Sensor replacement status prediction method and related device

Abstract

The implementation method of the present application provides a sensor replacement status prediction method and related device. The method obtains the deviation data of the target sensor; the deviation data comes from the slow time-varying object of the target sensor; updates the measurement monitoring matrix corresponding to the target sensor according to the deviation data; obtains the input state value of the current data node of the target sensor in the input matrix; based on the input matrix and the output matrix, updates the output layer weight parameter in the RBF neural network to obtain the state prediction value of the current data node in the updated RBF neural network; based on the difference between the input state value and the state prediction value, determines whether the target sensor issues a replacement instruction to the user. The present application updates the measurement monitoring matrix through a slow time-varying object to realize automatic monitoring of sensors, which helps to timely detect sensor failures or failures, improve the maintenance and replacement efficiency of sensors, and improve the reliability and stability of smart city systems.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of data processing. More specifically, the embodiments of the present application relate to a method for predicting the replacement status of a sensor and a related device. Background Art

[0002] At present, the construction of smart cities is in full swing. Smart cities refer to urban management methods that use information and communication technology and the Internet to make various fields in the city intelligent, digital and interconnected. For example, cities such as Xiamen, Quanzhou, and Putian have set up multiple micro-environment systems for urban monitoring. Each micro-environment is equipped with dozens of different types of sensors such as particle concentration, temperature, humidity, and weight, which continuously collect data 24 hours a day. However, due to the changing environmental factors in the city, the sensors are prone to aging and wear after a period of use. The existing maintenance methods focus on regular manual maintenance, which imposes a heavy burden on manpower.

[0003] In summary, there is an urgent need to provide a sensor replacement status prediction solution to solve the above-mentioned technical problems existing in the related technologies. Summary of the invention

[0004] In this context, the embodiments of the present application hope to provide a method and related device for predicting the replacement status of a sensor, so as to realize automatic monitoring of the sensor by updating the measurement monitoring matrix of a slow time-varying object, thereby greatly improving the monitoring accuracy, facilitating the timely detection of sensor faults or failures, improving the maintenance and replacement efficiency of the sensor, and improving the reliability and stability of the smart city system.

[0005] In a first aspect of the embodiments of the present application, a method for predicting a sensor replacement state is provided, comprising:

[0006] Obtaining deviation data of the target sensor; wherein the deviation data includes deviation input data and / or deviation output data of the target sensor; the deviation data is used to indicate the change of the sensor during the aging process; the deviation data comes from a slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sampling rate of the sensor;

[0007] The measurement monitoring matrix corresponding to the target sensor is updated according to the deviation data; the measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and a radial basis function (RBF) information matrix;

[0008] Get the input state value of the target sensor at the current data node in the input matrix;

[0009] Based on the input matrix and the output matrix, the output layer weight parameters in the RBF neural network are updated to obtain the state prediction value of the current data node in the updated RBF neural network;

[0010] Based on the difference between the input state value and the state prediction value, it is determined whether the target sensor issues a replacement instruction to the user.

[0011] In a second aspect of the embodiments of the present application, a device for predicting a sensor replacement state is provided, the device comprising:

[0012] An acquisition module is used to acquire deviation data of a target sensor; wherein the deviation data includes deviation input data and / or deviation output data of the target sensor; the deviation data is used to indicate the change of the sensor during the aging process; the deviation data comes from a slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sampling rate of the sensor;

[0013] An update module is used to update the measurement monitoring matrix corresponding to the target sensor according to the deviation data; the measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and an RBF neural network;

[0014] The prediction module is used to obtain the input state value of the current data node of the target sensor in the input matrix; based on the input matrix and the output matrix, update the output layer weight parameters in the RBF neural network to obtain the state prediction value of the current data node in the updated RBF neural network; based on the difference between the input state value and the state prediction value, determine whether the target sensor issues a target sensor replacement instruction to the user.

[0015] In a third aspect of the embodiments of the present application, a computing device is provided, the computing device comprising:

[0016] at least one processor, memory, and input-output unit;

[0017] The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the sensor replacement status prediction method of the first aspect.

[0018] In a fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided, which includes instructions. When the instructions are executed on a computer, the computer executes the sensor replacement status prediction method of the first aspect.

[0019] In the embodiment of the present application, a method and a related device for predicting the replacement status of a sensor are provided. In the embodiment of the present application, the deviation data of the target sensor is first obtained. Wherein, the deviation data includes the deviation input data and / or the deviation output data of the target sensor; the deviation data is used to indicate the change of the sensor during the aging process; the deviation data comes from the slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity with a change speed lower than the sensor sampling rate. Then, the measurement monitoring matrix corresponding to the target sensor is updated according to the deviation data. Wherein, the measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and an RBF neural network. Then, the input state value of the current data node of the target sensor in the input matrix is ​​obtained. And, based on the input matrix and the output matrix, the output layer weight parameters in the RBF neural network are updated to obtain the state prediction value of the current data node in the updated RBF neural network. Finally, based on the difference between the input state value and the state prediction value, it is determined whether the target sensor issues a replacement instruction to the user. In the implementation mode of the present application, automatic monitoring of sensors is achieved by updating the measurement monitoring matrix through slow time-varying objects, which greatly improves the monitoring accuracy, helps to promptly detect sensor faults or failures, improves the maintenance and replacement efficiency of sensors, and improves the reliability and stability of the smart city system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become readily understood. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-limiting manner, wherein:

[0021] Figure 1 A schematic diagram of a flow chart of a method for predicting a sensor replacement status provided in an embodiment of the present application;

[0022] Figure 2 A schematic diagram of a sensor replacement status prediction example provided in an embodiment of the present application;

[0023] Figure 3 A schematic diagram of the structure of a sensor replacement status prediction device provided in one embodiment of the present application;

[0024] Figure 4 A schematic diagram of the structure of a medium in an embodiment of the present application is schematically shown;

[0025] Figure 5 A schematic diagram of the structure of a computing device according to an embodiment of the present application is shown schematically.

[0026] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts. DETAILED DESCRIPTION

[0027] The principles and spirit of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present application, and are not intended to limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present disclosure more thorough and complete, and to fully convey the scope of the present disclosure to those skilled in the art.

[0028] Those skilled in the art know that the embodiments of the present application can be implemented as a system, device, equipment, method or computer program product. Therefore, the present disclosure can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0029] At present, the construction of smart cities is in full swing. Smart cities refer to urban management methods that use information and communication technology and the Internet to make various fields in the city intelligent, digital and interconnected. Through technologies such as the Internet of Things, big data, and artificial intelligence, smart cities can achieve intelligent and efficient operations in urban traffic management, energy management, environmental monitoring, public safety, smart medical care, and smart buildings. The goal of smart cities is to provide a more convenient, comfortable, and sustainable living environment, and to improve urban management efficiency and the quality of life of citizens.

[0030] For example, cities such as Xiamen, Quanzhou, and Putian have set up multiple micro-environment systems for urban monitoring. Each micro-environment is equipped with dozens of different types of sensors such as particle concentration, temperature, humidity, and weight, which continuously collect data 24 hours a day. However, due to the changing environmental factors in the city, the sensors are prone to aging and wear after a period of use. The existing maintenance method focuses on regular manual inspections, which imposes a heavy burden on manpower.

[0031] The applicant found that in order to reduce the manpower burden, the relevant technology mainly predicts the sensor status based on degradation experimental data, but this method is often not very accurate. This is because the sensor characteristics will change over time, and the model established based on the existing sensor characteristics cannot track the slow drift caused by this change phenomenon, which in turn causes the accuracy of the established model to decrease, or even the model to fail. This problem has become a common problem in the field of smart city construction.

[0032] The applicant found that in order to solve the problem of reduced accuracy or even failure of the above-mentioned model, in the related technology, the sensor can be adaptively compensated for deviation, the problem of data offset can be solved, and the problem of replacing the sensor once the data offset occurs can be avoided, which effectively extends the online working time of the sensor. However, the aging of the sensor still exists, and the deviation will gradually increase and produce more unstable fluctuations with aging until it cannot be compensated. In fact, the physical life of the sensor cannot be extended by data compensation.

[0033] In summary, there is an urgent need to provide a sensor replacement status prediction solution to solve the above-mentioned technical problems existing in the related technologies.

[0034] In order to overcome the above technical problems existing in the existing solutions, according to the implementation of the present application, a sensor replacement status prediction method and related devices are proposed.

[0035] In the implementation mode of the present application, the deviation data of the target sensor is first obtained. The deviation data includes the deviation input data and / or deviation output data of the target sensor; the deviation data is used to indicate the change of the sensor during the aging process; the deviation data comes from the slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sensor sampling rate. Then, the measurement monitoring matrix corresponding to the target sensor is updated according to the deviation data. The measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and an RBF neural network. Next, the input state value of the current data node of the target sensor in the input matrix is ​​obtained. And, based on the input matrix and the output matrix, the output layer weight parameters in the RBF neural network are updated to obtain the state prediction value of the current data node in the updated RBF neural network. Finally, based on the difference between the input state value and the state prediction value, it is determined whether the target sensor issues a replacement instruction to the user.

[0036] In the implementation mode of the present application, by obtaining the deviation data of the target sensor, the change of the sensor during the aging process can be understood. This helps to monitor the working state of the sensor and take timely measures, such as repairing or replacing the sensor, to ensure the reliability and accuracy of the system. By updating the measurement monitoring matrix according to the deviation data, the description and update of the state change of the measurement model can be achieved. This helps to improve the measurement accuracy and accuracy, and ensure that the data output by the sensor is more reliable and accurate. By updating the weight parameters of the RBF neural network and the output matrix, the prediction model can be optimized. This can improve the prediction accuracy of the state of the target sensor, thereby helping users understand the current working state of the sensor. Based on the difference between the input state value and the state prediction value, it can be determined whether the performance of the target sensor has dropped to the extent that it needs to be replaced. This can help users to find faulty or aging sensors in time, and take corresponding maintenance or replacement measures to reduce the negative impact on the operation of the system. In general, the implementation mode of the present application can improve the effect of sensor monitoring and fault diagnosis, improve the reliability and stability of the system, and provide more accurate measurement results.

[0037] The technical solution provided in the embodiments of the present application can be implemented by a server and / or a terminal device. It should be noted that the server involved in the embodiments of the present application can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, big data and artificial intelligence platforms.

[0038] The terminal device involved in the embodiments of the present application may be a device that provides voice and / or data connectivity to a user, a handheld device with a wireless connection function, or other processing devices connected to a wireless modem. For example, a mobile phone (or "cellular" phone) and a computer with a mobile terminal, for example, a portable, pocket-sized, handheld, computer-built-in or vehicle-mounted mobile device, which exchanges voice and / or data with a wireless access network. For example, Personal Communication Service (PCS) phones, cordless phones, Session Initiation Protocol (SIP) phones, Wireless Local Loop (WLL) stations, Personal Digital Assistants (PDAs) and other devices.

[0039] It should be understood herein that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction rather than having any limiting meaning.

[0040] The principles and spirit of the present application are explained in detail below with reference to several representative implementations of the present application.

[0041] Reference below Figure 1 , Figure 1 A flow chart of a method for predicting sensor replacement status provided in an embodiment of the present application. A method for predicting sensor replacement status provided in an embodiment of the present application includes:

[0042] Step S101, obtaining deviation data of a target sensor.

[0043] The deviation data involved in this application is used to indicate the change of the sensor during the aging process. The deviation data can be a deviation data sequence for the target sensor. The deviation data sequence can reflect the aging process of the sensor.

[0044] The parameters contained in the deviation data in the present application may belong to slow time-varying objects. That is, further optionally, the deviation data comes from the slow time-varying objects of the target sensor. Among them, the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sensor sampling rate. It can be understood that a slow time-varying object refers to an object that changes slowly within a certain time range, and the change speed is much slower than the frequency of sensor sampling. In the application of sensors, slow time-varying objects usually refer to the changes of some characteristics of the measured object on a time scale, such as temperature, humidity, light intensity, etc. The changes of these parameters are relatively slow, and it takes a certain amount of time to appear, and the speed of change is very small compared to the time scale, so they can be considered as slow time-varying objects. For slow time-varying objects, it is necessary to use appropriate algorithms and models for prediction or modeling to ensure the accuracy and stability of the prediction results. For the sensor life prediction method based on the online structure variable model, it is very important to accurately describe the changing characteristics of the slow time-varying object. Only through effective modeling and accurate prediction can the accurate evaluation and timely maintenance of the sensor state be achieved.

[0045] In order to fit the time-varying characteristics of the deviation data sequence, a model that can also change can be selected (i.e., the measurement monitoring matrix involved in this application, see below). In addition, in order to improve the real-time performance of the online monitoring model, the online monitoring model can select a variable model with a faster update speed. For example, a variable model whose model structure can be changed online is selected, that is, a variable model that is not limited to simple parameter adaptation, but a variable model that completes adaptive modeling by integrating structural flexibility and sparse modeling. For example, a new online learning RBF model with an adjustable structure can be selected in this application.

[0046] In the embodiment of the present application, the deviation data includes but is not limited to: deviation input data and / or deviation output data of the target sensor. Exemplarily, the deviation input data of the sensor may include the following aspects:

[0047] Sampling time: Record the time point of each sampling, which is used to establish a time series.

[0048] Sensor output value: The true value obtained by the sensor by measuring physical quantities (such as temperature, pressure, humidity, etc.).

[0049] Theoretical Expected Value: The predicted or expected theoretical value based on the sensor’s design parameters and operating environment.

[0050] Environmental factors: Record environmental conditions related to the sensor, such as temperature, humidity, air pressure, etc.

[0051] Workload: Records the workload or working conditions the sensor is exposed to, such as load, vibration, etc.

[0052] Maintenance operations: Record the sensor's maintenance operations, calibration and other processing processes.

[0053] The collection and analysis of these deviation input data can help determine the aging degree and remaining life of the sensor, thereby achieving prediction and evaluation of the sensor.

[0054] Exemplarily, the deviation output data of the sensor may include the following aspects:

[0055] Deviation of sensor output value: The difference between the actual value measured by the sensor and the theoretical expected value.

[0056] Measurement Error: The error value obtained when calibrating a sensor using standard quantities and calibration equipment.

[0057] Stability index: records the stability or changes of sensor output values, such as variance, standard deviation, etc.

[0058] Drift change: The change in sensor output value over time can be used to determine the long-term stability of the sensor.

[0059] Fault Occurrences: Records the number of times a sensor fails or fails within a specific time period.

[0060] By collecting and analyzing these deviation output data, the performance and fault status of the sensor can be evaluated and predicted, helping to promptly detect sensor problems in actual applications and take corresponding measures.

[0061] Step S102: updating the measurement monitoring matrix corresponding to the target sensor according to the deviation data.

[0062] The measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor.

[0063] Among them, the measurement monitoring matrix includes an input matrix, an output matrix, and a radial basis function (RBF) information matrix. The input matrix and RBF neural network are usually used to describe and extract the features of the input data for training and testing machine learning models. The output matrix is ​​used to record the results of the model output in order to train and verify the model. For classification problems, the output matrix is ​​usually used to record the different categories to which each sample belongs, while for regression problems, the output matrix usually records the target variable value of each sample. The above three matrices are used to describe different aspects of the online monitoring model, and there is a certain correlation between them. Therefore, by using these matrices reasonably, the optimization and update of the online monitoring model can be effectively implemented to improve the real-time performance of the online monitoring model and overcome the lag in related technologies.

[0064] Specifically, in an optional implementation of steps S101-S102, obtaining deviation data of the target sensor in step S101 can be implemented as follows:

[0065] Obtain the deviation input data according to the sampling time and the deviation output data ; Among them, the deviation data value of each data node corresponds one by one to the sampling value at each acquisition time.

[0066] Furthermore, in step S102, updating the measurement monitoring matrix corresponding to the target sensor according to the deviation data can be implemented as follows:

[0067] First, use the bias input data , update the input matrix ; Wherein, the number of information required in the input matrix . Then, the deviation output data is used , update the output matrix ; Wherein, the output matrix dimension ; is the input matrix dimension. Further, represents the information vector of the RBF neural network, express The corresponding weight coefficient is, express The corresponding error term.

[0068] Finally, using the bias input data And deviation output data , update the RBF neural network

[0069] .in, is the RBF regressor for the i-th node in the RBF neural network, is the RBF regressor of the Mth column in the RBF neural network, M is the number of RBF regressor columns in the RBF neural network, is the p-th row RBF regression vector in the RBF neural network.

[0070] It is understandable that the RBF model corresponding to the RBF neural network can adopt the following Gaussian model .in, is the input vector of the RBF neural network, is the center of each node in the RBF network node structure parameters, is the diagonal covariance matrix of the ith node. Further optionally, the number of RBF nodes included in the RBF model is greater than 3, and the number of RBF nodes included in the RBF model is less than 6.

[0071] It is worth mentioning that the input matrix is ​​an m×n matrix, where each row represents a sample and each column represents an input feature. This matrix is ​​used to describe the feature information of the input data, where m represents the number of samples and n represents the number of input features.

[0072] RBF Neural Network (RBF Information Matrix): It is an m×k matrix, where each row represents a sample and each column represents the basis center of a radial basis function. This matrix is ​​used to describe the location information of the radial basis function, and k represents the number of radial basis functions.

[0073] Output Matrix: It is an m×1 vector that represents the output value of the system model for each input sample. This vector is used to describe the output results of the modeling system.

[0074] When using the RBF model for modeling, the input matrix and RBF neural network are used to determine the parameters of the radial basis function, and the output matrix of the system is used to train and verify the accuracy of the model. By adjusting the basis center position of the RBF neural network and the weights between the input matrix and the RBF neural network, a reasonable prediction of the system output can be achieved.

[0075] Of course, the above update sequence is not limited, and the three matrices can be updated separately or simultaneously.

[0076] For example, in an optional example of steps S101-S102, see Figure 2The prediction process shown in the figure can also be initialized before step S101 to set the number of iterations. Then, in step S101, the deviation input data is collected according to the sampling time. and deviation output data In step S102, according to the deviation input data and deviation output data Calculate and update each matrix. The update process is as follows:

[0077] Assume that the input data matrix is , then the deviation input data is It is Item input matrix vector; is the number of information, which determines the number of error variables before the current sampling time used by the algorithm; is the input dimension, that is, the input matrix dimension; when When , the algorithm degenerates into the traditional RLS algorithm. Further, optionally, in this embodiment, .

[0078] Furthermore, it is assumed that the RBF neural network is expressed as follows:

[0079] ;

[0080] in, is the RBF regressor for the i-th node in the RBF neural network, is the RBF regressor of the Mth column in the RBF neural network, M is the number of RBF regressor columns in the RBF neural network, is the p-th row RBF regression vector in the RBF neural network.

[0081] Furthermore, it is assumed that the output matrix, that is, the predicted deviation matrix, is expressed as follows:

[0082] .

[0083] Among them, the weight vector ;

[0084] Residual vector ;

[0085] is the output of the RBF model, is the output; is the output dimension of the system. Further, optionally, in this embodiment, .

[0086] Further optionally, in this embodiment, the RBF model may adopt a Gaussian model, which is expressed as the following formula:

[0087] ; Among them, the center point vector can be expressed as: ;No. The covariance matrix of RBF nodes can be expressed as:

[0088] ;in, Nx is the model input dimension or the number of input channels, Indicates The covariance matrix of the RBF nodes corresponding to the last dimension or number of channels.

[0089] In order to adapt the RBF structure to changing scenarios, the model can be adjusted by replacing the node that contributes the least to the overall performance (i.e., the node with the smallest contribution) with a new node. The specific method is described below.

[0090] Further optionally, after step S102, it is also possible to detect whether the current data node of the target sensor in the measurement monitoring matrix belongs to the minimum contribution node. If the current data node belongs to the minimum contribution node, the current data node is replaced with a newly added data node, and a swarm intelligence algorithm is used to obtain the best center point and variance of the newly added data node. If the current data node does not belong to the minimum contribution node, jump to the following step S103.

[0091] As an optional embodiment, in the above steps, detecting whether the current data node of the target sensor in the measurement monitoring matrix belongs to the minimum contribution node can be implemented as follows: , find the data node in the measurement monitoring matrix ; The range of node quantity is expressed as , IEV N Function used to measure data nodes distance;

[0092] Determine the current data node Whether the following detection formula is satisfied: and ;

[0093] If the current data node If the detection formula is satisfied, the current data node is determined is the node with the minimum contribution; if the current data node If the detection formula is not satisfied, the current data node is determined is not the node with the minimum contribution. and It is the judgment threshold in the detection formula.

[0094] Further optionally, the judgment threshold in the detection formula is set to: .

[0095] Among them, The contribution of each RBF node to the overall performance is obtained as the increment of error variance (IEV), which can be expressed as . Further optionally, in this embodiment, 6 RBF nodes are set, and .

[0096] As an optional embodiment, in the above steps, the swarm intelligence algorithm can adopt at least one of the particle swarm algorithm, the non-dominated sorting algorithm, and the ant colony algorithm. The following embodiment takes the non-dominated sorting algorithm as an example, and first initializes the algorithm. That is, randomly initialize the starting population , and Perform non-dominated sorting and initialize the rank value of each individual. Assume that the initial setting is Next, select the Select individuals and perform selection crossover operations to generate a new generation of populations . Then, calculate the loss function of the new population Furthermore, by merging and Generate new populations

[0097] Furthermore, Perform non-dominated sorting and select N individuals through the elite retention strategy to form a new generation of population . Then, jump to the previous article on choosing from the binary tournament method Select the individual step and continue the loop until the center point and variance value that satisfy the loss function are found, and then stop the loop.

[0098] Step S103: acquiring the input state value of the current data node of the target sensor in the input matrix.

[0099] Step S104, based on the input matrix and the output matrix, updating the output layer weight parameters in the RBF neural network to obtain the state prediction value of the current data node in the updated RBF neural network.

[0100] Continuing with the above example, update the weight vector of the RBF output layer , output state prediction value. Among them, the weight value vector can be expressed as , the state prediction value of the current data node is expressed as

[0101] .

[0102] Step S105 , based on the difference between the input state value and the state prediction value, it is determined whether the target sensor issues a replacement instruction to the user.

[0103] As an optional embodiment, in the above steps, based on the difference between the input state value and the state prediction value, determining whether the target sensor issues a replacement instruction to the user can be implemented as follows: calculating the difference between the input state value and the state prediction value; judging the target sensor range to which the difference belongs; the target range is one of a plurality of preset sensor ranges; if the difference belongs to the first sensor range, determining that the sensor is in a healthy state; or, if the difference belongs to the second sensor range, determining that the sensor is in a sub-healthy state, and detecting whether the parameter change rate of the target sensor reaches a preset threshold within a preset time period; if the parameter change rate of the target sensor reaches the preset threshold, issuing the replacement instruction to the user; or, if the difference belongs to the third sensor range, determining that the sensor is in an unhealthy state, and issuing the replacement instruction to the user.

[0104] For example, if the difference between the input state value and the state prediction value is within ±10% of the range value of the target sensor (i.e., the range of the first sensor), it means that the target sensor is in a healthy state. (i.e., the second sensor range), indicating that the target sensor is in a sub-healthy state. In this case, within the preset period, if the state prediction value (i.e., parameter change rate) of the data node is between During the interval, users need to be notified to replace the sensors in time.

[0105] If the difference between the input state value and the state prediction value is greater than ±30% of the range value of the target sensor (i.e., the range of the third sensor), it means that the target sensor is in an unhealthy state. In this case, the user needs to be notified to replace the new sensor in time.

[0106] Optionally, after step S105, Figure 2 As shown, set the next sampling time , and return to step S102 to update each matrix.

[0107] In the implementation mode of the present application, through the above steps S101 to S105, the automatic monitoring of the sensor is realized by updating the measurement monitoring matrix through the slowly time-varying objects, which greatly improves the monitoring accuracy, helps to timely detect sensor faults or failures, improves the maintenance and replacement efficiency of the sensor, and improves the reliability and stability of the smart city system.

[0108] After introducing the method of the exemplary embodiment of the present application, next, refer to Figure 3 A sensor replacement status prediction device according to an exemplary embodiment of the present application is described. The sensor replacement status prediction device can be provided in an online monitoring device, and the device includes:

[0109] The acquisition module 301 is used to acquire the deviation data of the target sensor; wherein the deviation data includes the deviation input data and / or the deviation output data of the target sensor; the deviation data is used to indicate the change of the sensor during the aging process; the deviation data comes from the slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sampling rate of the sensor;

[0110] An updating module 302 is used to update the measurement monitoring matrix corresponding to the target sensor according to the deviation data; the measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and an RBF neural network;

[0111] The prediction module 303 is used to obtain the input state value of the current data node of the target sensor in the input matrix; based on the input matrix and the output matrix, update the output layer weight parameters in the RBF neural network to obtain the state prediction value of the current data node in the updated RBF neural network; based on the difference between the input state value and the state prediction value, determine whether the target sensor issues a target sensor replacement instruction to the user.

[0112] As an optional implementation manner, when the prediction module 303 determines whether the target sensor issues a replacement instruction of the target sensor to the user based on the difference between the input state value and the state prediction value, it is specifically used to:

[0113] Calculating a difference between the input state value and the state prediction value;

[0114] Determine the target sensor range to which the difference belongs; the target range is one of a plurality of preset sensor ranges;

[0115] If the difference falls within the range of the first sensor, it is determined that the sensor is in a healthy state; or

[0116] If the difference falls within the range of the second sensor, the sensor is determined to be in a sub-healthy state, and the parameter change rate of the target sensor is detected within a preset period of time to see whether it reaches a preset threshold; if the parameter change rate of the target sensor reaches the preset threshold, the replacement instruction is issued to the user; or

[0117] If the difference falls within the third sensor range, it is determined that the sensor is in an unhealthy state, and the replacement instruction is issued to the user.

[0118] As an optional implementation, the acquisition module 301 is specifically used for:

[0119] Obtain the deviation input data according to the sampling time and the deviation output data ; Among them, the deviation data value of each data node corresponds to the sampling value at each acquisition time;

[0120] The updating of the measurement monitoring matrix corresponding to the target sensor according to the deviation data includes:

[0121] Input data using the deviation , update the input matrix ; Wherein, the number of information required in the input matrix ; is the input matrix dimension;

[0122] Output data using the deviation , update the output matrix ; Wherein, the output matrix dimension ; represents the information vector of the RBF neural network, express The corresponding weight coefficient is, express The corresponding error term;

[0123] Input data using the deviation And the deviation output data , update the RBF neural network ;

[0124] in, is the RBF regressor for the i-th node in the RBF neural network, is the RBF regressor of the Mth column in the RBF neural network, M is the number of RBF regressor columns in the RBF neural network, is the pth row RBF regression vector in the RBF neural network; the RBF model corresponding to the RBF neural network adopts the Gaussian model ,in, is the input vector of the RBF neural network, is the center of each node in the RBF network node structure parameters, is the diagonal covariance matrix of the ith node.

[0125] As an optional implementation manner, the number of RBF nodes included in the RBF model is greater than 3, and the number of RBF nodes included in the RBF model is less than 6.

[0126] As an optional implementation, the device further includes a replacement module, specifically configured to:

[0127] Detecting whether the current data node of the target sensor in the measurement monitoring matrix belongs to the minimum contribution node;

[0128] If the current data node belongs to the minimum contribution node, the current data node is replaced with a newly added data node, and a swarm intelligence algorithm is used to obtain the optimal center point and variance of the newly added data node;

[0129] If the current data node does not belong to the minimum contribution node, the prediction module 303 is triggered to execute the function of obtaining the input state value of the current data node of the target sensor in the input matrix.

[0130] As an optional implementation manner, when the replacement module detects whether the current data node of the target sensor in the measurement monitoring matrix belongs to the minimum contribution node, it is specifically used to:

[0131] according to , find the data node in the measurement monitoring matrix ; The range of node quantity is expressed as , IEV N Function used to measure data nodes distance;

[0132] Determine the current data node Whether the following detection formula is satisfied: and ;in, and is the judgment threshold in the detection formula. If the current data node Satisfying the detection formula, is the node with the minimum contribution; if the current data node If the detection formula is not satisfied, the current data node is determined Not the minimum contribution node.

[0133] As an optional implementation, the judgment threshold in the detection formula is set to: .

[0134] In the embodiment of the present application, through the prediction device of the sensor replacement status, the automatic monitoring of the sensor is realized by updating the measurement monitoring matrix through the slow time-varying object, which greatly improves the monitoring accuracy, helps to timely detect the fault or failure of the sensor, improves the maintenance and replacement efficiency of the sensor, and improves the reliability and stability of the smart city system.

[0135] After introducing the devices, methods and apparatus of the exemplary embodiments of the present application, next, reference is made to Figure 4 The computer-readable storage medium of the exemplary embodiment of the present application is described. The computer-readable storage medium can be set in the prediction device of the sensor replacement state. Please refer to Figure 4 , the computer-readable storage medium shown is a CD 110, on which a computer program (i.e., a program product) is stored. When the computer program is executed by the processor, each step recorded in the above method implementation is implemented, for example, obtaining deviation data of the target sensor; wherein the deviation data includes deviation input data and / or deviation output data of the target sensor; the deviation data is used to indicate the change of the sensor during the aging process; the deviation data comes from a slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sensor sampling rate; the measurement monitoring matrix corresponding to the target sensor is updated according to the deviation data; the measurement The monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and a radial basis function RBF neural network; the input state value of the current data node of the target sensor in the input matrix is ​​obtained; based on the input matrix and the output matrix, the output layer weight parameters in the RBF neural network are updated to obtain the state prediction value of the current data node in the updated RBF neural network; based on the difference between the input state value and the state prediction value, it is determined whether the target sensor issues a replacement instruction to the user; the specific implementation method of each step will not be repeated here.

[0136] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0137] After introducing the devices, methods, media and apparatus of the exemplary embodiments of the present application, reference is now made to Figure 5 For the computing device for predicting the replacement status of a sensor according to the exemplary embodiment of the present application, the computing device may be provided in a device for predicting the replacement status of a sensor.

[0138] Figure 5 A block diagram of an exemplary computing device 120 suitable for implementing embodiments of the present application is shown. The computing device 120 may be a computer system or a server. Figure 5 The computing device 120 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0139] like Figure 5 As shown, the components of the computing device 120 may include, but are not limited to: one or more processors or processing units 1201 , a system memory 1202 , and a bus 1203 connecting different system components (including the system memory 1202 and the processing unit 1201 ).

[0140] The computing device 120 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computing device 120, including volatile and nonvolatile media, removable and non-removable media.

[0141] System memory 1202 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 12021 and / or cache memory 12022. Computing device 120 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, ROM 12023 may be used to read and write non-removable, non-volatile magnetic media ( Figure 5 is not shown in the Figure 5As shown in the figure, a disk drive for reading and writing a removable non-volatile disk (such as a "floppy disk") and an optical disk drive for reading and writing a removable non-volatile optical disk (such as a CD-ROM, DVD-ROM or other optical media) can be provided. In these cases, each drive can be connected to the bus 1203 through one or more data medium interfaces. The system memory 1202 may include at least one program product, which has a set (such as at least one) of program modules, which are configured to perform the functions of each embodiment of the present application.

[0142] A program / utility 12025 having a set (at least one) of program modules 12024 may be stored, for example, in system memory 1202, and such program modules 12024 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. Program modules 12024 generally perform the functions and / or methods of the embodiments described herein.

[0143] The computing device 120 may also communicate with one or more external devices 1204 (e.g., a keyboard, a pointing device, a display, etc.). Such communication may be performed via an input / output (I / O) interface 605. In addition, the computing device 120 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network such as the Internet) via a network adapter 1206. Figure 5 As shown, the network adapter 1206 communicates with other modules (such as the processing unit 1201, etc.) of the computing device 120 via the bus 1203. It should be understood that although Figure 5 Not shown, other hardware and / or software modules may be used in conjunction with computing device 120 .

[0144] The processing unit 1201 executes various functional applications and data processing by running the program stored in the system memory 1202, for example, obtaining the deviation data of the target sensor; wherein the deviation data includes the deviation input data and / or deviation output data of the target sensor; the deviation data is used to indicate the change of the sensor during the aging process; the deviation data comes from the slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity with a change speed lower than the sensor sampling rate; the measurement monitoring matrix corresponding to the target sensor is updated according to the deviation data; the measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and a radial basis function RBF neural network; obtaining the input state value of the current data node of the target sensor in the input matrix; based on the input matrix and the output matrix, updating the output layer weight parameter in the RBF neural network to obtain the state prediction value of the current data node in the updated RBF neural network; based on the difference between the input state value and the state prediction value, determining whether the target sensor issues a replacement instruction to the user. The specific implementation method of each step is not repeated here. It should be noted that, although several units / modules or sub-units / sub-modules of the prediction device for sensor replacement status are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiment of the present application, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided to be embodied by multiple units / modules.

[0145] In the description of the present application, it should be noted that the terms "first", "second" and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0146] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0147] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0148] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0149] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0150] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program codes.

[0151] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0152] In addition, although the operations of the method of the present application are described in a specific order in the drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

Claims

1. A method for predicting sensor replacement status, characterized in that: The method comprises: Obtaining deviation data of a target sensor; wherein the deviation data includes deviation input data and / or deviation output data of the target sensor; the deviation data is used to indicate changes in the sensor during aging; the deviation data comes from a slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sensor sampling rate; Update the measurement monitoring matrix corresponding to the target sensor according to the deviation data; the measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and a radial basis function RBF neural network; Obtaining an input state value of a current data node of the target sensor in the input matrix; Based on the input matrix and the output matrix, the output layer weight parameters in the RBF neural network are updated to obtain the state prediction value of the current data node in the updated RBF neural network; the weight value vector is expressed as , the state prediction value of the current data node is expressed as ; is the output matrix, Represents the information vector of the RBF neural network; Based on the difference between the input state value and the state prediction value, determining whether the target sensor issues a replacement instruction of the target sensor to the user; Wherein, after updating the measurement monitoring matrix corresponding to the target sensor according to the deviation data, the method further includes: Detecting whether the current data node of the target sensor in the measurement monitoring matrix belongs to the minimum contribution node; If the current data node belongs to the minimum contribution node, the current data node is replaced with a newly added data node, and a swarm intelligence algorithm is used to obtain the optimal center point and variance of the newly added data node; If the current data node does not belong to the minimum contribution node, jump to the step of obtaining the input state value of the current data node of the target sensor in the input matrix; The detecting whether the current data node of the target sensor in the measurement monitoring matrix belongs to a minimum contribution node includes: according to , find the data node in the measurement monitoring matrix ; The range of node quantity is expressed as , Function used to measure data nodes distance; Determine the current data node Whether the following detection formula is satisfied: and ; in, and is the judgment threshold in the detection formula; If the current data node If the detection formula is satisfied, the current data node is determined is the node with the minimum contribution; If the current data node If the detection formula is not satisfied, the current data node is determined Not the minimum contribution node; The judgment threshold in the detection formula is set as: .

2. The method according to claim 1, characterized in that The determining, based on the difference between the input state value and the state prediction value, whether the target sensor issues a replacement instruction to the user, comprises: Calculating a difference between the input state value and the state prediction value; Determine the target sensor range to which the difference belongs; the target sensor range is one of a plurality of preset sensor ranges; If the difference falls within the range of the first sensor, it is determined that the sensor is in a healthy state; or If the difference falls within the range of the second sensor, the sensor is determined to be in a sub-healthy state, and the parameter change rate of the target sensor is detected within a preset period of time to see whether it reaches a preset threshold; if the parameter change rate of the target sensor reaches the preset threshold, the replacement instruction is issued to the user; or If the difference falls within the third sensor range, it is determined that the sensor is in an unhealthy state, and the replacement instruction is issued to the user.

3. The method according to claim 1, characterized in that The step of obtaining the deviation data of the target sensor includes: Obtain the deviation input data according to the sampling time and the deviation output data ; Among them, the deviation data value of each data node corresponds to the sampling value at each acquisition time; The updating of the measurement monitoring matrix corresponding to the target sensor according to the deviation data includes: Input data using the deviation , update the input matrix ; Wherein, the number of information required in the input matrix ; is the input matrix dimension; Output data using the deviation , update the output matrix ;in, is the output matrix dimension, , represents the information vector of the RBF neural network, express The corresponding weight coefficient is, express The corresponding error term; Input data using the deviation And the deviation output data , update the RBF neural network ; in, is the RBF regressor for the ith node in the RBF neural network, is the Mth column of RBF regressors in the RBF neural network, M is the number of RBF regressor columns in the RBF neural network, is the pth row RBF regression vector in the RBF neural network, and the RBF model corresponding to the RBF neural network adopts the Gaussian model ,in, is the input vector of the RBF neural network, is the center of each node in the RBF network node structure parameters, is the diagonal covariance matrix of the ith node.

4. The method according to claim 3, characterized in that The number of RBF nodes included in the RBF model is greater than 3, and the number of RBF nodes included in the RBF model is less than 6.

5. A sensor replacement status prediction device, characterized in that: The device comprises: An acquisition module, used for acquiring deviation data of a target sensor; wherein the deviation data includes deviation input data and / or deviation output data of the target sensor; the deviation data is used for indicating the change of the sensor during the aging process; the deviation data comes from a slow time-varying object of the target sensor; the slow time-varying object is a measurement monitoring quantity whose change speed is lower than the sampling rate of the sensor; An updating module, used for updating the measurement monitoring matrix corresponding to the target sensor according to the deviation data; the measurement monitoring matrix is ​​used to describe the state change of the online monitoring model corresponding to the target sensor; the measurement monitoring matrix includes an input matrix, an output matrix, and a radial basis function RBF neural network; A prediction module is used to obtain the input state value of the current data node of the target sensor in the input matrix; update the weight parameters of each data node in the output matrix through the output of the RBF neural network, and obtain the state prediction value of the current data node of the target sensor in the updated output matrix; based on the difference between the input state value and the state prediction value, determine whether the target sensor issues a replacement instruction to the user; wherein the weight value vector is expressed as , the state prediction value of the current data node is expressed as ; The device further includes a replacement module for detecting whether the current data node of the target sensor in the measurement monitoring matrix belongs to the minimum contribution node; if the current data node belongs to the minimum contribution node, the current data node is replaced with a newly added data node, and a swarm intelligence algorithm is used to obtain the optimal center point and variance of the newly added data node; if the current data node does not belong to the minimum contribution node, the prediction module is triggered to execute the operation of obtaining the input state value of the current data node of the target sensor in the input matrix; The replacement module is specifically used according to , find the data node in the measurement monitoring matrix ; The range of node quantity is expressed as , Function used to measure data nodes Distance; Determine the current data node Whether the following detection formula is satisfied: and ;in, and is the judgment threshold in the detection formula; If the current data node If the detection formula is satisfied, the current data node is determined is the node with the minimum contribution; if the current data node If the detection formula is not satisfied, the current data node is determined It is not the node with the minimum contribution; the judgment threshold in the detection formula is set to: .

6. A computing device, characterized in that The computing device comprises: at least one processor, memory, and input-output unit; The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the method for predicting the sensor replacement status according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that: The method comprises instructions, which, when executed on a computer, cause the computer to execute the method for predicting a sensor replacement status according to any one of claims 1 to 4.

Images