Self-healing method of sensor system, storage medium and execution equipment

Abstract

The invention discloses a self-healing sensing method and a self-healing sensing device, and aims to improve the reliability and the accuracy of a sensor under the condition that a fault exists or the performance is reduced. According to the basic principle of the method, a self-detection and self-healing mechanism is started in time by monitoring the difference between sensor output and target physical quantity in real time, so that the stability of the system and the reliability of data are ensured. According to the self-healing sensing method and device, data unreliability caused by sensor faults can be effectively handled, and the self-adaptive capacity and robustness of the system are improved. Through the combination of real-time monitoring and machine learning, the system can automatically adjust the output when the sensor fails, and reduces the need of human intervention, thereby improving the stability and data quality of the whole system. The innovative method has wide application prospects in the fields of intelligent sensors, automatic control, Internet of Things and the like.

Description

Technical Field

[0001] The present invention relates to the field of sensor technology, and in particular to a self-healing method, storage medium and execution device of a sensor system. Background Art

[0002] In modern industrial and automation systems, sensors, as core components of data acquisition, undertake the important task of monitoring and feedback of key physical quantities (such as temperature, pressure, humidity, etc.). With the rapid development of the Internet of Things (IoT), smart manufacturing and automation technology, the application of sensors is becoming more and more extensive. However, sensors face many challenges in actual operation, especially their reliability and accuracy.

[0003] Current sensor fault detection solutions mainly include redundant sensor networks, adaptive signal processing technology, and model-based fault detection. Redundant sensor networks deploy multiple sensors of the same type to ensure functional succession, thereby monitoring data in real time and isolating faults, but this increases costs and system complexity, and cannot repair failed outputs. Adaptive signal processing uses filtering algorithms to identify and filter noise in real time to improve data quality, but lacks self-healing capabilities in severe fault conditions. Model-based fault detection improves accuracy and provides early warnings by comparing sensor outputs with model predictions, but the model is complex to build and is prone to drift in dynamic environments and cannot be corrected in real time. Although these solutions have their own advantages and disadvantages, none of them can effectively achieve automatic repair and self-healing of data.

[0004] Although the existing technology has made some progress in sensor monitoring and fault detection, it still faces many challenges, especially in real-time self-healing and intelligent processing capabilities. The present invention provides a more advanced solution through a unique self-healing mechanism and machine learning algorithm, which can effectively improve the working stability and data reliability of sensors in complex environments. Summary of the invention

[0005] 1. Technical issues to be resolved To solve the above problems, the present invention proposes a self-healing method, storage medium and execution device for a sensor system, aiming to solve the problem that the sensor lacks self-healing ability in the event of a serious fault.

[0006] (II) Technical solution A self-healing method for a sensor system of the present invention comprises: S100, obtaining a standard value of a target physical quantity, obtaining output values ​​of at least two sensors of the same type, comparing the output values ​​with the standard values ​​to obtain an error value, and outputting an abnormal signal based on the error value; S200, based on the abnormal signal, calibrating the label of the sensor that outputs the error value, correcting the error value using a corrector according to the information of the label and outputting a corrected value; S300, comparing the correction value with the standard value to obtain a correction result, and determining whether to output the correction value based on the correction result.

[0007] In the present invention, an evaluation cycle and a standard accuracy are preset, and whether the corrector reaches the standard accuracy is determined based on the evaluation cycle; If the standard accuracy is reached, continue to use the corrector to correct the error value; If the standard accuracy is not met, the corrector is optimized based on a loss function.

[0008] In the present invention, the corrector is constructed by a regression model or a neural network, and the regression model or the neural network is set as the corrector , input is the error value , the target output is the accurate physical quantity , ,in is the predicted value generated by the corrector.

[0009] In the present invention, the training process requires labeled data and real-time acquisition of the error value As the target output and target input, the training model is: ,in is the sample size, is the prediction generated by the corrector, is the actual physical value, is the orthotic parameter; Update the parameters after each training : , is the learning rate, is the gradient of the loss function with respect to the parameters.

[0010] In the present invention, the standard accuracy is set to , when the error between the correction result and the target output is less than When , the trainer is considered to meet the standard, where the training error is ,like If the training is completed, the training is considered complete, otherwise continue training; When the corrector outputs ,in is the final parameter obtained by the corrector, and the output value of the corrector is the correction value.

[0011] In the present invention, the standard value is a preset value and / or output values ​​of other sensors except the sensor that outputs an erroneous value and / or an average value of the sensor output values; When the output value of the sensor is greater than or less than the standard value, the output value is determined to be in an abnormal state; The labels include normal and abnormal states.

[0012] In the present invention, the step S300 further includes: The correction results include reaching the standard and not reaching the standard. If the correction value If the target is met, it is determined that the target is met; otherwise, it is determined that the target is not met and step S200 is continued, wherein: is the preset offset accuracy, is the standard value.

[0013] Another computer-readable storage medium of the present invention stores a computer program, which, when executed by a processor, implements the steps of the self-healing method of the sensor system described in the above technical solution.

[0014] Another execution device of the present invention includes an execution device body and a controller, wherein the controller includes a processor and a computer program stored in a memory and executable on the processor, and when the processor executes the program, the steps of the self-healing method of the sensor system described in the above technical solution are implemented.

[0015] (III) Beneficial effects Compared with the prior art, the present invention has the following beneficial effects: Through real-time self-detection and self-healing correction mechanisms, the present invention can automatically repair the output when the sensor fails or the performance is degraded, ensuring the continuous accuracy of the system output, and can still provide reliable operating data even when the sensor fails.

[0016] The present invention adopts a machine learning-driven corrector, which can perform self-training and optimization according to historical data and real-time data, continuously improve the correction accuracy, adapt to different environmental changes, and ensure that the system always maintains efficient operation.

[0017] The present invention corrects the sensor output in real time, eliminates data errors caused by sensor failures, improves data accuracy and reliability, ensures the quality of system decisions, and reduces system failures or misjudgments caused by erroneous data. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A schematic diagram of the logical structure of the self-healing method; Figure 2 This is a schematic diagram of the logical framework of the self-healing method; Figure 3 It is a schematic diagram of the module structure of the sensor system; Figure 4 It is a schematic diagram of the framework structure of the execution equipment.

[0020] 10-acquisition unit; 11-sensor; 12-acquisition module; 20-processing unit; 30-correction unit; 31-historical data storage module; 32-corrector; 40 - processor; 41 - memory; 42 - communication interface; 43 - communication bus. DETAILED DESCRIPTION

[0021] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Embodiment 1:

[0022] like Figure 1-Figure 2 A self-healing method for a sensor system is shown, which aims to improve the reliability and accuracy of the sensor in the event of a fault or performance degradation. The basic principle of this method is to monitor the difference between the sensor output and the target physical quantity in real time, and start the self-detection and self-healing mechanism in time to ensure the stability of the system and the reliability of the data.

[0023] The method flow of the present invention realizes the automation of real-time monitoring, fault detection and self-healing mechanism of sensors. The system continuously monitors the sensor output and determines the difference between it and the target physical quantity by setting a threshold. If the difference is significant, the self-detection program is automatically started to obtain accurate measurement values ​​and perform fault assessment. If the fault is confirmed, the system will generate a machine learning-driven corrector and train it using historical and real-time data until the predetermined accuracy is achieved. Finally, the trained corrector will post-process the sensor output and provide accurate data, thereby improving the stability and data quality of the overall system and reducing dependence on human intervention. The method comprises the following steps: S100, obtaining a standard value of a target physical quantity, obtaining output values ​​of the sensors based on at least two sensors of the same type, comparing the output values ​​with the standard values ​​to obtain error values, and outputting an abnormal signal based on the error values.

[0024] The standard value may be one or more combinations of a preset value, output values ​​of sensors other than the sensor that outputs an erroneous value, and an average value of sensor output values.

[0025] In particular, when the standard value is a preset value, the preset value can be set according to the provisions of product technical manuals, technical specifications, general product physical standards, physical parameter standards for specific industries or products, enterprise standard documents, international standard documents, and other documents or specific physical parameter standards.

[0026] When the standard value is the output value of other sensors except the sensor that outputs the wrong value, the output parameters of the cluster system composed of sensors of the same type are collected as the standard value, that is, sensors of the same type and with the same functions are deployed except the sensor that outputs the wrong value. This method realizes redundancy by deploying multiple sensors of the same type in the system to ensure that when a sensor fails, other sensors can take over its function. By comparing the output values ​​of each sensor, abnormal situations can be detected and faults can be isolated.

[0027] When the standard value is the average value of the sensor output, the sensor system needs to have a certain amount of data collection or a certain number of sensors as a basis to ensure the accuracy of the standard value.

[0028] Furthermore, the standard value can also be selected from a combination of three data values: a preset value, the output values ​​of sensors other than the sensor that outputs an erroneous value, and the average value of the sensor output value. That is, the detection mechanism is used to determine which of the three data types is more reasonable and use this data as the standard value.

[0029] When the sensor detects a significant difference between its output value and the pre-calibrated physical quantity, the system will automatically start the self-detection procedure. This process includes real-time monitoring of the sensor's output signal and using the set threshold value, i.e. the standard value, to determine whether the output signal is abnormal and the sensor enters the self-healing state.

[0030] S200, based on the abnormal signal, calibrate the label of the sensor that outputs the error value, use a corrector to correct the error value according to the information of the label and output a correction value.

[0031] After determining that the sensor output value is abnormal in step S100, the abnormality of the sensor is determined, that is, the label of the sensor that outputs the error value is calibrated. In particular, in this embodiment, the label includes normal and abnormal states. That is, when the error value is greater than or less than the standard value, it is determined to be an abnormal state, that is, the next self-healing repair step needs to be performed.

[0032] Furthermore, the tag may also include an abnormal state to be inspected, and its judgment standard depends on the degree of deviation between the error value and the standard value, that is, a judgment threshold is set by the user's own preset or calculated by a reasonable mathematical formula, and the judgment threshold is an interval value that includes the standard value. When the error value is between the judgment threshold and the standard value, it is judged to be an abnormal state to be inspected. At this time, the data output by the sensor can still be used, but the operation results are detected in real time through the alarm system. If the result value is still abnormal, the next step of disposal is carried out. When the error value is greater than or less than the judgment threshold, it is judged to be an abnormal state, that is, the next step of self-healing repair is required, which includes generating a machine learning-driven corrector to correct the output of the sensor in real time.

[0033] The self-healing mechanism generates a corrector through machine learning techniques. This process involves selecting an appropriate model (such as a regression model or a neural network) whose input is the sensor's erroneous output and the target output is the accurate physical quantity.

[0034] Set regression model or neural network as corrector , input is the error value , the target output is the accurate physical quantity , ,in is the predicted value generated by the corrector.

[0035] The training process requires labeled data and real-time acquisition of the error value As the target output and target input, the training model is: ,in is the sample size, is the prediction generated by the corrector, is the actual physical value, is the orthotic parameter; Update the parameters after each training : , is the learning rate, is the gradient of the loss function with respect to the parameters.

[0036] Corrector output value ,in is the final parameter obtained by the corrector, and the output value of the corrector is the correction value.

[0037] The above are all correction processes of error values ​​in the corrector. In order to ensure the correction accuracy of the corrector, the user presets an evaluation cycle and standard accuracy by himself or according to the system judgment, and judges whether the corrector reaches the standard accuracy based on the evaluation cycle. If the standard accuracy is reached, continue to use the corrector to correct the error value; if the standard accuracy is not met, optimize the corrector based on the loss function.

[0038] Set the standard accuracy to , when the error between the correction result and the target output is less than When , the trainer is considered to meet the standard, where the training error is ,like If the training is completed, the training is considered complete, otherwise continue training; when When the model training is completed and the accuracy meets the requirements, the training mode is turned off and the trained corrector is used for real-time data correction. This method can not only accurately correct sensor faults through the adaptive ability of machine learning, but also continuously optimize the ability of data correction as the environment changes, thereby greatly improving the reliability and intelligence level of the system.

[0039] The trained corrector starts to correct the sensor output in real time. At this time, the sensor output may still be faulty or abnormal, and the corrector's task is to adjust these erroneous data into more accurate physical quantities.

[0040] Error output for a given sensor , the trained corrector C will map the error output to the corrected physical value Therefore, the role of the corrector is to map the erroneous data to the correct data space: ,in, are the final model parameters obtained through training, is the sensor output value, This process ensures that even if a sensor fails, the system will still give an accurate output.

[0041] Specifically, the output of the corrector is used to monitor and control the process in real time, minimizing the impact of faulty sensors. The entire system is able to automatically correct data deviations without human intervention, ensuring continued stable operation.

[0042] After self-healing correction, the system will output the corrected data. The corrected data can be used for subsequent decision-making or control logic of the system to ensure the accuracy and reliability of the decision-making basis.

[0043] Corrected data The corrected data will continue to flow into the decision-making layer or control module for further use. The corrected data is not only used for real-time control and decision-making, but also for record storage or analysis to verify the stability of the system or perform subsequent model optimization.

[0044] By implementing self-healing output, the trained corrector can correct the sensor's erroneous output in real time to ensure the stability and reliability of the system. The corrected data is then provided to subsequent decision or control modules for accurate decision support or automated control. This process enables the system to continuously provide accurate output data without human intervention, thereby improving the intelligence level and adaptability of the entire system.

[0045] The process design of this method aims to realize the real-time monitoring and self-healing capability of the sensor. Through the collaborative work of various modules, it ensures a rapid response when a sensor fails, maintaining the reliability of the system and the accuracy of the data.

[0046] S300, comparing the correction value with the standard value to obtain a correction result, and determining whether to output the correction value based on the correction result.

[0047] If the correction value If the target is met, it is determined that the target is met; otherwise, it is determined that the target is not met and step S200 is continued, wherein: is the preset offset accuracy, Through real-time self-detection and self-healing correction mechanisms, the present invention can automatically repair the output when the sensor fails or performance degrades, ensuring the continuous accuracy of the system output, and can still provide reliable operating data even when the sensor fails.

[0048] The self-healing sensing method and device of the present invention can effectively deal with the data unreliability caused by sensor failure, and improve the system's adaptability and robustness. By combining real-time monitoring with machine learning, the system can automatically adjust the output when the sensor fails, reducing the need for human intervention, thereby improving the stability and data quality of the overall system. This innovative method has broad application prospects in the fields of intelligent sensors, automated control, and the Internet of Things. Embodiment 2:

[0049] like Figure 3 The difference between the embodiment 1 and the embodiment 1 is that the embodiment provides a sensor system with a self-healing function, which can monitor the sensor output value in real time, judge its accuracy, and make corrections through a correction unit when necessary, thereby ensuring the stability and accuracy of the sensor system. The specific implementation of the present invention will be described in detail below with reference to the accompanying drawings.

[0050] In this embodiment, the sensor system mainly includes a collection unit 10, a processing unit 20 and a correction unit 30. The collection unit 10 is equipped with a plurality of sensors 11 and a collection module 12. The sensor 11 is used to sense external physical quantities, such as temperature, pressure, humidity, etc., and convert them into electrical signals for output. The collection module 12 is responsible for obtaining the output values ​​of these sensors in real time and transmitting them to the processing unit 20. In this embodiment, the collection module 12 can be a data acquisition chip.

[0051] The processing unit 20 is connected to the acquisition unit 10 via a data bus or other communication methods to receive the sensor output values ​​from the acquisition module 12. The processing unit 20 has built-in judgment logic for comparing these output values ​​with the predetermined target physical quantity to determine whether there is a difference. If the difference exceeds a preset threshold, the processing unit 20 will determine that the self-healing mechanism needs to be activated and send an instruction to the correction unit 30.

[0052] The correction unit 30 is connected to the processing unit 20 and receives the instructions and sensor output values ​​from the processing unit 20. When it is determined that there is a difference between the sensor output value and the predetermined target physical quantity, the correction unit 30 will correct the sensor output value. To achieve this function, the correction unit 30 has a built-in machine learning algorithm and can improve the accuracy of its data mapping through machine learning training. The output value of the correction unit 30 is the actual output value of the sensor system, which is closer to the real physical quantity after correction.

[0053] Furthermore, the correction unit 30 uses the historical data stored in the historical data storage module 31 and the real-time data of the acquisition unit 10 for machine learning training. By constructing a regression model or other appropriate algorithms, the correction unit 30 can continuously optimize its parameters and improve the accuracy of the correction. The process of machine learning training is continuous until the predetermined accuracy standard is reached.

[0054] In this embodiment, the correction unit 30 also includes a corrector 32. The corrector 32 is generated by the processing unit 20 according to the difference between the sensor output value and the predetermined target physical quantity, and is used to perform a specific correction operation on the sensor output value. The generation and parameter adjustment of the corrector 32 are based on a machine learning algorithm to ensure the accuracy and stability of the correction. Embodiment 3:

[0055] An embodiment of the present invention provides a computer-readable storage medium.

[0056] The computer-readable storage medium provided in the embodiment of the present invention stores a computer program, and when the computer program is executed by a processor, the steps of any one of the above-mentioned self-healing methods for the sensor system can be implemented.

[0057] The computer-readable storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0058] For an introduction to the computer-readable storage medium provided in an embodiment of the present invention, please refer to the above method embodiment, and the present invention will not elaborate on it here.

[0059] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0060] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0061] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. Embodiment 4:

[0062] An embodiment of the present invention provides an execution device.

[0063] Please refer to Figure 4 , Figure 4 This is a schematic diagram of the structure of an execution device provided by the present invention, and the execution device may include: Memory for storing computer programs; The processor is used to implement the steps of any one of the above-mentioned self-healing methods for the sensor system when executing the computer program.

[0064] like Figure 3 As shown, it is a schematic diagram of the composition structure of the execution device, and the execution device may include: a processor 40, a memory 41, a communication interface 42 and a communication bus 43. The processor 40, the memory 41, and the communication interface 42 all communicate with each other through the communication bus 43.

[0065] In the embodiment of the present invention, the processor 40 may be a central processing unit (CPU), an application specific integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic devices.

[0066] The processor 40 may call a program stored in the memory 41 . Specifically, the processor 40 may execute the operations in the embodiment of the push button switch fault detection method.

[0067] The memory 41 is used to store one or more programs, which may include program codes, and the program codes include computer operation instructions. In the embodiment of the present invention, the memory 41 at least stores programs for implementing the following functions: Obtaining a standard value of a target physical quantity, based on at least two sensors of the same type, obtaining output values ​​of the sensors and comparing the output values ​​with the standard values ​​to obtain an error value, and outputting an abnormal signal based on the error value; Based on the abnormal signal, calibrate the label of the sensor that outputs the error value, correct the error value using a corrector according to the information of the label and output a corrected value; The correction value is compared with a standard value to obtain a correction result, and whether to output the correction value is determined based on the correction result.

[0068] In a possible implementation, the memory 41 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function, etc.; the data storage area may store data created during use.

[0069] In addition, the memory 41 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device or other volatile solid-state storage device.

[0070] The communication interface 42 may be an interface of a communication module, and is used to connect to other devices or systems.

[0071] Of course, it should be noted that Figure 3 The structure shown does not constitute a limitation on the execution device in the embodiment of the present invention. In actual applications, the execution device may include Figure 3 More or fewer components than shown, or combinations of certain components.

[0072] Finally, the evaluation results and evaluation content are sent to the end user and the cloud for data storage.

[0073] The above-described embodiments are merely descriptions of preferred implementations of the present invention, and are not intended to limit the concept and scope of the present invention. Without departing from the design concept of the present invention, various modifications and improvements made by ordinary persons in the art to the technical solution of the present invention should fall within the protection scope of the present invention, and the technical contents for which protection is sought in the present invention have been fully recorded in the claims.

Claims

1. A self-healing method for a sensor system, characterized in that: The method comprises: S100, obtaining a standard value of a target physical quantity, obtaining output values ​​of at least two sensors of the same type, comparing the output values ​​with the standard values ​​to obtain an error value, and outputting an abnormal signal based on the error value; S200, based on the abnormal signal, calibrating the label of the sensor that outputs the error value, correcting the error value using a corrector according to the information of the label and outputting a corrected value; S300, comparing the correction value with the standard value to obtain a correction result, and determining whether to output the correction value based on the correction result.

2. The self-healing method of the sensor system according to claim 1, characterized in that: Preset an evaluation cycle and standard accuracy, and determine whether the corrector reaches the standard accuracy based on the evaluation cycle; If the standard accuracy is reached, continue to use the corrector to correct the error value; If the standard accuracy is not met, the corrector is optimized based on a loss function.

3. The self-healing method of the sensor system according to claim 1 or 2, wherein the corrector is constructed by a regression model or a neural network, and the regression model or the neural network is set as the corrector. , input is the error value , the target output is the accurate physical quantity , ,in is the predicted value generated by the corrector.

4. The self-healing method of the sensor system according to claim 3, characterized in that: The training process requires labeled data and real-time acquisition of the error value As the target output and target input, the training model is: ,in is the sample size, is the prediction generated by the corrector, is the actual physical value, is the orthotic parameter; Update the parameters after each training : , is the learning rate, is the gradient of the loss function with respect to the parameters.

5. The self-healing method of the sensor system according to claim 4, characterized in that: Set the standard accuracy to , when the error between the correction result and the target output is less than When , the trainer is considered to meet the standard, where the training error is ,like If the training is completed, the training is considered complete, otherwise continue training; When the corrector outputs ,in is the final parameter obtained by the corrector, and the output value of the corrector is the correction value.

6. The self-healing method of the sensor system according to claim 5, characterized in that: The standard value is a preset value and / or the output values ​​of other sensors except the sensor that outputs an erroneous value and / or the average value of the sensor output value; When the output value of the sensor is greater than or less than the standard value, the output value is determined to be in an abnormal state; The labels include normal and abnormal states.

7. The self-healing method of the sensor system according to claim 6, characterized in that: The step S300 further includes: The correction results include reaching the standard and failing to reach the standard. If the target is met, it is determined that the target is met; otherwise, it is determined that the target is not met and step S200 is continued, wherein: is the preset offset accuracy, is the standard value.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the self-healing method of the sensor system described in any one of claims 1 to 7 are implemented.

9. An execution device, comprising an execution device body and a controller, characterized in that: The controller includes a processor and a computer program stored in a memory and executable on the processor. When the processor executes the program, the steps of the self-healing method for the sensor system according to any one of claims 1 to 7 are implemented.

Images