A solar insecticidal lamp internet of things node detection method and system

Abstract

The application discloses a kind of solar insecticidal lamp internet of things internet of things node detection method and system, wherein the method includes obtaining the characteristic value of each internet of things node of solar insecticidal lamp internet of things n time, monitoring value, geographic coordinate point and the characteristic value of each effective element of each internet of things node;Using the geographic coordinate point obtained and the distance threshold value preset, the neighbor node of each internet of things node is determined;The weight of each internet of things node is determined using the characteristic value of each internet of things node and its neighbor node;The characteristic value of each internet of things node, the characteristic value of each effective element of each internet of things node and the weight of each internet of things node are input into the trained space regression model to obtain the predicted value of each internet of things node, the size relationship of the difference between the predicted value and monitoring value of each internet of things node and monitoring threshold value is compared, and the detection result of each internet of things node is determined.The application can monitor whether internet of things node fails.

Description

TECHNICAL FIELD

[0001] The application relates to a solar insecticidal lamp Internet of Things node detection method and system, and belongs to the technical field of agricultural Internet of Things device detection. BACKGROUND

[0002] The solar insecticidal lamp is an energy-saving and environment-friendly agricultural pest control and monitoring device. After the Internet of Things module is configured, corresponding insecticidal data at different time periods, climate conditions and component states and other information can be transmitted to the user end, so that pest information statistics and prediction and device maintenance are facilitated. However, in most cases, the solar insecticidal lamp Internet of Things node is deployed in a harsh environment. Therefore, they are prone to aging, theft or damage, and these factors will cause the solar insecticidal lamp Internet of Things node to malfunction and be abnormal, affecting its stable operation. On the other hand, the solar insecticidal lamp Internet of Things node is deployed in remote areas, and the maintenance personnel inspection cost is high. Therefore, it is very important to take remote and automatic fault detection to ensure the reliability of the solar insecticidal lamp Internet of Things node.

[0003] However, the solar insecticidal lamp Internet of Things node fault detection cannot be accurately performed through single Internet of Things node information. The current mainstream fault diagnosis strategy includes centralized detection and distributed detection, wherein the centralized detection needs to transmit all the information of the node to the background, which will cause a large amount of data transmission and is not suitable for the solar insecticidal lamp Internet of Things which is widely deployed.

[0004] Therefore, the application provides a solar insecticidal lamp Internet of Things node detection method and system. SUMMARY

[0005] The application aims to overcome the deficiencies in the prior art, and provides a solar insecticidal lamp Internet of Things node detection method and system, which can monitor whether the Internet of Things node malfunctions.

[0006] To achieve the above-mentioned purpose, the application is implemented by using the following technical scheme:

[0007] On one hand, the application provides a solar insecticidal lamp Internet of Things node detection method, which comprises the following steps:

[0008] obtaining characteristic values, monitoring values, geographic coordinate points of each Internet of Things node at n moments of the solar insecticidal lamp Internet of Things, and characteristic values of each effective element of each Internet of Things node;

[0009] determining the neighbor nodes of each Internet of Things node by using the obtained geographic coordinate points and a preset distance threshold value;

[0010] determining the weight of each Internet of Things node by using the characteristic values of each Internet of Things node and its neighbor nodes;

[0011] The characteristic values of each Internet of Things node, the characteristic values of each effective element of each Internet of Things node, and the weights of each Internet of Things node are input into the trained spatial regression model to obtain predicted values of each Internet of Things node, and a size relationship between a difference between the predicted values and the monitoring values of each Internet of Things node and a monitoring threshold is compared to determine a detection result of each Internet of Things node.

[0012] Further, the effective elements are selected from one or more combinations of atmospheric temperature, relative humidity, temperature in an electrical box, battery voltage, battery current, grid voltage, grid current, lamp tube voltage, lamp tube current, solar panel voltage, solar panel current, light intensity, insecticidal voltage count, insecticidal sound count, and rain control module voltage.

[0013] Further, the neighbor nodes of each Internet of Things node are determined by using the obtained geographic coordinate points and the preset distance threshold, and the determination includes:

[0014] The distance between each Internet of Things node is determined according to the geographic coordinate points of the Internet of Things nodes, and when the distance between any two Internet of Things nodes is less than or equal to a distance threshold, the two Internet of Things nodes are determined to be neighbor nodes of each other.

[0015] Further, the weights of each Internet of Things node are determined by using the characteristic values of the Internet of Things nodes and their neighbor nodes, and the determination includes:

[0016] The root mean square error of each Internet of Things node is determined according to the characteristic values of the Internet of Things nodes, and the determination includes the following formula:

[0017] s={s1,s2,……s i ……}

[0018]

[0019] In the formula, s is the root mean square error of the Internet of Things node, s i is the root mean square error of the effective element of the i-th neighbor node of the Internet of Things node, y k is the characteristic value of the current Internet of Things node at the k-th moment, y i，k is the characteristic value of the i-th neighbor node of the current Internet of Things node at the k-th moment, 0

[0020] The reciprocal of the root mean square error of each Internet of Things node is taken as the weight of each Internet of Things node, and the determination includes the following formula:

[0021]

[0022] w={w1,w2,……w i ……}

[0023] In the formula, w is the weight of the Internet of Things node, wi a weight of an effective element of an i-th neighbor node of the IoT node, a sum of inverse square roots of errors of effective elements of all neighbor nodes of the IoT node.

[0024] Further, the spatial regression model comprises the following formula:

[0025]

[0026] wherein w i a weight of an effective element of an i-th neighbor node of the IoT node, y i an eigenvalue of an effective element of an i-th neighbor node of the IoT node, and y' is a predicted value of the IoT node.

[0027] Further, the comparison of the difference between the predicted value and the monitored value and the size of the monitoring threshold determines the detection result, which comprises:

[0028] When the difference between the predicted value and the monitored value of the same IoT node is greater than or equal to the monitoring threshold, it is determined that the IoT node has a fault corresponding to the monitored value; when the difference between the predicted value and the monitored value of the same IoT node is less than the monitoring threshold, it is determined that the IoT node does not have a fault corresponding to the monitored value.

[0029] Further, the detection result is a mismatch between the solar cell panel current and the light intensity value, a mismatch between the light intensity value and the rain control module, a mismatch between the atmospheric temperature and the temperature in the electrical box, and / or the solar insecticidal lamp does not normally light up.

[0030] Further, the monitored value comprises atmospheric temperature, relative humidity, temperature in the electrical box, battery voltage, battery current, grid voltage, grid current, lamp voltage, lamp current, solar cell panel voltage, solar cell panel current, light intensity, insecticidal voltage count, insecticidal sound count, and / or rain control module voltage.

[0031] Further, the Pearson correlation coefficient value of the effective element between each IoT node is greater than or equal to 0.7.

[0032] In another aspect, the present application provides a solar insecticidal lamp IoT node detection system of an IoT network, comprising:

[0033] An acquisition module is configured to acquire eigenvalues, monitored values, geographical coordinate points of each IoT node at n time points of a solar insecticidal lamp IoT network, and eigenvalues of each effective element of each IoT node;

[0034] A node module is configured to determine neighbor nodes of each IoT node by using the acquired geographical coordinate points and a preset distance threshold.

[0035] The weight module is configured to determine the weight of each IOT node by using the characteristic value of each IOT node and its neighbor nodes.

[0036] The judging module is configured to input the characteristic value of each IOT node, the characteristic value of each effective element of each IOT node and the weight of each IOT node into the trained spatial regression model to obtain the predicted value of each IOT node, compare the size relationship between the difference between the predicted value and the monitoring value of each IOT node and the monitoring threshold, and determine the detection result of each IOT node.

[0037] Compared with the prior art, the present application has the following beneficial effects:

[0038] The present application determines the weight of each IOT node by using the characteristic value of each IOT node, the monitoring value and the characteristic value of each effective element of each IOT node, obtains the predicted value of each IOT node, and determines the detection result of each IOT node according to the relationship between the predicted value and the monitoring value, so that the detection result of the present application has higher reliability compared with the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 An embodiment flow chart of training the spatial regression model of the present application is shown;

[0040] Figure 2 An embodiment matrix diagram of the spatial correlation of each effective element between each IOT node of the present application is shown;

[0041] Figure 3 A comparison bar chart of the fault detection accuracy of the method, DFD and DSFD of the present application under different fault conditions is shown;

[0042] Figure 4 An embodiment flow chart of the IOT node detection method of the IOT of the solar insecticidal lamp of the present application is shown;

[0043] Figure 5 A comparison bar chart of the fault detection accuracy of the method, DFD and DSFD of the present application under different number of neighbor nodes is shown. DETAILED DESCRIPTION

[0044] The present application will be further described below in conjunction with the drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and cannot be used to limit the protection scope of the present application.

[0045] In the description of the present application, it needs to be understood that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like is based on the orientation or positional relationship shown in the drawings, which is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element indicated must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second" and the like are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" and the like can be explicitly or implicitly included one or more. In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0046] In the description of the present application, it needs to be understood that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connection" should be understood in a broad sense, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through intermediate medium, or the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood through specific circumstances.

[0047] Embodiment 1

[0048] The present embodiment provides a kind of solar insecticidal lamp Internet of Things node detection method of Internet of Things.

[0049] Reference Figure 4 The solar insecticidal lamp Internet of Things node detection method of Internet of Things of the present embodiment includes the following steps:

[0050] S1 obtains the characteristic value, monitoring value, geographic coordinate point of each Internet of Things node of solar insecticidal lamp Internet of Things n time, and the characteristic value of each effective element of each Internet of Things node.

[0051] When applied, the effective element is selected from one or more combinations of atmospheric temperature, relative humidity, temperature in electrical box, battery voltage, battery current, grid voltage, grid current, lamp voltage, lamp current, solar panel voltage, solar panel current, light intensity, insecticidal voltage count, insecticidal sound count and rain control module voltage.

[0052] In actual application, the spatial correlation of each effective element is high.

[0053] S2 determines the neighbor node of each Internet of Things node by using the obtained geographic coordinate point and the preset distance threshold.

[0054] In the application, the distance between each Internet of Things node is determined according to the geographic coordinate points of the Internet of Things nodes, and when the distance between any two Internet of Things nodes is less than or equal to the distance threshold, the two Internet of Things nodes are defined as neighbor nodes.

[0055] In actual application, a virtual circle is drawn with the geographic coordinate point of any Internet of Things node as the center and the preset distance threshold as the radius, and other Internet of Things nodes within the virtual circle are neighbor nodes of the current Internet of Things node, that is, the number of other Internet of Things nodes within the virtual circle is the number of neighbor nodes of the current Internet of Things node.

[0056] S3 determines the weight of each Internet of Things node by using the characteristic value of the Internet of Things node and its neighbor nodes.

[0057] (1) The root mean square error of the Internet of Things node is determined according to the characteristic value of the Internet of Things node, including the following formula:

[0058] s={s1,s2,……s i ……}

[0059]

[0060] In the formula, s is the root mean square error of the Internet of Things node, s i is the root mean square error of the effective element of the i th neighbor node of the Internet of Things node, y k is the characteristic value of the current Internet of Things node at the k th moment, y i，k is the characteristic value of the i th neighbor node of the current Internet of Things node at the k th moment, 0 < k ≤ n, and n > 0.

[0061] (2) The reciprocal of the root mean square error of each Internet of Things node is defined as the weight of each Internet of Things node, including the following formula:

[0062]

[0063] w={w1,w2,……w i ……}

[0064] In the formula, w is the weight of the Internet of Things node, w i is the weight of the effective element of the i th neighbor node of the Internet of Things node, and is the sum of the reciprocal of the root mean square error of all effective elements of the neighbor nodes of the Internet of Things node.

[0065] S4 inputs the characteristic value of each Internet of Things node, the characteristic value of each effective element of each Internet of Things node, and the weight of each Internet of Things node into the trained spatial regression model to obtain the predicted value of each Internet of Things node, compares the difference between the predicted value and the monitoring value of each Internet of Things node with the monitoring threshold, and determines the detection result of each Internet of Things node.

[0066] In the application, the predicted value of the Internet of Things node is obtained according to the following formula:

[0067]

[0068] In the formula, w i is the weight of the effective element of the i th neighbor node of the Internet of Things node, y i is the characteristic value of the effective element of the i th neighbor node of the Internet of Things node, and y' is the predicted value of the Internet of Things node.

[0069] In actual application, the detection result is a fault that the solar cell panel current and the light intensity value are mismatched, the light intensity value and the rain control module are mismatched, the atmospheric temperature and the temperature in the electrical box are mismatched, and / or the solar insecticidal lamp is not normally turned on.

[0070] The above faults are difficult to detect by a single effective element. Among them, the mismatch between the solar cell panel current and the light intensity value is caused by the solar cell panel fault F1 CS and / or the light intensity sensor fault F1 L ; the mismatch between the light intensity value and the rain control module is caused by the light intensity sensor fault and / or the rain control module fault F2 VR ; the mismatch between the atmospheric temperature and the temperature in the electrical box is caused by the sensor fault F3 T0 for measuring the atmospheric temperature and / or the sensor fault F3 T1 for measuring the temperature in the electrical box; the solar insecticidal lamp is not normally turned on, which is caused by the lamp tube current anomaly F4 CL and / or the grid current anomaly F4 CM .

[0071] Therefore, the monitoring values include the atmospheric temperature T 0 , the relative humidity H, the temperature T 1 in the electrical box, the battery voltage V B , the battery current C B , the grid voltage V M , the grid current C M , the lamp tube voltage V L , the lamp tube current C L , the solar cell panel voltage V S , the solar cell panel current C S , the light intensity L, the insecticidal voltage count V C , the insecticidal sound count S C , and / or the rain control module voltage V R

[0072] When the difference between the predicted value and the monitored value of the same Internet of Things node is greater than or equal to the monitoring threshold, it is determined that the Internet of Things node has a fault corresponding to the monitored value; when the difference between the predicted value and the monitored value of the same Internet of Things node is less than the monitoring threshold, it is determined that the Internet of Things node does not have a fault corresponding to the monitored value. Those skilled in the art can adjust the size of the monitoring threshold according to actual conditions or experience.

[0073] The present application is directed to Internet of Things nodes with different spatial correlations, and the eigenvalues of the Internet of Things nodes, the monitored values, and the eigenvalues of each effective element of the Internet of Things nodes are used to determine the weights of the Internet of Things nodes to obtain the predicted values of the Internet of Things nodes, and the detection results of the Internet of Things nodes are determined according to the relationship between the predicted values and the monitored values. Compared with the prior art, the detection results of the present application have higher reliability.

[0074] Embodiment 2

[0075] Based on embodiment 1, the method for training the spatial regression model is described in detail.

[0076] Reference Figure 1 The method for training the spatial regression model of the present embodiment comprises the following steps:

[0077] Step 1: Obtain the historical eigenvalues of each Internet of Things node and the historical eigenvalues of each element of each Internet of Things node during the normal operation period of the solar insecticidal lamp Internet of Things.

[0078] In application, the historical total eigenvalues of each Internet of Things node and the historical total eigenvalues of each element of each Internet of Things node of the solar insecticidal lamp Internet of Things are first obtained, and then the historical fault data of each Internet of Things node and the historical fault eigenvalues of each element of each Internet of Things node during the fault period of the solar insecticidal lamp Internet of Things are filtered out to obtain the historical eigenvalues of each Internet of Things node and the historical eigenvalues of each element of each Internet of Things node during the normal operation period of the solar insecticidal lamp Internet of Things.

[0079] The historical eigenvalues of each Internet of Things node of the present embodiment each comprise a plurality of elements, and each element is: atmospheric temperature T 0 , relative humidity H, temperature T 1 in the electrical box, battery voltage V B , battery current C B , grid voltage V M , grid current C M , lamp voltage V L , lamp current C L , solar panel voltage V S , solar panel current C S , light intensity L, insecticidal voltage count V C , insecticidal sound count S CAnd rain control module voltage V R .

[0080] Step 2: Obtain the geographic coordinate points of each Internet of Things node, and determine the neighbor nodes of each Internet of Things node by using the obtained geographic coordinate points and the preset distance threshold.

[0081] In the application, a virtual circle is drawn with the geographic coordinate point of any Internet of Things node as the center and the preset distance threshold as the radius, and other Internet of Things nodes within the virtual circle are all neighbor nodes of the current Internet of Things node, that is, the number of other Internet of Things nodes within the virtual circle is the number of neighbor nodes of the current Internet of Things node.

[0082] The distance threshold of the embodiment is less than or equal to 80 meters, and a person skilled in the art can adjust the size of the distance threshold according to the actual situation.

[0083] Step 3: Based on the Pearson correlation coefficient formula, the spatial correlation of each element between each Internet of Things node is determined by using the historical characteristic values of each element, and the element with high spatial correlation is determined as an effective element.

[0084] S31 calculates the Pearson correlation coefficient value of each element between each Internet of Things node by using the obtained historical characteristic values of each element, and the calculation formula is as follows:

[0085]

[0086] r represents the Pearson correlation coefficient value, x represents the historical characteristic value of the target element, y represents the historical characteristic value of the other comparison element, and are the mean values of x and y, respectively.

[0087] S32 determines the high and low of the spatial correlation of each element according to the Pearson correlation coefficient value calculated by A and the preset correlation threshold, and filters out the elements with low spatial correlation, and the remaining elements are effective elements.

[0088] In the application, when the Pearson correlation coefficient value of any element between each Internet of Things node is greater than or equal to the correlation threshold, the spatial correlation of the corresponding element is high; when the Pearson correlation coefficient value of any element between each Internet of Things node is less than the correlation threshold, the spatial correlation of the corresponding element is low.

[0089] Since the low spatial correlation indicates that the trend of data change of the element is inconsistent between the Internet of Things nodes, therefore, the element with low spatial correlation and its corresponding historical characteristic value are filtered out, which will have a negative effect on the subsequent detection result.

[0090] The correlation threshold of the embodiment is 0.7, and a person skilled in the art can adjust the size of the correlation threshold according to the actual situation.

[0091] Step 4: determining the weight of each IoT node by using the eigenvalue of the IoT node and its neighbor nodes.

[0092] (1) determining the root mean square error of the IoT node according to the historical eigenvalue of the IoT node.

[0093] (2) using the reciprocal of the root mean square error of the IoT node as the weight of the effective element of the neighbor node of the IoT node.

[0094] Step 5: inputting the historical eigenvalue of each IoT node, the effective element, the historical eigenvalue of the effective element, and the weight of each IoT node into a spatial regression model to obtain the predicted value of each IoT node, comparing the difference between the predicted value and the monitoring value of each IoT node with the monitoring threshold to determine the detection result of each IoT node.

[0095] 5.1 obtaining the historical monitoring value of each IoT node.

[0096] 5.2 determining whether the IoT node has a fault by using the predicted value of the IoT node, the historical monitoring value, and the preset monitoring threshold.

[0097] When the difference between the predicted value and the historical monitoring value of the IoT node is greater than or equal to the monitoring threshold, it is determined that the IoT node has a fault corresponding to the historical monitoring value; when the difference between the predicted value and the historical monitoring value of the IoT node is less than the monitoring threshold, it is determined that the IoT node does not have a fault corresponding to the historical monitoring value.

[0098] Wherein, the fault specifically includes: solar cell panel current and light intensity value mismatch, light intensity value and rain control module mismatch, atmospheric temperature and temperature inside the electrical box mismatch, and solar insecticidal lamp not normally lighting.

[0099] In application, the solar cell panel current and light intensity value mismatch is caused by solar cell panel fault F1 CS and / or light intensity sensor fault F1 L ; the light intensity value and rain control module mismatch is caused by light intensity sensor fault and / or rain control module fault F2 VR ; the atmospheric temperature and temperature inside the electrical box mismatch is caused by sensor fault F3 T0 for measuring atmospheric temperature and / or sensor fault F3 T1 for measuring the temperature of the electrical box; the solar insecticidal lamp not normally lighting is caused by lamp tube current anomaly F4 CL and / or power grid current anomaly F4 CM .

[0100] In addition, the historical monitoring values include atmospheric temperature, relative humidity, temperature in the electrical box, battery voltage, battery current, grid voltage, grid current, lamp voltage, lamp current, solar panel voltage, solar panel current, light intensity, insecticidal voltage count, insecticidal sound count, and / or rain control module voltage.

[0101] In actual application, in order to take into account the two adverse situations of fault false detection and fault missed detection, the product of the standard deviation of each Internet of Things node and the intersection of the fault false detection rate and the fault missed detection rate is used as the monitoring threshold.

[0102] Wherein, the fault missed detection refers to the case that the current data is judged as no fault although the fault occurs; the fault false detection refers to the case that the current data is judged as fault although no fault occurs.

[0103] The person skilled in the art can estimate the fault false detection rate and the fault missed detection rate according to the historical situation, as follows:

[0104] e α = FP / (TP+FP)

[0105] e β = FN / (TP+FN)

[0106] In the formula, e α is the fault false detection rate, e β is the fault missed detection rate, TP is the number of times that the detection result is accurate in the historical statistics, FP is the number of times that the detection result is inaccurate in the historical statistics, and FN is the number of times that the detection result is inaccurate in the historical statistics.

[0107] Step 6: The detection result of the spatial regression model is sent to the background or user end through the ZigBee device, reducing unnecessary data transmission.

[0108] Embodiment 3

[0109] On the basis of Embodiment 1 or 2, this embodiment introduces an Internet of Things node detection method of a solar insecticidal lamp Internet of Things.

[0110] In this embodiment, seven Internet of Things nodes of the solar insecticidal lamp Internet of Things are actually deployed, and historical data from August 14, 2021 to October 4, 2021 are collected for verification.

[0111] Firstly, the historical characteristic values of the IoT node and its neighbor nodes are used to calculate the root mean square error of the IoT node by using the elements with high spatial correlation as effective elements, and the inverse of the root mean square error of the IoT node is used as the weight of the IoT node, and the historical characteristic data of the IoT node, the historical characteristic data of each effective element of the IoT node, and the weight of the IoT node are input into the trained spatial regression model to obtain the predicted value of the IoT node.

[0112] Therefore, in the embodiment, the atmospheric temperature T 0 , the temperature T 1 in the electrical box, the grid current C M , the lamp voltage C L , the solar panel current C S , the light intensity L, and the rain control module voltage V R are used as the effective elements of each IoT node. Figure 2 The effective elements of the seven IoT nodes have high Pearson correlation coefficients, that is, the effective elements have high spatial correlation between the IoT nodes.

[0113] Then, the difference between the obtained predicted value and the corresponding historical detection value is compared with a monitoring threshold to determine a detection result, and the detection result is sent to a background or a user end through a ZigBee device.

[0114] The detection results of the embodiment, the detection results of a distributed fault detection algorithm (DFD), and the detection results of a distributed self-fault diagnosis algorithm (DSFD) are compared and analyzed. Figure 3 Under different fault conditions, the fault detection accuracy of the embodiment has good performance, and compared with the DFD and the DSFD, the fault detection accuracy of the embodiment is more stable and has higher feasibility.

[0115] Further, the fault detection accuracies of the embodiment, the DFD, and the DSFD under different numbers of neighbor nodes are compared and analyzed. Figure 5 When the number of neighbor nodes is less than two, compared with the DFD and the DSFD, the embodiment has obvious advantages in fault detection accuracy.

[0116] In summary, compared with the prior art, the method of the embodiment can perform more accurate fault detection on the IoT nodes lacking neighbor nodes or having few neighbor nodes, and it can be seen that the application is suitable for accurately detecting faults that are difficult to determine by using single node single element information by using neighbor node information or effective elements.

[0117] Embodiment 4

[0118] The embodiment provides a solar insecticidal lamp Internet of Things node detection system, comprising:

[0119] an acquisition module, configured to acquire characteristic values, monitoring values, geographic coordinate points of each Internet of Things node at n moments of a solar insecticidal lamp Internet of Things, and characteristic values of each effective element of each Internet of Things node;

[0120] a node module, configured to determine neighbor nodes of each Internet of Things node by using the acquired geographic coordinate points and a preset distance threshold value;

[0121] a weight module, configured to determine weights of each Internet of Things node by using the characteristic values of each Internet of Things node and neighbor nodes thereof;

[0122] a judgment module, configured to input the characteristic values of each Internet of Things node, the characteristic values of each effective element of each Internet of Things node and the weights of each Internet of Things node into a trained spatial regression model to obtain predicted values of each Internet of Things node, compare a size relationship between a difference between the predicted values of each Internet of Things node and the monitoring values and a monitoring threshold value, and determine detection results of each Internet of Things node

[0123] The specific function implementation of each functional module is referable to the related content in the methods in Embodiments 1-3.

[0124] The above description is only the preferred embodiments of the present application, and it should be pointed out that, for ordinary skilled in the art, without departing from the technical principles of the present application, a number of improvements and modifications can be made, and these improvements and modifications should be considered as the protection scope of the present application.

Claims

1. A method for detecting IoT nodes in a solar-powered insecticidal lamp IoT system, characterized in that, Includes the following steps: Acquire the feature values, monitoring values, geographic coordinates, and feature values ​​of each effective element of each IoT node at n time points in the solar insecticidal lamp IoT system. Using the acquired geographic coordinates and preset distance thresholds, the neighboring nodes of each IoT node are determined; The weight of each IoT node is determined by using the feature values ​​of each IoT node and its neighboring nodes; The feature values ​​of each IoT node, the feature values ​​of each effective element of each IoT node, and the weights of each IoT node are input into the trained spatial regression model to obtain the predicted value of each IoT node. The difference between the predicted value and the monitored value of each IoT node is compared with the monitoring threshold to determine the detection result of each IoT node. The process of determining the neighboring nodes of each IoT node using the acquired geographic coordinates and a preset distance threshold includes: The distance between each IoT node is determined based on the geographical coordinates of the IoT nodes. When the distance between any two IoT nodes is less than or equal to the distance threshold, the two IoT nodes are designated as neighbor nodes. The process of determining the weight of each IoT node using the feature values ​​of each IoT node and its neighboring nodes includes: The root mean square error of an IoT node is determined based on its characteristic values, including the following formula: ， ， In the formula, s represents the root mean square error of the IoT node. For the first IoT node The root mean square error of the effective features of each neighboring node. Let be the feature value of the current IoT node at time k. For the current IoT node The feature values ​​of the neighboring nodes at time k, 0 < k ≤ n, n > 0; Let the reciprocal of the root mean square error of each IoT node be used as the weight of each IoT node, including the following formula: , , In the formula, For the weight of IoT nodes, For the first IoT node The weights of the effective elements of each neighboring node. It is the sum of the reciprocals of the root mean square errors of the effective elements of all neighboring nodes of the Internet of Things node; The spatial regression model includes the following formula: ， In the formula, For IoT node number The weights of effective features in each neighboring node. For the first IoT node The feature values ​​of the effective elements of each neighboring node, where y' is the predicted value of the IoT node; The effective elements have a Pearson correlation coefficient value of greater than or equal to 0.7 among the IoT nodes.

2. The IoT node detection method for the solar-powered insecticidal lamp IoT according to claim 1, characterized in that, The effective elements are selected from one or more combinations of atmospheric temperature, relative humidity, electrical box temperature, battery voltage, battery current, grid voltage, grid current, lamp voltage, lamp current, solar panel voltage, solar panel current, light intensity, insecticidal voltage count, insecticidal sound count, and rain control module voltage.

3. The IoT node detection method for the solar-powered insecticidal lamp IoT according to claim 1, characterized in that, The step of comparing the difference between the predicted and monitored values ​​of each IoT node with the monitoring threshold to determine the detection result includes: When the difference between the predicted value and the monitored value of the same IoT node is greater than or equal to the monitoring threshold, it is determined that the IoT node has a fault corresponding to the monitored value; when the difference between the predicted value and the monitored value of the same IoT node is less than the monitoring threshold, it is determined that the IoT node does not have a fault corresponding to the monitored value.

4. The IoT node detection method for the solar-powered insecticidal lamp IoT according to claim 1, characterized in that, The test results indicated a mismatch between the solar panel current and light intensity, a mismatch between the light intensity and the rain control module, a mismatch between the atmospheric temperature and the temperature inside the electrical box, and / or that the solar insecticidal lamp was not turned on properly.

5. The IoT node detection method for the solar-powered insecticidal lamp IoT according to claim 1, characterized in that, The monitored values ​​include atmospheric temperature, relative humidity, electrical box temperature, battery voltage, battery current, grid voltage, grid current, lamp voltage, lamp current, solar panel voltage, solar panel current, light intensity, insecticidal voltage count, insecticidal sound count, and / or rain control module voltage.

6. An IoT node detection system for a solar-powered insecticidal lamp, characterized in that, include: The acquisition module is used to acquire the feature values, monitoring values, geographic coordinates, and feature values ​​of each effective element of each IoT node at n time points in the solar insecticidal lamp IoT system. The node module is used to determine the neighboring nodes of each IoT node by using the acquired geographic coordinates and preset distance thresholds. The weighting module is used to determine the weight of each IoT node by utilizing the feature values ​​of each IoT node and its neighboring nodes. The judgment module is used to input the feature values ​​of each IoT node, the feature values ​​of each effective element of each IoT node, and the weights of each IoT node into the trained spatial regression model to obtain the predicted value of each IoT node, compare the difference between the predicted value and the monitored value of each IoT node with the monitoring threshold, and determine the detection result of each IoT node. The process of determining the neighboring nodes of each IoT node using the acquired geographic coordinates and a preset distance threshold includes: The distance between each IoT node is determined based on the geographical coordinates of the IoT nodes. When the distance between any two IoT nodes is less than or equal to the distance threshold, the two IoT nodes are designated as neighbor nodes. The process of determining the weight of each IoT node using the feature values ​​of each IoT node and its neighboring nodes includes: The root mean square error of an IoT node is determined based on its characteristic values, including the following formula: ， ， In the formula, s represents the root mean square error of the IoT node. For the first IoT node The root mean square error of the effective features of each neighboring node. Let be the feature value of the current IoT node at time k. For the current IoT node The feature values ​​of the neighboring nodes at time k, 0 < k ≤ n, n > 0; Let the reciprocal of the root mean square error of each IoT node be used as the weight of each IoT node, including the following formula: , , In the formula, For the weight of IoT nodes, For the first IoT node The weights of the effective elements of each neighboring node. It is the sum of the reciprocals of the root mean square errors of the effective elements of all neighboring nodes of the Internet of Things node; The spatial regression model includes the following formula: ， In the formula, For IoT node number The weights of effective features in each neighboring node. For the first IoT node The feature values ​​of the effective elements of each neighboring node, where y' is the predicted value of the IoT node; The effective elements have a Pearson correlation coefficient value of greater than or equal to 0.7 among the IoT nodes.

Images