A real-time abnormal data detection and early warning method and system in a sensor network

Abstract

The present invention discloses a real-time abnormal data detection and early warning method and system in a sensor network, which relates to the field of automatic control technology. The method comprises S1, performing precision calibration on an end effector of a mechanical arm; S2, collecting and preprocessing raw data of each sensor node; S3, generating a standard reference data set based on the corrected sensor installation position information; S4, monitoring the data of each sensor node in real time, and comparing it with the standard reference data set; S5, issuing an abnormal data early warning signal in real time according to the comparison result; the real-time abnormal data detection and early warning method and system in a sensor network solve the problems that the end effector of the prior art has insufficient precision, resulting in large deviation in the sensor installation position; the installation deviation affects the data acquisition accuracy of the sensor node, reduces the overall performance of the system; the adaptability to environmental changes is poor, and it is difficult to ensure the reliability and stability of the installation.

Description

Technical Field

[0001] The present invention relates to the field of automatic control technology, and in particular to a real-time abnormal data detection and early warning method and system in a sensor network. Background Art

[0002] In the fields of intelligent manufacturing, automation equipment and industrial Internet, robotic arms are widely used in the installation and maintenance of sensor nodes. Traditional robotic arms complete the positioning and installation of sensor nodes through preset motion trajectories and posture adjustments, and fix the sensors based on the coordination between the end effector of the robotic arm and the target position. However, due to the limitations of the accuracy of the end effector of the robotic arm, large position deviations often occur during the installation of sensor nodes.

[0003] Specifically, when the existing robotic arms perform sensor node installation tasks, the repeatability of the end effector is not accurate enough, making it difficult to meet the strict requirements of high-precision sensor nodes for the installation position. This deviation not only causes an error between the sensor installation position and the design position, but also significantly affects the data acquisition performance and reliability of the sensor. In addition, existing robotic arms usually rely on trajectory planning with fixed parameters, and have poor adaptability to changes in installation position or environmental interference, further exacerbating the installation deviation problem. Therefore, the deficiencies of the prior art are mainly reflected in the following aspects: insufficient accuracy of the end effector, resulting in large deviations in the sensor installation position; installation deviations affect the data acquisition accuracy of the sensor node and reduce the overall performance of the system; poor adaptability to environmental changes makes it difficult to ensure the reliability and stability of the installation. Summary of the invention

[0004] The purpose of the present invention is to provide a real-time abnormal data detection and early warning method and system in a sensor network to solve the problems that the end effector in the prior art is insufficient in precision, resulting in large deviations in the sensor installation position; the installation deviation affects the data collection accuracy of the sensor node and reduces the overall performance of the system; the adaptability to environmental changes is poor, and it is difficult to ensure the reliability and stability of the installation.

[0005] To achieve the above object, the present invention provides the following technical solution: a real-time abnormal data detection and early warning method in a sensor network, the method comprising:

[0006] S1. Perform precision calibration on the end effector of the robot arm, specifically using a high-resolution position sensor to measure the initial position and speed of the robot arm, substituting the motion parameters of the robot arm into the trajectory prediction formula, calculating the end trajectory, comparing the actual trajectory with the predicted trajectory, calculating the deviation and applying the correction factor to update the motion control model;

[0007] S2, collect the raw data of each sensor node and perform preprocessing;

[0008] S3, generating a standard reference data set based on the corrected sensor installation position information;

[0009] S4. Monitor the data of each sensor node in real time and compare it with the standard reference data set. Specifically, the real-time sensor data and the standard reference data set are modeled as a dynamic competitive relationship, and the interaction dynamics between the two are calculated to observe whether abnormal growth or decline occurs. If any variable deviates from the equilibrium point, it is judged as abnormal. The specific formula is:

[0010] , ;

[0011] in, and Indicates the dynamic state of real-time data and reference data, and represents the growth rate, and Indicates the carrying capacity, and represents the mutual influence coefficient;

[0012] S5. Based on the comparison results, an abnormal data warning signal is issued in real time.

[0013] Preferably, S2 includes:

[0014] The raw data collected by the sensor node is fast Fourier transformed to obtain the frequency domain signal, the power spectrum density is calculated, the noise frequency component is identified and filtered out, and the noise-reduced signal is converted back to the time domain. The specific formula is:

[0015] ;

[0016] in, Indicates that the signal is at a certain frequency The power distribution on Indicates frequency, represents the number of data points, Indicates the index number.

[0017] Preferably, S3 includes:

[0018] The symmetry model is established by using the corrected position information of the sensor. Through the symmetry operation, the position information of the sensor is mapped to a standard reference coordinate system to generate a standard reference data set that conforms to the geometric symmetry and is used as a benchmark for anomaly detection. The specific formula is:

[0019] ;

[0020] in, represents the position after conversion, represents the rotation matrix, represents the original sensor position, Represents the translation vector.

[0021] Preferably, S5 includes:

[0022] The sensor data is converted into a spectrum, the main frequency components are extracted, and the deviation between the real-time data frequency and the reference frequency is calculated. If the deviation exceeds the set threshold, a change signal is generated as an early warning prompt. The specific formula is:

[0023] ;

[0024] in, Indicates the frequency deviation, Indicates the real-time data frequency, Indicates the reference data frequency.

[0025] Preferably, S3 includes:

[0026] The correction value of the end effector of the robot arm is obtained through the high-precision calibration module;

[0027] Adjust the installation position of each sensor node according to the correction value;

[0028] Collect initial data with the adjusted installation position;

[0029] The standard deviation of each sensor is calculated based on the initial data and a standard reference data set is generated.

[0030] Preferably, the calculating the standard deviation of each sensor based on the initial data comprises:

[0031] Collect sensor data over multiple different time periods;

[0032] Calculate the average value μ of the data in each time period;

[0033] Calculate the standard deviation σ of the data in each time period;

[0034] The standard deviation is calculated as: ;

[0035] Among them, x i represents the data value of the i-th time period, N represents the number of time periods, μ represents the mean, and σ represents the standard deviation.

[0036] Preferably, calculating the average value of data in each time period includes:

[0037] Initialize the sum S to 0;

[0038] Traverse the data of all time periods and accumulate the data values ​​to the total S;

[0039] Based on the total S and the number of time periods N, calculate the average μ=S / N;

[0040] If S=0 or N≤1, the average value μ is set to 0.

[0041] Preferably, the calculation deviation includes:

[0042] Collect actual trajectory data and compare it with the predicted trajectory;

[0043] Based on position error |r(t) 实际 −r(t) 预测 ∣Calculation deviation;

[0044] If the deviation exceeds a preset threshold, a corrective action is triggered.

[0045] Preferably, the specific formula for calculating the terminal trajectory in S1 is:

[0046] ;

[0047] in, represents the real-time position of the end effector, represents the initial position, represents the initial velocity, represents acceleration, Indicates time.

[0048] A real-time abnormal data detection and early warning system in a sensor network, using the real-time abnormal data detection and early warning method in a sensor network, the system comprises:

[0049] High-precision calibration module, used to calibrate the end effector of the robot arm;

[0050] Data acquisition module, used to collect raw data from each sensor node and perform preprocessing;

[0051] A reference data generation module, used to generate a standard reference data set based on the corrected sensor installation position information;

[0052] Real-time monitoring module, used to monitor the data of each sensor node in real time and compare it with the standard reference data set;

[0053] The early warning signal module is used to issue abnormal data early warning signals in real time according to the comparison results.

[0054] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0055] The real-time abnormal data detection and early warning method and system in the sensor network, by precision correction of the end effector of the mechanical arm, collects and preprocesses the original data of each sensor node, generates a standard reference data set based on the corrected sensor installation position information, monitors the data of each sensor node in real time, and compares it with the standard reference data set, and issues an abnormal data early warning signal in real time according to the comparison result, significantly reduces the position deviation of the end effector of the mechanical arm during the installation of the sensor node, ensures the high consistency between the sensor installation position and the design position, improves the installation accuracy of the sensor node, reduces the influence of the position error on the data acquisition performance, ensures the accuracy and consistency of the collected data, improves the overall work efficiency and stability, improves the flexibility and reliability of the sensor installation and maintenance process, and significantly optimizes the overall performance of the sensor network, provides a stable foundation for subsequent abnormal data detection and early warning, reduces the operation and maintenance costs, and solves the problems that the end effector of the prior art is insufficient in precision, resulting in large deviation in the sensor installation position; the installation deviation affects the data acquisition accuracy of the sensor node and reduces the overall performance of the system; the adaptability to environmental changes is poor, and it is difficult to ensure the reliability and stability of the installation. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a flow chart of the method of the present invention;

[0057] Figure 2 It is a schematic diagram of the connection of the system modules of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] like Figure 1 As shown, the present invention provides a technical solution: a real-time abnormal data detection and early warning method in a sensor network, the method comprising:

[0060] S1. Perform precision calibration on the end effector of the robot arm. Specifically, use a high-resolution position sensor to measure the initial position and speed of the robot arm, bring the motion parameters of the robot arm into the trajectory prediction formula, calculate the end trajectory, compare the actual trajectory with the predicted trajectory, calculate the deviation and apply the correction factor to update the motion control model. The end trajectory calculation formula is:

[0061] ;

[0062] in, represents the real-time position of the end effector, represents the initial position, represents the initial velocity, represents acceleration, Indicates time;

[0063] S2, collect the raw data of each sensor node and perform preprocessing;

[0064] S3, generating a standard reference data set based on the corrected sensor installation position information;

[0065] S4. Monitor the data of each sensor node in real time and compare it with the standard reference data set. Specifically, the real-time sensor data and the standard reference data set are modeled as a dynamic competitive relationship, and the interaction dynamics between the two are calculated to observe whether abnormal growth or decline occurs. If any variable deviates from the equilibrium point, it is judged as abnormal. The specific formula is:

[0066] , ;

[0067] in, and Indicates the dynamic state of real-time data and reference data, and represents the growth rate, and Indicates the carrying capacity, and represents the mutual influence coefficient;

[0068] S5. Based on the comparison results, an abnormal data warning signal is issued in real time.

[0069] The above method realizes dynamic monitoring and adjustment of the robot arm's running trajectory through high-precision measurement and trajectory prediction of the robot arm's end effector position. On the basis of collecting sensor node data, a standard reference data set is generated, and the real-time data is compared and analyzed in combination with a dynamic competition model to effectively judge the abnormal state of the data. Among them, the dynamic model formula captures the growth trend and dynamic changes of the data by considering the competition and mutual influence relationship between the data, ensuring that the abnormal phenomenon of deviation from the equilibrium point is discovered in time. Furthermore, by issuing early warning signals, potential problems can be responded to quickly, thereby realizing effective monitoring of the sensor network. Through the trajectory prediction and correction method, the operation accuracy of the robot arm's end effector is improved, and the trajectory deviation problem caused by cumulative errors is avoided. The abnormal changes of data can be quickly captured, and the real-time detection and early warning of sensor network data can be realized. By introducing the carrying capacity and mutual influence coefficient, the data modeling is ensured to be closer to the actual operating environment and the misjudgment rate is reduced. The real-time early warning signal can effectively reduce the risk of system failure caused by data anomalies. The method can be adapted to a variety of sensor networks and data types and is suitable for different industrial scenarios.

[0070] S2 includes:

[0071] The raw data collected by the sensor node is fast Fourier transformed to obtain the frequency domain signal, the power spectrum density is calculated, the noise frequency component is identified and filtered out, and the noise-reduced signal is converted back to the time domain. The specific formula is:

[0072] ;

[0073] in, Indicates that the signal is at a certain frequency The power distribution on Indicates frequency, represents the number of data points, Indicates the index number.

[0074] This implementation method converts the signal from the time domain to the frequency domain by applying the fast Fourier transform (FFT) to the raw data collected by the sensor node, and obtains the power spectrum density distribution corresponding to each frequency component. The power spectrum density can reveal the energy distribution characteristics of the signal and help identify abnormal frequency components. The identification and filtering of noise frequency components are based on the fact that their power characteristics are significantly lower than the characteristic frequency of the signal and they show randomness. After noise reduction processing, the signal is restored to the time domain through the inverse Fourier transform, so as to obtain higher quality signal data, which provides a basis for subsequent abnormal detection and early warning. The frequency domain analysis and noise filtering are realized by the fast Fourier transform, which effectively improves the signal-to-noise ratio of the sensor data. The power spectrum density analysis can accurately identify the abnormal components of a specific frequency, which helps to improve the sensitivity and accuracy of abnormal data detection. The fast Fourier transform algorithm is used to greatly reduce the computational complexity of data processing, which is suitable for real-time processing scenarios. This method can be applied to a variety of sensor signal types, including but not limited to vibration signals, temperature signals and acoustic signals. By filtering out the noise frequency components, the anti-interference ability of the sensor network is improved, and the data reliability is ensured.

[0075] S3 includes:

[0076] The symmetry model is established by using the corrected position information of the sensor. Through the symmetry operation, the position information of the sensor is mapped to a standard reference coordinate system to generate a standard reference data set that conforms to the geometric symmetry and is used as a benchmark for anomaly detection. The specific formula is:

[0077] ;

[0078] in, represents the position after conversion, represents the rotation matrix, represents the original sensor position, Represents the translation vector.

[0079] This implementation method converts the position information of the sensor after correction into data points in the standard reference coordinate system by establishing a symmetry model, ensuring that the mapping of different sensor positions has consistency and geometric symmetry. The rotation matrix R is used to realize the rotation transformation of the sensor position, and the translation vector Used for calibration compensation of position translation. Through this model, the interference of sensor installation errors or environmental changes on position information can be eliminated, thereby generating a standard reference data set that conforms to geometric symmetry. This reference data set provides a stable and reliable benchmark for subsequent anomaly detection. Through the geometric correction of position information, the position information deviation caused by sensor installation deviation or environmental factors is eliminated. The standardized reference data set ensures the accuracy of the benchmark data during anomaly detection, thereby improving the credibility of the detection results. The symmetry model is suitable for sensor networks with different geometric distributions and can be flexibly adapted to various scenarios. The geometric standardization of sensor data helps to improve the adaptability and stability of the system in complex environments. Through precise symmetry modeling and coordinate mapping, the false alarm of anomaly detection caused by position information errors is reduced.

[0080] S5 includes:

[0081] The sensor data is converted into a spectrum, the main frequency components are extracted, and the deviation between the real-time data frequency and the reference frequency is calculated. If the deviation exceeds the set threshold, a change signal is generated as an early warning prompt. The specific formula is:

[0082] ;

[0083] in, Indicates the frequency deviation, Indicates the real-time data frequency, Indicates the reference data frequency.

[0084] This embodiment converts the real-time data collected by the sensor into frequency domain, extracts the main frequency component of the data (such as peak frequency), and compares it with the preset reference frequency. By calculating the absolute difference between the two, the system generates a change signal as an abnormal warning when the deviation exceeds the set threshold range. This method can capture abnormal changes in data frequency in real time, prompt possible faults or abnormal behaviors, and provide users with timely warning information. Through frequency deviation calculation, the abnormal frequency characteristics hidden in the data can be quickly discovered, and efficient real-time anomaly detection can be achieved. The main frequency components are used as the basis for data analysis to effectively avoid the interference of noise on the abnormal detection results. By setting the deviation threshold, the warning sensitivity can be flexibly adjusted to adapt to different application scenarios and data characteristics. This method can be widely used in mechanical vibration monitoring, acoustic signal analysis, and other sensor data processing with obvious frequency characteristics. The analysis of frequency characteristics avoids the misjudgment that may occur when relying solely on time domain data, and improves the reliability of system anomaly detection.

[0085] S3 includes obtaining the correction value of the end effector of the robot arm through a high-precision calibration module; adjusting the installation position of each sensor node according to the correction value; collecting initial data through the adjusted installation position; calculating the standard deviation of each sensor based on the initial data, and generating a standard reference data set.

[0086] In this embodiment, the correction value of the end effector of the robot arm is obtained through a high-precision calibration module to ensure the accuracy of the position and motion trajectory of the robot arm. The position of the sensor node is readjusted according to the correction value to reduce the data deviation caused by installation errors. Using the initial data collected at the adjusted position, by calculating statistical parameters such as standard deviation, a standard reference data set reflecting the data characteristics of the sensor node is generated to provide a reliable benchmark for subsequent anomaly detection. Through high-precision calibration and adjustment, the accuracy and consistency of sensor data are improved, and the generated standard reference data set has higher reliability and can be effectively used for benchmark comparison of anomaly detection. The method can adapt to different sensor network configurations to meet the needs of multiple scenarios.

[0087] Calculating the standard deviation of each sensor based on the initial data includes collecting sensor data of multiple different time periods; calculating the average value μ of the data in each time period; calculating the standard deviation σ of the data in each time period;

[0088] The standard deviation is calculated as: ;

[0089] Among them, x i represents the data value of the i-th time period, N represents the number of time periods, μ represents the mean, and σ represents the standard deviation.

[0090] By collecting and analyzing data from different time periods, the data fluctuation characteristics are quantified using the calculation method of the mean value and standard deviation. The calculation of the standard deviation σ reflects the degree of data dispersion, which can provide a statistical basis for generating a standard reference data set and ensure the stability and accuracy of the data benchmark. The standard deviation is used to characterize data fluctuations and provide a reliable indicator for anomaly detection. The method can effectively eliminate the impact of short-term fluctuations and improve the stability of benchmark data. The standard deviation calculation process is fast and can be applied to real-time data processing.

[0091] Calculating the average value of data in each time period includes initializing the total S to 0; traversing the data of all time periods and accumulating each data value to the total S; calculating the average value μ=S / N based on the total S and the number of time periods N; judging that if S=0 or N≤1, the average value μ is set to 0.

[0092] This implementation method calculates the average value by traversing the data of each time period and accumulating the sum, and combining the number of time periods. The judgment logic for abnormal conditions (such as missing data or insufficient quantity) is added to ensure the rationality and stability of the average value calculation result. By judging abnormal conditions, calculation errors caused by missing data or invalid input are avoided, and the calculation process is efficient based on simple addition and division operations.

[0093] Calculating the deviation involves collecting actual trajectory data and comparing it with the predicted trajectory; based on the position error | r(t) 实际 −r(t) 预测 ∣Calculate the deviation; if the deviation exceeds the preset threshold, the correction action is triggered.

[0094] By comparing the actual trajectory data with the predicted trajectory, the deviation between the two is quantified. Combined with the preset threshold, the abnormal situation of the trajectory is judged in real time and the correction operation is triggered to ensure the accuracy of the robot arm trajectory.

[0095] The correction operation includes adjusting the driving parameters of the robot arm. When the position deviation exceeds the threshold, the motion trajectory is dynamically corrected by adjusting the driving parameters of the robot arm (such as speed and acceleration) to eliminate the deviation.

[0096] A real-time abnormal data detection and early warning system in a sensor network is also provided, which adopts a real-time abnormal data detection and early warning method in a sensor network, and the system comprises:

[0097] High-precision calibration module, used to calibrate the end effector of the robot arm;

[0098] Data acquisition module, used to collect raw data from each sensor node and perform preprocessing;

[0099] A reference data generation module, used to generate a standard reference data set based on the corrected sensor installation position information;

[0100] Real-time monitoring module, used to monitor the data of each sensor node in real time and compare it with the standard reference data set;

[0101] The early warning signal module is used to issue abnormal data early warning signals in real time according to the comparison results.

[0102] High-precision calibration module: During the system initialization phase, this module calibrates the position of the robot end effector through high-precision positioning equipment (such as laser rangefinder or inertial measurement unit) and records its initial position and motion parameters. The calibration value serves as a reference for subsequent sensor installation and adjustment to ensure the relative position of the sensor and the robot arm is accurate and reduce signal deviation caused by installation errors.

[0103] Data acquisition module: This module collects raw data from sensor nodes in real time and performs denoising and format standardization operations locally. The pre-processed data ensures the consistency of data quality of the input reference data generation module and the real-time monitoring module, effectively avoiding misjudgments caused by signal noise or irregular data formats.

[0104] Reference data generation module: Generates a standard reference data set based on the installation location information after high-precision calibration. The generation of reference data extracts the characteristic components of the sensor through statistical analysis (such as mean, standard deviation, power spectrum analysis) and establishes a dynamic data model. This module ensures that the system has a stable and reliable benchmark for comparison and anomaly detection of real-time monitoring modules.

[0105] Real-time monitoring module: compares and analyzes the data collected by sensor nodes in real time with the standard reference data set to capture abnormal feature points. It can quickly respond to subtle abnormal changes in data.

[0106] Early warning signal module: Based on the comparison results of the real-time monitoring module, this module determines the severity of data anomalies and triggers early warning signals. If the data deviation exceeds the preset threshold, the early warning signal module notifies the user through various communication means (such as SMS, email, alarm sound). At the same time, the module automatically records the log information of abnormal events, including the type of abnormality, the time of occurrence and the relevant sensor nodes, to provide data support for subsequent fault diagnosis.

[0107] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A real-time abnormal data detection and early warning method in a sensor network, characterized in that: The method comprises: S1. Perform precision calibration on the end effector of the robot arm, specifically using a high-resolution position sensor to measure the initial position and speed of the robot arm, substituting the motion parameters of the robot arm into the trajectory prediction formula, calculating the end trajectory, comparing the actual trajectory with the predicted trajectory, calculating the deviation and applying the correction factor to update the motion control model; S2, collect the raw data of each sensor node and perform preprocessing; S3, generating a standard reference data set based on the corrected sensor installation position information; S4. Monitor the data of each sensor node in real time and compare it with the standard reference data set. Specifically, the real-time sensor data and the standard reference data set are modeled as a dynamic competitive relationship, and the interaction dynamics between the two are calculated to observe whether abnormal growth or decline occurs. If any variable deviates from the equilibrium point, it is judged as abnormal. The specific formula is: ， ； in, and Indicates the dynamic state of real-time data and reference data, and represents the growth rate, and Indicates the carrying capacity, and represents the mutual influence coefficient; S5. Based on the comparison results, an abnormal data warning signal is issued in real time.

2. The method for real-time abnormal data detection and early warning in a sensor network according to claim 1, characterized in that: The S2 includes: The raw data collected by the sensor node is fast Fourier transformed to obtain the frequency domain signal, the power spectrum density is calculated, the noise frequency component is identified and filtered out, and the noise-reduced signal is converted back to the time domain. The specific formula is: ； in, Indicates that the signal is at a certain frequency The power distribution on Indicates frequency, represents the number of data points, Indicates the index number.

3. The real-time abnormal data detection and early warning method in a sensor network according to claim 1, characterized in that: The S3 includes: The symmetry model is established by using the corrected position information of the sensor. Through the symmetry operation, the position information of the sensor is mapped to a standard reference coordinate system to generate a standard reference data set that conforms to the geometric symmetry and is used as a benchmark for anomaly detection. The specific formula is: ； in, represents the position after conversion, represents the rotation matrix, represents the original sensor position, Represents the translation vector.

4. The real-time abnormal data detection and early warning method in a sensor network according to claim 1, characterized in that: The S5 includes: The sensor data is converted into a spectrum, the main frequency components are extracted, and the deviation between the real-time data frequency and the reference frequency is calculated. If the deviation exceeds the set threshold, a change signal is generated as an early warning prompt. The specific formula is: ； in, Indicates the frequency deviation, Indicates the real-time data frequency, Indicates the reference data frequency.

5. The real-time abnormal data detection and early warning method in a sensor network according to claim 1, characterized in that: The S3 includes: The correction value of the end effector of the robot arm is obtained through the high-precision calibration module; Adjust the installation position of each sensor node according to the correction value; Collect initial data with the adjusted installation position; The standard deviation of each sensor is calculated based on the initial data and a standard reference data set is generated.

6. The real-time abnormal data detection and early warning method in a sensor network according to claim 5, characterized in that: Calculating the standard deviation of each sensor based on the initial data includes: Collect sensor data over multiple different time periods; Calculate the average value μ of the data in each time period; Calculate the standard deviation σ of the data in each time period; The standard deviation is calculated as: ; Among them, x i represents the data value of the i-th time period, N represents the number of time periods, μ represents the mean, and σ represents the standard deviation.

7. The real-time abnormal data detection and early warning method in a sensor network according to claim 6, characterized in that: Calculating the average value of data in each time period includes: Initialize the sum S to 0; Traverse the data of all time periods and accumulate the data values ​​to the total S; Based on the total S and the number of time periods N, calculate the average μ=S / N; If S=0 or N≤1, the average value μ is set to 0.

8. The real-time abnormal data detection and early warning method in a sensor network according to claim 1, characterized in that: The calculated deviations include: Collect actual trajectory data and compare it with the predicted trajectory; Based on position error |r(t) 实际 −r(t) 预测 ∣Calculation deviation; If the deviation exceeds a preset threshold, a corrective action is triggered.

9. The real-time abnormal data detection and early warning method in a sensor network according to claim 1, characterized in that: The specific formula for calculating the terminal trajectory in S1 is: ； in, represents the real-time position of the end effector, represents the initial position, represents the initial velocity, represents acceleration, Indicates time.

10. A real-time abnormal data detection and early warning system in a sensor network, using the real-time abnormal data detection and early warning method in a sensor network according to any one of claims 1 to 9, characterized in that: The system comprises: High-precision calibration module, used to calibrate the end effector of the robot arm; Data acquisition module, used to collect raw data from each sensor node and perform preprocessing; A reference data generation module, used to generate a standard reference data set based on the corrected sensor installation position information; Real-time monitoring module, used to monitor the data of each sensor node in real time and compare it with the standard reference data set; The early warning signal module is used to issue abnormal data early warning signals in real time according to the comparison results.

Images