A method for detecting abnormal state of environmental monitoring sensor

By using spatial mutual checking and time self-checking methods in environmental monitoring sensors, data abnormalities are detected and corrected, the problem of sensor data abnormalities is solved, and monitoring accuracy and stability are improved.

CN114659672BActive Publication Date: 2024-12-06ROCKET FORCE UNIV OF ENG
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210219691.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-08
Publication Date
2024-12-06
Estimated Expiration
2042-03-08

AI Technical Summary

Technical Problem

Environmental monitoring sensors are prone to data abnormalities during use, including data loss and singular values, resulting in accuracy and functional abnormalities. It is difficult for the prior art to effectively detect and solve these problems.

Method used

Three sensors are used for spatial mutual verification and time self-checking. Through spatial self-checking, whether the data abnormality comes from the sensor itself, and the sensor accuracy spatial and temporal consistency verification is performed to determine the data abnormality sampling point.

Benefits of technology

Effectively detect and correct sensor data abnormalities, improve monitoring accuracy and stability, and ensure the reliability of environmental monitoring data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114659672B_ABST
    Figure CN114659672B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of sensor detection technology, and specifically relates to an abnormal state detection method for an environmental monitoring sensor. The method can be applied to data anomaly detection of wired or wireless sensors, and performs detection on two aspects: accuracy and functional anomaly of the sensor. Compared with the existing sensor data anomaly detection method, the present invention introduces time self-verification to perform two-stage verification, and further determines the data anomaly sampling point; uses space self-verification to determine whether the data anomaly comes from the sensor itself, and verifies the monitoring accuracy of the sensor.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of sensor detection, and in particular relates to a method for detecting abnormal state of an environmental monitoring sensor. Background Art

[0002] The current environmental monitoring system based on wireless sensor networks can realize continuous and long-term monitoring of environmental factors such as temperature and humidity in the monitoring area, saving manpower and material resources, and making it more convenient for people to perceive the environmental status and deal with it in a timely manner.

[0003] However, due to the limitation of production technology or the aging of sensor equipment as the use time increases, the data monitored by the sensor will inevitably be abnormal. In addition to the abnormal data monitoring caused by environmental changes, the unstable operation of the sensor and its own unreliability caused by the deployment environment will also cause sensor accuracy and function abnormalities. Among them, the sensor data mainly includes data loss and singular values. For example, the environmental monitoring sensor is unstable due to environmental changes, or it is interfered by electromagnetic radiation during operation, which will cause the sensor to produce data deviations of varying degrees. Based on this, the present invention designs a method for detecting abnormal conditions of environmental monitoring sensors. Summary of the invention

[0004] The purpose of the present invention is to provide a method for detecting abnormal state of an environmental monitoring sensor, which adopts three sensors for spatial mutual verification; compared with the existing sensor data abnormality verification method, it introduces time self-verification to perform two-stage verification, and then further determines the data abnormality sampling point; uses spatial self-verification to determine whether the data anomaly comes from the sensor itself, and verifies the monitoring accuracy of the sensor.

[0005] In order to achieve the above technical objectives, the technical solution adopted by the present invention is as follows:

[0006] According to the technical solution of the present invention, a method for detecting abnormal state of an environmental monitoring sensor includes the following steps:

[0007] Step 1: Perform spatial mutual verification on three sensors evenly distributed in the same space and time;

[0008] Step 2: Perform spatial mutual verification on the three sensors. The three sensors are marked as sensor A, sensor B, and sensor C. Under the same environmental state and no abnormal data, the sum of the data differences between the three sensors should be less than the detection accuracy value ε of the sensor;

[0009] Right now If the spatial mutual verification is normal, the time self-verification will be turned on. Otherwise, if the spatial mutual verification is abnormal, the sensor data will be abnormal.

[0010] Among them, T A is the measurement data of sensor A, T B is the measurement data of sensor B, T C is the measurement data of sensor C;

[0011] Step 3: Perform time self-verification. If the time self-verification is correct, enter the sensor accuracy time and space self-verification. Otherwise, the sensor data is abnormal.

[0012] Step 4: Check the temporal and spatial consistency of sensor accuracy; assume that the working status monitoring sequences of the three sensors in the same time and space are {A i},i∈(1,N),{B i},i∈(1,N),{C i}, i∈(1,N), the monitored value of the temperature at one sampling point is T i The temperature monitoring value T of the previous sampling point i-1 The value between is recorded as ΔT i ,

[0013] If the sensor accuracy is normal,

[0014] In the step 2, when When it is not true, that is, the data of a certain sensor is abnormal, its monitoring data will be greatly different from the monitoring data of the other two sensors, and the measurement threshold T is set. max , the distance between sensor A and sensor B is d AB , the distance between sensor B and sensor C is d BC , the distance between sensor A and sensor C is d AC , d AB and d BC and d AC Approximately equal, the effective measurement range of the sensor is D, the error coefficient is λ;

[0015] like Then the data of A is judged to be abnormal;

[0016] The error coefficient λ is 0.05.

[0017] The method of time self-verification in step 3 is: at a certain time threshold, the nearest neighbor algorithm is used to check the neighbor difference threshold and the measurement value. The nearest neighbor algorithm formula is as follows:

[0018] D ij ={X i -X j ,1<i≤N,1<j≤N}

[0019] Among them, D ijis a data object, i and j represent two time points in the time series, X i represents the temperature value at time point i, X j represents the temperature value at time point j;

[0020] Taking temperature detection as an example, the monitoring value in the time series is, {T i}, i∈(1,N) is the monitored value of the temperature at one of the sampling points T i The temperature monitoring value T of the previous sampling point i-1 The value between is recorded as ΔT i , the temperature difference threshold and time threshold are T max and t max ;

[0021] At the time threshold t max The temperature difference threshold T max and D ij For comparison, if D ij >T max If it is established, the sensor data in this time series is abnormal. Then the ij data between the i-th sampling point and the j-th sampling point are tested for abnormal data points to find ΔT k ,(i<k≤j)greater than T max state point, judge T k Dot and T i-1 and T j+1 If the difference between max , then the sampling data point T k Defined as an outlier data point;

[0022] If D ij >T max If it is not true, then the time self-verification is correct and the sensor accuracy time-space self-verification is entered;

[0023] The range of the time threshold is set according to the sampling frequency of the sensor. If the sampling frequency is high, the threshold is set to 30s, and if the sampling frequency is low, the threshold is set to 30min.

[0024] The nearest neighbor algorithm measures the proximity between data objects, and an abnormal data object refers to a data object with a larger proximity.

[0025] If the accuracy of a sensor in step 4 is abnormal, for example, the accuracy of sensor A is abnormal, the above conditions are not met and

[0026]

[0027] The present invention designs a method for detecting abnormal state of environmental monitoring sensors, which can be applied to data anomaly detection of wired or wireless sensors, and performs detection on two aspects: accuracy and functional anomaly of the sensor. Compared with the existing sensor data anomaly detection method, the present invention introduces time self-verification to perform two-stage verification, and then further determines the data anomaly sampling point; uses space self-verification to determine whether the data anomaly comes from the sensor itself, and verifies the monitoring accuracy of the sensor. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present invention can be further illustrated by the non-limiting examples given in the accompanying drawings;

[0029] Figure 1 The present invention is a flow chart of a method for detecting abnormal state of an environmental monitoring sensor. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0031] A method for detecting abnormal state of an environmental monitoring sensor of the present invention comprises the following steps:

[0032] Step 1: Perform spatial mutual verification on three sensors evenly distributed in the same space and time;

[0033] Step 2: Perform spatial mutual verification on the three sensors. The three sensors are marked as sensor A, sensor B, and sensor C. Under the same environmental state and no abnormal data, the sum of the data differences between the three sensors should be less than the detection accuracy value ε of the sensor;

[0034] Right now If the spatial mutual verification is normal, the time self-verification will be turned on. Otherwise, if the spatial mutual verification is abnormal, the sensor data will be abnormal.

[0035] Among them, T A is the measurement data of sensor A, T B is the measurement data of sensor B, T C is the measurement data of sensor C;

[0036] Step 3: Perform time self-verification. If the time self-verification is correct, enter the sensor accuracy time and space self-verification. Otherwise, the sensor data is abnormal.

[0037] Step 4: Check the temporal and spatial consistency of sensor accuracy; suppose the working status monitoring sequences of the three sensors in the same time and space are {A i},i∈(1,N),{B i},i∈(1,N),{C i}, i∈(1,N), the monitored value of the temperature at one sampling point is T i The temperature monitoring value T of the previous sampling point i-1 The value between is recorded as ΔT i ,

[0038] If the sensor accuracy is normal,

[0039] When using environmental monitoring sensors to monitor environmental factors, more than two sensors are often used, evenly distributed in the spatial range; taking temperature monitoring as an example, the present invention uses three sensors for spatial mutual verification; compared with the existing sensor data anomaly verification method, the present invention introduces time self-verification to perform two-stage verification, and then further determines the data anomaly sampling point; uses spatial self-verification to determine whether the data anomaly comes from the sensor itself, and verifies the monitoring accuracy of the sensor.

[0040] In the step 2, when When it is not true, that is, the data of a certain sensor is abnormal, its monitoring data will be greatly different from the monitoring data of the other two sensors, and the measurement threshold T is set. max , the distance between sensor A and sensor B is d AB , the distance between sensor B and sensor C is d BC , the distance between sensor A and sensor C is d AC , d AB and d BC and d AC Approximately equal, the effective measurement range of the sensor is D, the error coefficient is λ;

[0041] like Then the data of A is judged to be abnormal;

[0042] Sensors always read data periodically. In the process of perceiving the physical environment, discrete points that are continuous in time can be obtained. Therefore, the detection data obtained by a sensor in a certain period of time can be regarded as a data sequence, which is correlated with time and has its own changing rules over time. This is called a time series. Therefore, time self-verification is added to detect sensor data anomalies.

[0043] The error coefficient λ is 0.05.

[0044] The method of time self-verification in step 3 is: at a certain time threshold, the nearest neighbor algorithm is used to check the neighbor difference threshold and the measurement value. The nearest neighbor algorithm formula is as follows:

[0045] D ij ={X i -X j,1<i≤N,1<j≤N}

[0046] Among them, D ij is a data object, i and j represent two time points in the time series, X i represents the temperature value at time point i, X j represents the temperature value at time point j;

[0047] Taking temperature detection as an example, the monitoring value in the time series is, {T i}, i∈(1,N) is the monitored value of the temperature at one of the sampling points T i The temperature monitoring value T of the previous sampling point i-1 The value between is recorded as ΔT i , the temperature difference threshold and time threshold are T max and t max ;

[0048] At the time threshold t max The temperature difference threshold T max and D ij For comparison, if D ij >T max If it is established, the sensor data in this time series is abnormal. Then the ij data between the i-th sampling point and the j-th sampling point are tested for abnormal data points to find ΔT k ,(i<k≤j)greater than T max state point, judge T k Dot and T i-1 and T j+1 If the difference between max , then the sampling data point T k Defined as an outlier data point;

[0049] If D ij >T max If it is not true, then the time self-verification is correct and the sensor accuracy time-space self-verification is entered;

[0050] The range of the time threshold is set according to the sampling frequency of the sensor. If the sampling frequency is high, the threshold is set to 30s, and if the sampling frequency is low, the threshold is set to 30min.

[0051] The nearest neighbor algorithm measures the proximity between data objects. Abnormal data objects refer to data objects with large proximity. The advantage of the algorithm is that it has low time complexity and can be effectively applied to time series anomaly detection.

[0052] If the accuracy of a sensor in step 4 is abnormal, for example, the accuracy of sensor A is abnormal, the above conditions are not met and

[0053]

[0054] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A method for detecting abnormal state of an environmental monitoring sensor, characterized in that: The steps include: Step 1: Perform spatial mutual verification on three sensors evenly distributed in the same space and time; Step 2: Perform spatial mutual verification on the three sensors. The three sensors are marked as sensor A, sensor B, and sensor C. Under the same environmental state and no abnormal data, the sum of the data differences between the three sensors should be less than the detection accuracy value ε of the sensor; Right now If the spatial mutual verification is normal, the time self-verification will be turned on. Otherwise, if the spatial mutual verification is abnormal, the sensor data will be abnormal. Among them, T A is the measurement data of sensor A, T B is the measurement data of sensor B, T C is the measurement data of sensor C; Step 3: Perform time self-verification. If the time self-verification is correct, enter the sensor accuracy time and space self-verification. Otherwise, the sensor data is abnormal. Step 4: Check the temporal and spatial consistency of sensor accuracy; assume that the working status monitoring sequences of the three sensors in the same time and space are {A i },i∈(1,N),{B i },i∈(1,N),{C i }, i∈(1,N), the monitored value of the temperature at one sampling point is T i The temperature monitoring value T of the previous sampling point i-1 The value between is recorded as ΔT i , If the sensor accuracy is normal, In the step 2, when When it is not true, that is, the data of a certain sensor is abnormal, its monitoring data will be greatly different from the monitoring data of the other two sensors, and the measurement threshold T is set. max , the distance between sensor A and sensor B is d AB , the distance between sensor B and sensor C is d BC , the distance between sensor A and sensor C is d AC , d AB and d BC and d AC Approximately equal, the effective measurement range of the sensor is D, the error coefficient is λ; like Then the data of A is judged to be abnormal; The method of time self-verification in step 3 is: at a certain time threshold, the nearest neighbor algorithm is used to check the neighbor difference threshold and the measurement value. The nearest neighbor algorithm formula is as follows: D ij ={X i -X j ,1<i≤N,1<j≤N} Among them, D ij is a data object, i and j represent two time points in the time series, X i represents the temperature value at time point i, X j represents the temperature value at time point j; Taking temperature detection as an example, the monitoring value in the time series is, {T i }, i∈(1,N) is the monitored value of the temperature at one of the sampling points T i The temperature monitoring value T of the previous sampling point i-1 The value between is recorded as ΔT i , the temperature difference threshold and time threshold are T max and t max ; At the time threshold t max The temperature difference threshold T max and D ij For comparison, if D ij >T max If it is established, the sensor data in this time series is abnormal. Then the ij data between the i-th sampling point and the j-th sampling point are tested for abnormal data points to find ΔT k ,(i<k≤j)greater than T max state point, judge T k Dot and T i-1 and T j+1 If the difference between max , then the sampling data point T k Defined as an outlier data point; If D ij >T max If it is not true, then the time self-verification is correct and the sensor accuracy time-space self-verification is entered; If the accuracy of a sensor in step 4 is abnormal, for example, the accuracy of sensor A is abnormal, the above conditions are not met and 2. The method for detecting abnormal state of an environmental monitoring sensor according to claim 1, characterized in that: The error coefficient λ is 0.

05.

3. The method for detecting abnormal state of an environmental monitoring sensor according to claim 1, characterized in that: The range of the time threshold is set according to the sampling frequency of the sensor. If the sampling frequency is high, the threshold is set to 30s, and if the sampling frequency is low, the threshold is set to 30min.

4. The method for detecting abnormal state of an environmental monitoring sensor according to claim 1, characterized in that: The nearest neighbor algorithm measures the proximity between data objects, and an abnormal data object refers to a data object with a larger proximity.