; statistical skewness and kurtosis on the fused feature curve, and obtaining a normality index based on the test statistic, skewness, and kurtosis; selecting the type of membership function according to the normality index, and thus constructing the membership functions of different fuzzy sets. The types of membership functions include Gaussian membership functions and other membership functions, where the other membership functions include triangular membership functions and trapezoidal membership functions.

[0041] In this embodiment, multiple data are fused to obtain a parameter of a fused feature. The distribution map is analyzed to determine whether it conforms to a normal distribution. Statistical test methods such as the Shapiro-Wilk test, Kolmogorov-Smirnov test (K-S test), or Anderson-Darling test are used to formally test whether the data follows a normal distribution. These test methods will give a p-value (test statistic). If the p-value is greater than a certain significance level (such as 0.05), the hypothesis that the data follows a normal distribution cannot be rejected. Skewness measures the symmetry of the data distribution, and the skewness of a normal distribution is 0. Kurtosis measures the peakedness of the data distribution, and the kurtosis of a normal distribution is 3 (0 in some definitions, depending on whether 3 is subtracted for standardization). Combining the test statistic, skewness, and kurtosis generates a normality index to determine whether it is a normal distribution.

[0042] In this embodiment, according to the distribution of the fusion features, a suitable type of membership function is selected. For example, if the data presents a normal distribution, a Gaussian membership function can be selected. Otherwise, a membership function of triangle or trapezoid is selected.

[0043] In some embodiments of the present application, fuzzy control rules for the target electronic device are established, including Fuzzy control rules for each control objective of the target electronic device are established according to multiple fuzzy sets and multiple membership functions.

[0044] In this embodiment, for example, for a smart home, the fuzzy sets may include environmental comfort level (low, medium, high), user activity level (low, medium, high), operation mode (comfortable, energy-saving), and wind speed. The fuzzy rule is: IF the environmental comfort level is "high" AND the user activity level is "low" THEN the operation mode is "energy-saving". Implementation: When the membership function value of the environmental comfort level is within the range of "high", and the membership function value of the user activity level is within the range of "low", the membership function value of the output operation mode is within the range of "energy-saving".

[0045] Step S104, all data of the target electronic device of the Internet of Things are collected in real time, fuzzy inference is performed according to the fuzzy control rules to generate a control instruction, and the target electronic device is intelligently controlled according to the control instruction.

[0046] In this embodiment, the fuzzy inference process is as follows: Fuzzification: Map the fusion features to the fuzzy sets.

[0047] Rule matching: Activate the fuzzy rules that meet the conditions.

[0048] Inference synthesis: Calculate the activation strength of each rule.

[0049] Defuzzification: Obtain an accurate control instruction through the centroid method or the maximum membership degree method.

[0050] Execution of control instruction Send the instruction to the device actuator (such as adjusting the frequency of the air conditioner compressor and changing the fan speed).

[0051] Monitor the execution effect and feedback it to the system log for subsequent optimization.

[0052] In some embodiments of the present application, fuzzy inference is performed according to the fuzzy control rules to generate a control instruction, including Draw the change curve of each type of data based on all the data collected over a period of time according to the facts, analyze the change curve to determine the confidence level of each type of data, and add the confidence level to the fuzzy inference to assist in generating the control instruction.

[0053] In this embodiment, to ensure the reliability of fuzzy inference, the credibility of real-time collected data is analyzed to determine the confidence level, and a plotting tool is selected: plotting libraries such as Matplotlib and Seaborn, or spreadsheet software such as Excel are used. Plotting curves: For each type of data, with its time series as the horizontal axis and the data value as the vertical axis, a change curve is plotted.

[0054] Confidence level evaluation: Stability: Data with small fluctuations and a stable trend in the curve has a higher confidence level; data with large fluctuations and an unstable trend has a lower confidence level.

[0055] Consistency: Data that is consistent with other relevant data or the expected trend has a higher confidence level; data with obvious deviations has a lower confidence level.

[0056] Data integrity: Data with less missing data and a complete coverage time period has a higher confidence level; data with more missing data and an incomplete coverage time period has a lower confidence level.

[0057] The calculation formula for the confidence level ("confidence level") of each type of data is as follows: ; Where is the confidence level of the th type of data, , , are the conversion coefficients of the stability, consistency, and integrity of the data respectively, , , are the confidence levels determined by the stability, consistency, and integrity of the th type of data respectively, , are the maximum and minimum values in respectively, is the third constant of the th type of data, represents the correction of the average value determined by the maximum and minimum values to the average value of the sum of the three.

[0058] Confidence level fusion: During the fuzzy inference process, the confidence level of each type of data is used as an additional input variable or weight factor to affect the result of fuzzy inference.

[0059] Processing of fuzzy inference results: According to the output result of the fuzzy inference system, defuzzification processing is performed to obtain specific control instruction parameters.

[0060] Instruction generation and sending: The control instruction parameters are converted into actual control instructions and sent to the actuator (such as intelligent devices, automation systems, etc.).

[0061] Correspondingly, the present application further provides an intelligent control system for electronic devices based on the Internet of Things, such as Figure 2 shown, including A first module for collecting all data categories and control objectives of the target electronic device of the Internet of Things, and analyzing the contribution degree of each data category to the control objective; A second module for setting Kalman filter parameters according to the contribution degree of the data category to the control objective, and performing data fusion on all data categories to obtain a fusion feature; A third module for defining a fuzzy set and a membership function on the basis of the fusion feature, and establishing a fuzzy control rule for the target electronic device; A fourth module for real-time collecting all data of the target electronic device of the Internet of Things, performing fuzzy reasoning according to the fuzzy control rule, generating a control instruction, and performing intelligent control on the target electronic device according to the control instruction.

[0062] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Analyze the contribution degree of each data category to the control objective, consider the contribution or association of different data categories corresponding to different control objectives, and provide a reliable basis for subsequent data fusion and device intelligent control. Set Kalman filter parameters according to the contribution degree of the data category to the control objective, and jointly set Kalman filter parameters according to the attributes and contribution degrees of the data categories, reasonably reduce noise, and help the real-time performance, robustness and scalability of data analysis and processing.

[0063] 2. Define a fuzzy set and a membership function on the basis of the fusion feature, select the type of membership function according to the characteristics of the fusion feature, and improve the pertinence of fuzzy reasoning. Perform fuzzy reasoning according to the fuzzy control rule to generate a control instruction, and consider the confidence of the data to assist in determining the control instruction. Improve the adaptability and accuracy of device intelligent control, ensure the normal automatic and safe operation of device intelligence, and improve the working efficiency of the device.

[0064] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present invention.

[0065] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the drawings are not necessarily essential for implementing the present invention.

[0066] Those skilled in the art can understand that the modules in the system of the implementation scenario can be distributed in the system of the implementation scenario according to the description of the implementation scenario, or can be correspondingly changed and located in one or more systems different from this implementation scenario. The modules of the above implementation scenario can be combined into one module, or can be further split into multiple sub-modules.

[0067] As described above, the above are only the preferred specific implementation manners of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered within the protection scope of the present invention.

Claims

1. An intelligent control method for electronic devices based on the Internet of Things, characterized in that, Including, Collect all data categories and control objectives of the target electronic device in the Internet of Things, and analyze the contribution degree of each data category to the control objective; Set Kalman filter parameters according to the contribution degree of the data category to the control objective, and perform data fusion on all data categories to obtain a fusion feature; Define a fuzzy set and a membership function based on the fusion feature, and establish a fuzzy control rule for the target electronic device; Collect all data of the target electronic device in the Internet of Things in real time, perform fuzzy inference according to the fuzzy control rule, generate a control instruction, and perform intelligent control on the target electronic device according to the control instruction.

2. The intelligent control method of an electronic device based on the Internet of Things according to claim 1, wherein After collecting all data categories and control objectives of the target electronic device in the Internet of Things, the method further includes, Classify the control objectives of the target electronic device, determine the matching relationship between all data categories and some data categories of each type of control objective, and establish a category matching relationship between data and control objectives; Collect the effect parameters describing the effect situation of each type of control objective, and add the corresponding effect parameters on the basis of the category matching relationship between data and control objectives to form a category matching relationship between data-control objective-effect parameters.

3. The intelligent control method for an electronic device based on the Internet of Things according to claim 2, wherein Analyze the contribution degree of each data category to the control objective, including, On the basis of the category matching relationship between data-control objective-effect parameters, draw a scatter plot between each data category and each effect parameter category, connect the points on the scatter plot for smoothing processing, determine the shape of the line on the scatter plot, and thus determine the linear relationship and non-linear relationship; For the linear relationship, describe the contribution degree between each data category and each effect parameter category through the correlation coefficient; For the non-linear relationship, describe the contribution degree between each data category and each effect parameter category through the SHAP value; Integrate the contribution degrees according to the effect parameter category to obtain the contribution degree of each data category; Among them, the contribution degree between each data category and each effect parameter category is the contribution degree of the data category to the effect parameter.

4. The intelligent control method of the electronic device based on the Internet of Things according to claim 3, wherein, Set Kalman filter parameters according to the contribution degree of the data category to the control objective, Including, Calculate the measurement noise variance and standard deviation under each data category, and determine the process noise variance under each data category according to the standard deviation; Construct a measurement noise covariance matrix and a process noise covariance matrix according to the measurement noise variance and the process noise variance respectively; Allocate element values for each data category according to the contribution degree of the data category to the control objective, and construct an observation matrix; Determine the Kalman gain based on the measurement noise covariance matrix, the process noise covariance matrix and the observation matrix; Among them, the Kalman filter parameters include a preset initial error covariance matrix, a measurement noise covariance matrix, a process noise covariance matrix, an observation matrix and a Kalman gain.

5. The intelligent control method of an electronic device based on the Internet of Things according to claim 4, characterized in that, And perform data fusion on all data categories to obtain a fusion feature, including, Under the category matching relationship between data and control objectives, perform the prediction step, update step and fusion feature step of the corresponding data category according to the Kalman filter parameters to obtain the fusion feature under the category matching relationship of each group of data-control objectives.

6. The intelligent control method for an electronic device based on the Internet of Things according to claim 1, characterized in that Define a fuzzy set and a membership function based on the fusion feature, including, Define the fuzzy sets of the fusion features according to the categories of the control objectives; Construct the distribution diagrams of each fusion feature, connect the points on the distribution diagrams, perform smoothing processing to obtain the fusion feature curves, and conduct a normality test on the fusion feature curves according to the normality test method to obtain the test statistic; Statistically analyze the skewness and kurtosis on the fusion feature curves, and obtain the normality index based on the test statistic, skewness, and kurtosis; Select the type of membership function according to the normality index, and construct the membership functions of different fuzzy sets accordingly. The types of membership functions include Gaussian membership functions and other membership functions, where the other membership functions include triangular membership functions and trapezoidal membership functions.

7. The intelligent control method for an electronic device based on the Internet of Things according to claim 1, characterized in that, Establish the fuzzy control rules of the target electronic device, including, Establish the fuzzy control rules for each control objective of the target electronic device according to multiple fuzzy sets and multiple membership functions.

8. The intelligent control method of an electronic device based on the Internet of Things according to claim 1, characterized in that, Conduct fuzzy inference according to the fuzzy control rules to generate control instructions, including, Draw the change curves of each type of data based on the data collected in a period of time according to the facts, analyze the change curves to determine the confidence level of each type of data, and add the confidence level to the fuzzy inference to assist in generating control instructions.

9. An intelligent control system for electronic devices based on the Internet of Things, characterized in that, including, The first module is used to collect all data categories and control objectives of the target electronic device in the Internet of Things, and analyze the contribution degree of each data category to the control objective; The second module is used to set the Kalman filter parameters according to the contribution degree of the data category to the control objective, and perform data fusion on all data categories to obtain the fusion features; The third module is used to define fuzzy sets and membership functions based on the fusion features, and establish the fuzzy control rules of the target electronic device; The fourth module is used to collect all data of the target electronic device in the Internet of Things in real time, conduct fuzzy inference according to the fuzzy control rules to generate control instructions, and perform intelligent control on the target electronic device according to the control instructions.

Images