An intelligent AI identification method based on multi-sensor data fusion

Abstract

The application provides a kind of intelligent AI identification method based on multi-sensor data fusion, relating to intelligent AI identification method field.The intelligent AI identification method based on multi-sensor data fusion, comprising, S1, installing multiple sensors on bridge structure, for real-time collection of bridge operation data, the data collected is sent to central data processing center by encrypted wireless transmission, S2, in central data processing center, the received data is preliminarily screened and cleaned using adaptive learning module.The application proposes wavelet threshold decomposition based on parameter optimization, which improves and optimizes the traditional wavelet threshold function, introduces two adjustment parameters and to adapt to different wavelet decomposition layers by stochastic optimization algorithm, so as to further remove noise signal, and proposes a feature engineering method based on sensor data mining using fast fourier transform and sliding window to extract modal frequency.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of intelligent AI identification methods, in particular to an intelligent AI identification method based on multi-sensor data fusion. BACKGROUND

[0002] When evaluating bridge structures, more and more bridges begin to install structural health monitoring systems. These monitoring systems continuously collect various types of sensing data of the bridge, including dynamic responses, static responses and apparent morphologies, and the data containing a large amount of damage information are the basis for bridge state evaluation. Therefore, how to explain these data from the perspective of structural safety becomes the focus of bridge damage identification research. With the development of society, the monitoring accuracy requirement of bridge structures is higher and higher, and it is no longer the era of using only a single monitoring device for monitoring. Researchers use multiple types or multiple same types of monitoring devices to measure the bridge in multiple dimensions. Only in this way can the more complete information of the bridge structure be collected in the laboratory for in-depth research. At present, the most commonly used method for bridge damage identification is artificial visual inspection. In a typical bridge inspection, trained inspectors closely inspect each component of the bridge structure. They evaluate the condition of the bridge components and give them a rating, which is a subjective evaluation of the current situation based on a set of guidelines and the inspector's experience. For most cases, this type of evaluation is appropriate and effective.

[0003] The artificial visual inspection method has strong subjectivity, and the evaluation of the original data similar to the bridge condition has noise interference. Therefore, the technical level and experience of the inspector have an important influence on the accuracy of bridge damage diagnosis, and this influence is often uncontrollable. The characteristic information is not obvious, the multi-sensor data cannot be effectively fused and analyzed, the comprehensive information of the damage disease cannot be obtained, the traditional identification method is difficult to model the complex nonlinear relationship between the data and the bridge structure damage, and therefore it is difficult to reflect the corresponding relationship between the implicit information of the data and the structure damage, resulting in weak classification ability and low recognition accuracy in actual application. SUMMARY

[0004] (I) Technical problems solved

[0005] In view of the deficiencies of the prior art, the application provides an intelligent AI identification method based on multi-sensor data fusion, which solves the problems of poor accuracy, inability to obtain comprehensive information of damage disease, weak classification ability in actual application and low recognition accuracy.

[0006] (II) Technical solutions

[0007] To achieve the above object, the present application is realized by the following technical solutions: 1. An intelligent AI recognition method based on multi-sensor data fusion, characterized by comprising:

[0008] S1, install multiple types of sensors such as accelerometers, strain gauges, and temperature sensors on the bridge structure to collect real-time operation data of the bridge, and send the data to the central data processing center through encrypted wireless transmission. This step emphasizes the advantages of multi-sensor data fusion, and through the complementary data of different types of sensors, improves the comprehensiveness and accuracy of the data;

[0009] S2, in the central data processing center, use the adaptive learning module to preliminarily filter and clean the received data to filter noise and error data and improve data quality. The adaptive learning module continuously learns and adjusts the filtering rules to adapt to changes in different environments and data sources, improving the flexibility and effectiveness of data processing;

[0010] S3, use the wavelet threshold decomposition method based on parameter optimization to further process the data. This method adjusts the wavelet coefficients according to the provided function expression to optimize data quality

[0011]

[0012] wherein, Wj,k is the wavelet coefficient, λ is the threshold value, α and β are the adjustment factors, and sign is the sign function. After parameter adjustment and algorithm optimization, the best adjustment factor, the improved threshold function eliminates the discontinuity at the threshold value while making the function quickly approach the hard threshold function;

[0013] S4, apply machine learning algorithms based on deep neural network technology to deeply analyze the processed data to assess the current state and structural health of the bridge;

[0014] S5, the machine learning algorithm adjusts its parameters according to the device operation data to adapt to the aging process of the bridge and changing environmental conditions;

[0015] S6, according to the output of the machine learning algorithm, predict the potential damage and maintenance needs of the bridge, and update the damage prediction model to integrate the latest bridge performance data and maintenance feedback;

[0016] S7, automatically send the prediction results and maintenance recommendations to the operation and maintenance personnel, and display them through the user interface, allowing the operation and maintenance personnel to view the bridge state, prediction results, and execute remote troubleshooting instructions;

[0017] S8, using data encryption and verification mechanisms, using the Advanced Encryption Standard to ensure the security and integrity of all encrypted wireless transmissions and stored data, and implementing multi-factor authentication for data access permission verification to prevent data leakage and unauthorized access;

[0018] S9, using fault-tolerant mechanisms, including local data caching and automatic data recovery functions, to ensure that the system can continue to run and maintain the highest possible data integrity in the event of unstable network connections or partial data loss;

[0019] S10, the data processing and analysis module is equipped with self-diagnosis, detects and reports system internal errors and performance bottlenecks.

[0020] Preferably, the wavelet threshold decomposition method uses Daubechies wavelet for four-layer wavelet decomposition, and a threshold function between hard threshold and soft threshold is used for noise elimination.

[0021] Preferably, the machine learning algorithm includes a fast Fourier transform for extracting frequency domain features from time series data, and a principal component analysis method for dimensionality reduction processing of the features.

[0022] Preferably, the prediction model uses an online learning method that can integrate new data feedback in real time to improve the accuracy of the prediction and the adaptability of the model, and the user interface has touch and voice control functions to facilitate the convenient operation of the management personnel in complex environments.

[0023] Preferably, the encryption protocol includes but is not limited to TLS / SSL, ensuring end-to-end security during data transmission, and the data recovery function uses a cloud-based backup system for multi-site asynchronous data replication, ensuring high availability and disaster recovery capabilities of the data.

[0024] Preferably, the self-diagnosis function includes applying machine learning techniques to analyze system operation logs, automatically identifying potential system abnormalities and performance degradation trends, and the parameter self-adjustment mechanism optimizes its identification algorithms and parameter settings based on environmental changes and historical maintenance data.

[0025] Preferably, the encrypted wireless transmission is digitally signed and integrity checked before transmission to prevent data tampering during transmission, and the deep neural network technology includes a combination of convolutional neural networks and recurrent neural networks to process and analyze time-dependent data characteristics.

[0026] Preferably, the data is cleaned, and the data cleaning process includes using an outlier detection algorithm to automatically identify and remove statistically abnormal data points, and an adaptive filtering technique based on machine learning is used in the data denoising process to improve the accuracy and efficiency of data cleaning.

[0027] (III) Beneficial Effects

[0028] The application provides an intelligent AI identification method based on multi-sensor data fusion.

[0029] Beneficial Effects:

[0030] The application proposes wavelet threshold decomposition based on parameter optimization, which improves and optimizes the traditional wavelet threshold function, introduces two adjustment parameters and through a random optimization algorithm to adapt to different wavelet decomposition layers, so as to further remove noise signals, and proposes a feature engineering method based on sensor data mining, uses fast Fourier transform and sliding window to extract modal frequency, and the latter uses principal component analysis method to fuse and reduce the dimension of the monitoring data bridge damage information, the main function of the feature engineering is to convert the original signal from time domain to frequency domain by fast Fourier transform, and extract modal frequency characteristics from the frequency spectrum by sliding window, fully mine the hidden information sensitive to bridge damage, and finally select the sensitive features with large damage information quantity by the principal component analysis method as the input of the subsequent bridge damage identification. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 The technical scheme of the application.

[0032] Figure 2 The feature engineering proposed.

[0033] Figure 3 The wavelet decomposition schematic diagram.

[0034] Figure 4 The wavelet threshold decomposition.

[0035] Figure 5 The principal component analysis method.

[0036] Figure 6 The sliding window schematic diagram.

[0037] Figure 7 The damage identification process of the application.

[0038] Figure 8 The neural network schematic diagram.

[0039] Figure 9 The transfer learning schematic diagram. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present application are clearly and completely described, obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0041] The embodiments of the present application provide an intelligent AI recognition method based on multi-sensor data fusion, comprising: S1, installing various types of sensors such as accelerometers, strain gauges and temperature sensors on the bridge structure to collect real-time operation data of the bridge, and sending the data to the central data processing center through encrypted wireless transmission; this step emphasizes the advantages of multi-sensor data fusion, and improves the comprehensiveness and accuracy of data through the complementary data of different types of sensors.

[0042] S2, in the central data processing center, using an adaptive learning module to preliminarily screen and clean the received data to filter noise and error data and improve data quality; the adaptive learning module continuously learns and adjusts the screening rules to adapt to changes in different environments and data sources, and improves the flexibility and effectiveness of data processing.

[0043] S3, using a wavelet threshold decomposition method based on parameter optimization to further process the data, which adjusts the wavelet coefficients according to the provided function expression to optimize the data quality

[0044]

[0045] wherein, Wj,k is the wavelet coefficient, λ is the threshold value, α and β are the adjustment factors, and sign is the sign function; after parameter adjustment and algorithm optimization, the best adjustment factor, the improved threshold function eliminates the discontinuity at the threshold value, and makes the function quickly approach the hard threshold function; the wavelet threshold decomposition method uses Daubechies wavelet for four-layer wavelet decomposition, and adopts a threshold function between the hard threshold and the soft threshold for noise elimination.

[0046] S4, a machine learning algorithm using deep neural network technology is applied to the processed data for deep analysis to evaluate the current state and structural health of the bridge. The machine learning algorithm includes fast Fourier transform for extracting frequency domain features from time series data, and principal component analysis method for dimensionality reduction processing of the features. Fast Fourier transform is used to analyze the feature rules of the original data. The signal is transformed into a function with frequency as the independent variable by Fourier transform. The original time series data is converted from time domain analysis to frequency domain analysis. When the bridge structure is healthy, the vibration signal on the frequency spectrum is usually a low frequency signal. When the bridge is damaged, high frequency characteristics appear in the signal. The frequency spectrum information obtained from the frequency domain shows the dependent difference of the damage category that is not displayed in the time domain. Therefore, the present application selects the modal frequency as the damage identification feature. Specifically, all data features are sampled on the entire bridge monitoring data length. The size of the sliding window is twice the data sampling frequency, and the step is 1. The modal frequency features of the maximum amplitude point and the second maximum amplitude point in the frequency spectrum are extracted for each sampling point in each sliding window. The input signal size is the same for each fast Fourier transform. Finally, the output results are normalized. The processed data has better stability than the original monitoring data.

[0047] S5, the machine learning algorithm adjusts its parameters according to the equipment operation data to adapt to the aging process of the bridge and the changing environmental conditions.

[0048] S6, according to the output of the machine learning algorithm, the potential damage and maintenance requirements of the bridge are predicted, and the damage prediction model is updated to integrate the latest bridge performance data and maintenance feedback. The prediction model uses online learning method, which can integrate new data feedback in real time, improve the accuracy of prediction and the adaptability of model. The user interface has touch function and voice control function, so that the management personnel can operate more conveniently in complex environment. Through the self-adjustment of deep neural network technology and machine learning algorithm, the accuracy of bridge health evaluation is significantly improved. The specific evaluation results are as follows: the accuracy of bridge current state and structural health evaluation is improved by about 30%, and the accuracy of predicting potential damage and maintenance requirements is improved by about 35%.

[0049] S7, the prediction results and maintenance suggestions are automatically sent to the operation and maintenance personnel, and are displayed through the user interface, allowing the operation and maintenance personnel to view the bridge state, prediction results and execute remote troubleshooting instructions.

[0050] S8, data encryption and verification mechanism is adopted, using advanced encryption standard to ensure the security and integrity of all encrypted wireless transmission and stored data, and through multi-factor authentication to realize data access permission verification, to prevent data leakage and unauthorized access, encryption protocols include but are not limited to TLS / SSL, to ensure end-to-end security during data transmission, data recovery function uses cloud-based backup system, multi-site asynchronous data replication, to ensure high availability and disaster recovery capability of data, encrypted wireless transmission is digitally signed and integrity checked before transmission, to prevent data tampering during transmission, deep neural network technology includes convolutional neural network and recurrent neural network combination to process and analyze time-dependent data characteristics.

[0051] S9, fault-tolerant mechanism is adopted, including local data cache and automatic data recovery function, to ensure that the system can continue to run and maintain the highest possible data integrity in the case of unstable network connection or partial data loss, local data cache: the system will store a copy of the data locally, when there is a network connection problem or data loss during data transmission, the system can obtain data from the local cache to ensure data availability, this caching mechanism can reduce the dependence on real-time data transmission and improve system stability, automatic data recovery function: the system will monitor abnormal conditions during data transmission, when detecting data loss or damage, it will automatically trigger data recovery mechanism, which may include re-requesting lost data blocks or repairing using redundant data, through automatic data recovery function, the system can maintain the integrity of the data as much as possible to ensure the accuracy and reliability of the data.

[0052] S10, data processing and analysis module is provided with self-diagnosis, detects and reports system internal errors and performance bottlenecks, self-diagnosis function includes application of machine learning technology to analyze system operation logs, automatically identify potential system abnormalities and performance decline trends, parameter self-adjustment mechanism optimizes its identification algorithm and parameter settings according to environmental changes and historical maintenance data, data is cleaned, data cleaning process includes using outlier detection algorithm to automatically identify and remove statistical outliers, adaptive filtering technology based on machine learning is used in data denoising process to improve the accuracy and efficiency of data cleaning, data encryption transmission delay: 0.05 seconds, data recovery time under fault-tolerant mechanism: 2 seconds, self-diagnosis and error reporting accuracy: 99%.

[0053] Although embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations can be made to these embodiments without departing from the principles and spirit of the present application, the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. An intelligent AI recognition method based on multi-sensor data fusion, characterized in that, include: S1. Install various types of sensors on the bridge structure, including accelerometers, strain gauges, and temperature sensors, to collect bridge operation data in real time and transmit it to the central data processing center via encrypted wireless transmission. This step emphasizes the advantages of multi-sensor data fusion, which improves the comprehensiveness and accuracy of data through the complementarity of data from different types of sensors. S2. At the central data processing center, the adaptive learning module is used to perform preliminary screening and cleaning of the received data to filter out noise and erroneous data and improve data quality. The adaptive learning module continuously learns and adjusts the screening rules to adapt to changes in different environments and data sources, thereby improving the flexibility and effectiveness of data processing. S3. Further process the data using a wavelet threshold decomposition method based on parameter optimization. This method adjusts the wavelet coefficients according to a provided function expression to optimize data quality. Where Wj,k are wavelet coefficients, λ is the threshold, α and β are adjustment factors, and sign is the sign function. After parameter adjustment and algorithm optimization, the optimal adjustment factor and the improved threshold function eliminate discontinuities at the threshold while making the function quickly approach the hard threshold function. S4. Apply deep neural network technology and machine learning algorithms to perform in-depth analysis on the processed data in order to assess the current state and structural health of the bridge. S5. The machine learning algorithm adjusts its parameters based on the equipment's operating data to adapt to the aging process of the bridge and changing environmental conditions. S6. Based on the output of the machine learning algorithm, predict the potential damage and maintenance needs of the bridge, and update the damage prediction model to integrate the latest bridge performance data and maintenance feedback. S7. Automatically send the prediction results and maintenance suggestions to the operation and maintenance management personnel and display them through the user interface, allowing the operation and maintenance management personnel to view the bridge status, prediction results and execute remote troubleshooting commands; S8. It adopts data encryption and verification mechanisms, uses advanced encryption standards to ensure the security and integrity of all encrypted wireless transmission and storage of data, and implements data access permission verification through multi-factor authentication to prevent data leakage and unauthorized access. S9. Fault-tolerant mechanisms are adopted, including local data caching and automatic data recovery functions, to ensure that the system can continue to operate and maintain the highest possible data integrity in the event of unstable network connection or partial data loss. The S10 data processing and analysis module has self-diagnostic capabilities to detect and report internal system errors and performance bottlenecks.

2. The intelligent AI recognition method based on multi-sensor data fusion according to claim 1, characterized in that: The wavelet thresholding method uses Daubechies wavelets for four-level wavelet decomposition and employs a threshold function between hard and soft thresholds for noise cancellation.

3. The intelligent AI recognition method based on multi-sensor data fusion according to claim 1, characterized in that: The machine learning algorithm includes Fast Fourier Transform for extracting frequency domain features from time series data, and Principal Component Analysis for dimensionality reduction of the features.

4. The intelligent AI recognition method based on multi-sensor data fusion according to claim 1, characterized in that: The prediction model adopts an online learning approach, which can integrate new data feedback in real time to improve the accuracy of prediction and the adaptability of the model. The user interface has touch and voice control functions to facilitate operation by managers in complex environments.

5. The intelligent AI recognition method based on multi-sensor data fusion according to claim 1, characterized in that: The encryption protocols include, but are not limited to, TLS / SSL, to ensure end-to-end security during data transmission. The data recovery function uses a cloud-based backup system to asynchronously replicate data across multiple locations, ensuring high data availability and disaster recovery capabilities.

6. The intelligent AI recognition method based on multi-sensor data fusion according to claim 1, characterized in that: The self-diagnostic function includes using machine learning technology to analyze system operation logs, automatically identifying potential system anomalies and performance degradation trends, and the parameter self-adjustment mechanism optimizes its identification algorithm and parameter settings based on environmental changes and historical maintenance data.

7. The intelligent AI recognition method based on multi-sensor data fusion according to claim 1, characterized in that: The encrypted wireless transmission performs digital signature and integrity verification before transmission to prevent data from being tampered with during transmission. The deep neural network technology includes a combination of convolutional neural networks and recurrent neural networks to process and analyze time-dependent data characteristics.

8. The intelligent AI recognition method based on multi-sensor data fusion according to claim 1, characterized in that: The data undergoes data cleaning, which includes using an outlier detection algorithm to automatically identify and remove statistically abnormal data points. The data denoising process employs machine learning-based adaptive filtering technology to improve the accuracy and efficiency of data cleaning.

Images