Intelligent adjustment method and system for bridge rotation posture based on multi-source data fusion

Abstract

The present invention provides a method and system for intelligently adjusting the bridge rotation posture based on multi-source data fusion. The method obtains a multi-source monitoring data set generated during the bridge rotation process, performs spatiotemporal alignment and fusion processing on the multi-source monitoring data set, and generates fused monitoring data. Based on the fused monitoring data and a preset target posture parameter set, the method identifies the posture deviation characteristics between the bridge rotation posture and the target posture. Based on the posture deviation characteristics, a posture adjustment strategy for the bridge rotation actuator is generated, and the posture adjustment strategy is executed to perform real-time posture adjustment operations on the bridge rotation process. The present invention can improve adjustment efficiency while enhancing system reliability, solving the problems of insufficient control accuracy and response delay caused by data isolation and static strategies in traditional methods.

Description

Technical Field

[0001] The present invention relates to the fields of data processing and intelligent control, and in particular to a method and system for intelligently adjusting the rotation posture of a bridge based on multi-source data fusion. Background Art

[0002] Adjusting the rotational posture of a bridge is crucial in bridge construction. The goal is to ensure that the rotating structure's spatial posture precisely matches a preset target trajectory through real-time monitoring and dynamic control. Existing adjustment methods typically rely on simple monitoring data (such as displacements of key bridge points or actuator hydraulic parameters) for posture feedback, using preset fixed adjustment commands to drive the actuator to complete the rotation operation. However, such methods struggle to accurately quantify the coupled effects of environmental disturbances (such as wind and temperature variations) and actuator response lag on the rotational posture, resulting in unrecognized dynamic deviations between the actual and target postures. Furthermore, fixed adjustment strategies are unable to adapt to the nonlinear changes in structural deformation and external disturbances during the rotation process, easily leading to a mismatch between adjustment commands and real-time operating conditions. This results in reduced adjustment accuracy, frequent actuator starts and stops, and even the risk of local overload, hindering the safety and efficiency of long-span bridge rotation construction. Summary of the Invention

[0003] The present invention provides a method and system for intelligently adjusting the rotation posture of a bridge based on multi-source data fusion.

[0004] In a first aspect, an embodiment of the present invention provides a method for intelligently adjusting a bridge rotation posture based on multi-source data fusion, comprising:

[0005] Acquire a multi-source monitoring data set generated during the bridge rotation process, wherein the multi-source monitoring data set includes bridge structure response monitoring data, environmental impact factor monitoring data, and rotation actuator operation status monitoring data;

[0006] Performing spatiotemporal alignment and fusion processing on the multi-source monitoring data set to generate fused monitoring data;

[0007] Based on the fused monitoring data and a preset target posture parameter set, identifying posture deviation characteristics between the bridge rotation posture and the target posture;

[0008] generating a posture adjustment strategy for the bridge rotation actuator according to the posture deviation characteristics;

[0009] The posture adjustment strategy is executed to perform real-time posture adjustment operations on the bridge rotation process.

[0010] In a second aspect, an embodiment of the present invention provides an intelligent bridge rotation posture adjustment system based on multi-source data fusion, comprising:

[0011] a memory storing a computer program;

[0012] The processor is used to load the computer program to implement the above-mentioned intelligent adjustment method for bridge rotation posture based on multi-source data fusion.

[0013] The present invention provides an intelligent bridge rotation posture adjustment method based on multi-source data fusion. By acquiring a multi-source monitoring data set during the bridge rotation process, integrating multi-dimensional information on structural response, environmental impact, and actuator operation status, and performing spatiotemporal alignment and fusion processing on the multi-source data, the spatiotemporal reference differences in the sensor acquisition process are eliminated, and spatiotemporal consistent fusion monitoring data is generated, providing a reliable data basis for the accurate identification of posture deviations. Based on the dynamic comparison of fused monitoring data and target posture parameters, the posture deviation characteristics are collaboratively analyzed from multiple dimensions such as three-dimensional coordinate offset, overall torsion angle, and cross-sectional deformation degree, and the comprehensive deviation degree is comprehensively quantified, breaking through the limitations of traditional single parameter deviation analysis and significantly improving the accuracy of deviation identification. By combining the posture adjustment strategy generation model with structural mechanics constraints and actuator operation rules, an adjustment instruction combination adapted to the current rotation state is dynamically generated, driving multiple execution units to collaboratively adjust according to action type, direction, amplitude, and timing, realizing adaptive optimization of the deviation elimination process, and effectively avoiding local overload or adjustment lag problems. It performs closed-loop control and provides real-time feedback of multi-source monitoring data to ensure continuous optimization and precise execution of adjustment strategies, maintains high-precision control of rotation posture under conditions of complex environmental interference and dynamic structural changes, improves adjustment efficiency while enhancing system reliability, and solves the problems of insufficient control accuracy and response delay caused by data isolation and static strategies in traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] Figure 1 This is a flow chart of a method for intelligently adjusting a bridge rotation posture based on multi-source data fusion provided by an embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of the composition of an intelligent bridge rotation posture adjustment system based on multi-source data fusion provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] See also Figure 1 , Figure 1 A flowchart of a method for intelligently adjusting a bridge rotation posture based on multi-source data fusion provided by an embodiment of the present invention. The method can be executed by a computer system and may include the following steps:

[0019] Step S100: Acquire a multi-source monitoring data set generated during the bridge rotation process, wherein the multi-source monitoring data set includes bridge structure response monitoring data, environmental impact factor monitoring data, and rotation actuator operation status monitoring data.

[0020] Bridge structure response monitoring data reflects the mechanical response and deformation of a bridge structure during its rotation. During a bridge rotation, its structure is subjected to various forces, generating corresponding responses. This response data can reflect the actual operating status of the bridge structure. For example, data such as displacement, strain, and stress at key bridge locations are all considered bridge structure response monitoring data. Displacement data can be collected using displacement sensors, which can be installed at key nodes or target locations on the bridge to measure the bridge's position changes in real time during rotation. Strain data can be acquired using strain gauges, which are affixed to the surface of the bridge structure. When the structure deforms, the resistance of the strain gauge changes, and by measuring this resistance change, the corresponding strain value can be obtained. Stress data can be calculated based on the strain data combined with the mechanical properties of the material.

[0021] Environmental impact factor monitoring data describes the environmental conditions during the bridge rotation process. Environmental factors can have a significant impact on bridge rotation. Different environmental conditions may cause changes in the mechanical properties of the bridge structure and even affect the safety and stability of the rotation. Environmental impact factor monitoring data includes air temperature, wind speed, humidity, and other factors. Air temperature data can be collected using temperature sensors installed at various locations around the bridge to obtain accurate ambient temperature information. Wind speed data can be measured using anemometers, which are typically installed higher on the bridge to prevent interference from surrounding objects. Humidity data can be collected using humidity sensors, which can monitor the ambient humidity in real time.

[0022] The operating status monitoring data for the swivel actuator is data on the working conditions of the bridge's swivel actuator. The swivel actuator is a key device for achieving bridge rotation, and its operating status is directly related to the smooth progress of the rotation. This data covers information such as the speed and torque of the actuator's drive motor, as well as the pressure and flow of the hydraulic system. The drive motor's speed can be measured using a speed sensor, which is installed on the motor shaft and monitors the motor's rotational speed in real time. Torque data can be obtained using a torque sensor, which is installed on the motor's output shaft and measures the torque output. Hydraulic system pressure data can be collected using a pressure sensor, which is installed in the hydraulic pipeline and monitors pressure changes in the hydraulic system in real time. Flow data can be measured using a flow sensor, which is installed at the outlet of the hydraulic pump or other key locations to monitor the flow of hydraulic oil.

[0023] In practice, to obtain accurate multi-source monitoring data, multiple monitoring points can be deployed at the bridge rotation site, and corresponding sensor equipment can be installed. For example, displacement sensors and strain gauges can be installed at different spans and heights of the bridge to comprehensively monitor the response of the bridge structure. Temperature sensors, anemometers, and humidity sensors can be installed at multiple locations around the bridge to obtain accurate data on environmental factors. Speed ​​sensors, torque sensors, pressure sensors, and flow sensors can be installed at key locations on the rotation actuator to monitor the actuator's operating status in real time. Furthermore, to ensure data reliability and stability, the sensor equipment requires regular calibration and maintenance to guarantee measurement accuracy.

[0024] Step S200: performing spatiotemporal alignment and fusion processing on the multi-source monitoring data set to generate fused monitoring data.

[0025] Spatiotemporal alignment and fusion processing comprehensively considers the temporal and spatial differences among different types of data in a multi-source monitoring data set, unifying them for temporal and spatial consistency and comparability. This process extracts data that comprehensively reflects the coordinated changes in the bridge's structural status, environmental impacts, and actuator status. Fused monitoring data eliminates temporal and spatial differences and comprehensively reflects the coordinated changes in the bridge's structural status, environmental impacts, and actuator status. This data provides more comprehensive and accurate information for subsequent bridge rotation analysis and adjustments.

[0026] As an embodiment, the step S200 of performing spatiotemporal alignment and fusion processing on the multi-source monitoring data set to generate fused monitoring data may specifically include the following steps S210 to S240:

[0027] Step S210: performing data validity verification on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotary actuator operation status monitoring data, respectively, eliminating invalid data items, and obtaining a valid monitoring data set.

[0028] Data validation checks and verifies collected monitoring data to determine whether it meets pre-defined validity criteria. Invalid data items may be caused by sensor failure, data transmission errors, or other factors, resulting in outliers or missing values. For example, when a sensor is subject to external interference, it may output values ​​outside the normal range, resulting in invalid data. Data validation improves data quality and prevents invalid data from impacting subsequent analysis and processing.

[0029] For bridge structure response monitoring data, validity checks can include checking whether displacement data is within a reasonable range and whether strain data is consistent with the material's mechanical properties. If the displacement data exceeds the maximum displacement allowed by the bridge structure design, or if the strain data differs significantly from the theoretical value calculated based on the material's mechanical properties, the data is considered invalid. For environmental factor monitoring data, validity checks can include checking whether temperature data is within a reasonable range of local historical temperatures and whether wind speed data is consistent with actual meteorological conditions. If temperature data shows abnormally high or low temperatures, or if wind speed data is significantly inconsistent with the actual surrounding environment, the data may be invalid. For rotary actuator operating status monitoring data, validity checks can include checking whether the drive motor speed and torque are within the motor's rated operating range and whether the hydraulic system pressure and flow meet the system's design requirements. If the drive motor speed exceeds the rated speed, or if the hydraulic system pressure is too high or too low, the data is considered invalid.

[0030] In practice, statistical analysis methods can be used to verify data validity. For example, statistical parameters such as the mean and standard deviation of various monitoring data can be calculated and used to determine the normal range of the data. Data outside this range can be marked as invalid and discarded. Furthermore, a comprehensive assessment can be made based on the sensor's operating status and historical data to improve the accuracy of data validation.

[0031] Step S220: Perform timestamp alignment processing on the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data in the effective monitoring data set, so that the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data have a synchronous correspondence in the time dimension.

[0032] Timestamp alignment aligns different types of monitoring data so that they all have valid values ​​at the same point in time. Because different sensors may have different data collection frequencies and time bases, the collected data may exhibit temporal discrepancies. Timestamp alignment eliminates these discrepancies, providing accurate, time-synchronized data for subsequent data analysis and fusion.

[0033] As an embodiment, the step S220 of performing timestamp alignment processing on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotary actuator operation status monitoring data in the valid monitoring data set may specifically include the following steps S221 to S224:

[0034] Step S221: Obtain a first timestamp sequence corresponding to the bridge structure response monitoring data, a second timestamp sequence corresponding to the environmental impact factor monitoring data, and a third timestamp sequence corresponding to the rotary actuator operation status monitoring data in the valid monitoring data set.

[0035] A timestamp sequence records the collection time of each piece of monitoring data. The first timestamp sequence corresponds to the collection time of the bridge structure response monitoring data, the second timestamp sequence corresponds to the collection time of the environmental impact factor monitoring data, and the third timestamp sequence corresponds to the collection time of the rotary actuator operating status monitoring data. By obtaining these timestamp sequences, the collection time sequence and time interval of each type of monitoring data can be clearly determined.

[0036] In practice, sensors record the acquisition time of each data point when collecting data. Arranging these acquisition times in the order in which the data was collected creates a corresponding timestamp sequence. For example, when a displacement sensor collects displacement data from a bridge structure, it records the time of each acquisition. Arranging these times in sequence forms the first timestamp sequence.

[0037] Step S222: Time resampling processing is performed on the first timestamp sequence, the second timestamp sequence and the third timestamp sequence at a preset reference time interval to generate a first resampling data sequence of the bridge structure response monitoring data, a second resampling data sequence of the environmental impact factor monitoring data and a third resampling data sequence of the rotating actuator operation status monitoring data, respectively.

[0038] Time resampling converts data with different time intervals into data with a uniform time interval. The preset baseline time interval is a fixed interval determined based on the needs of data analysis and the characteristics of each type of monitoring data. Time resampling ensures that all types of monitoring data have the same sampling interval, facilitating subsequent time synchronization.

[0039] For example, assume that the collection time interval for bridge structure response monitoring data is 10 seconds, the collection time interval for environmental impact factor monitoring data is 20 seconds, and the collection time interval for rotary actuator operation status monitoring data is 15 seconds. The preset reference time interval is 5 seconds. When performing time resampling processing, for the first timestamp sequence, the original timestamp sequence is interpolated or sampled according to the reference time interval to generate a new timestamp sequence, and the bridge structure response monitoring data is resampled according to the new timestamp sequence to obtain a first resampled data sequence. The second and third timestamp sequences are processed in the same way to generate a second resampled data sequence and a third resampled data sequence, respectively.

[0040] Time resampling can be performed using a linear interpolation algorithm. This algorithm uses the values ​​of two adjacent data points and their time interval to calculate the data value at the intermediate time point. For example, consider two adjacent displacement data points, A and B, acquired at times t1 and t2, respectively, and with data values ​​x1 and x2. If you want to interpolate at a time point t between t1 and t2, you can use the linear interpolation formula: x = x1 + (x2 - x1) × (t - t1) / (t2 - t1) to obtain the displacement data value x corresponding to time point t.

[0041] Step S223: Time synchronization is performed on the first resampled data sequence, the second resampled data sequence, and the third resampled data sequence to ensure that corresponding valid data points exist in the first resampled data sequence, the second resampled data sequence, and the third resampled data sequence at the same reference time point.

[0042] Time synchronization further adjusts the resampled data series so that all types of monitoring data have corresponding valid data at the same benchmark time point. During the time resampling process, some data series may not have valid data at certain time points. Time synchronization can solve this problem.

[0043] In practice, a data alignment algorithm can be used for time synchronization. This algorithm compares the data of different resampled data sequences at various benchmark time points. For sequences lacking data at a particular benchmark time point, interpolation or extrapolation can be used to supplement the data. For example, if the first resampled data sequence has data at a particular benchmark time point, while the second resampled data sequence does not, linear interpolation can be performed based on the data from adjacent time points in the second resampled data sequence to obtain approximate data for that benchmark time point.

[0044] Step S224: The first resampled data sequence, the second resampled data sequence and the third resampled data sequence that have completed time synchronization processing are respectively used as the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data after timestamp alignment processing.

[0045] After time synchronization processing, various types of monitoring data have been aligned in time. These synchronized data sequences are used as corresponding data after timestamp alignment processing, providing a temporally consistent data basis for subsequent spatial position registration processing and data fusion operations.

[0046] Step S230: Perform spatial position registration processing on the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data in the valid monitoring data set after timestamp alignment processing, so that the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data have a unified spatial reference benchmark in the spatial dimension.

[0047] Spatial registration is the process of spatially calibrating different types of monitoring data so that they share a common reference. Since different types of monitoring data may be collected in different spatial reference frames, spatial discrepancies may exist. Spatial registration eliminates these discrepancies, providing accurate, spatially synchronized data for subsequent data analysis and fusion.

[0048] As an embodiment, the step S230 of performing spatial position registration processing on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotary actuator operation status monitoring data in the valid monitoring data set after the timestamp alignment processing may specifically include the following steps S231 to S236:

[0049] Step S231: establishing a unified spatial coordinate system as a registration reference coordinate system.

[0050] A unified spatial coordinate system is a defined reference framework used to describe spatial locations. The registration reference coordinate system is a standard spatial coordinate system determined based on the actual conditions of bridge rotation and the needs of data analysis. The purpose of establishing a registration reference coordinate system is to convert monitoring data from different spatial reference systems into a unified coordinate system for unified comparison and analysis of spatial locations.

[0051] For example, a Cartesian coordinate system can be established with a fixed point on the bridge as the origin and the bridge axis as the coordinate axis. In this coordinate system, the positions of various parts of the bridge and its surroundings can be accurately represented using coordinate values.

[0052] Step S232: Identify the first spatial reference system corresponding to the bridge structure response monitoring data, the second spatial reference system corresponding to the environmental impact factor monitoring data, and the third spatial reference system corresponding to the rotation actuator operation status monitoring data.

[0053] The first, second, and third spatial reference systems are used for collecting data on bridge structural response monitoring, environmental impact monitoring, and the operating status of the rotating actuator, respectively. Different types of monitoring data may use different spatial reference systems due to differences in sensor installation locations and measurement methods.

[0054] For example, sensors monitoring bridge structural response data may be installed in different locations within the bridge structure, and their spatial reference system may be based on the local structure where the sensors are located. Sensors monitoring environmental impact factors may be installed in various locations around the bridge, and their spatial reference system may be based on local geographic coordinates. Sensors monitoring the operating status of a rotary actuator may be installed on various components of the actuator, and their spatial reference system may be based on the actuator's mechanical structure.

[0055] Step S233: Calculating spatial transformation parameters between the first spatial reference system, the second spatial reference system, the third spatial reference system, and the registration reference coordinate system respectively.

[0056] Spatial transformation parameters are a set of parameters used to transform data in different spatial reference systems into the registration reference coordinate system. By calculating spatial transformation parameters, coordinate transformation between different spatial reference systems can be achieved.

[0057] As an implementation manner, the step S233 of respectively calculating the spatial transformation parameters between the first spatial reference system, the second spatial reference system, and the third spatial reference system and the registration reference coordinate system may specifically include the following steps S2331 to S2334:

[0058] Step S2331: Acquire a reference coordinate set of a preset number of common control points in the registration reference coordinate system.

[0059] Common control points are points whose positions can be accurately determined in both the registration reference coordinate system and various spatial reference systems. The preset number of common control points is the target number determined based on the spatial transformation calculations. The reference coordinate set is the coordinate values ​​of these common control points in the registration reference coordinate system.

[0060] For example, several obvious and easily identifiable points are selected as common control points at the bridge rotation site, such as the vertices of the bridge piers, preset markers, etc. The reference coordinate set is obtained by measuring the coordinates of these common control points in the registration reference coordinate system.

[0061] Step S2332: For the first spatial reference system, obtain the first source coordinate set of the common control point in the first spatial reference system, perform coordinate matching between the reference coordinate set and the first source coordinate set to generate a first coordinate pair set, calculate the first rotation matrix and the first translation vector from the first spatial reference system to the alignment reference coordinate system based on the first coordinate pair set, and combine the first rotation matrix and the first translation vector into the first spatial transformation parameters.

[0062] The first source coordinate set is the coordinate values ​​of the common control points in the first spatial reference system. The first source coordinate set is obtained by measuring the coordinates of the common control points in the first spatial reference system. Coordinate matching is a one-to-one correspondence between each reference coordinate in the reference coordinate set and the source coordinates of the corresponding common control point in the first source coordinate set. The first coordinate pair set consists of the matched reference coordinates and source coordinates.

[0063] The first rotation matrix and the first translation vector are parameters used to describe the rotation and translation relationship from the first spatial reference system to the registration reference coordinate system. Based on the first coordinate pair set, the first rotation matrix and the first translation vector can be calculated using methods such as the least squares method. The least squares method is a method for solving parameters by minimizing the sum of squares of errors. Specifically, based on the coordinate values ​​in the first coordinate pair set, an error equation is established, and the optimal values ​​of the first rotation matrix and the first translation vector are solved by minimizing the sum of squares of the error equation. By combining the first rotation matrix and the first translation vector, the first spatial transformation parameters are obtained.

[0064] Step S2333: For the second spatial reference system, obtain the second source coordinate set of the common control point in the second spatial reference system, perform coordinate matching between the base coordinate set and the second source coordinate set to generate a second coordinate pair set, calculate the second rotation matrix and second translation vector from the second spatial reference system to the registration base coordinate system based on the second coordinate pair set, and combine the second rotation matrix and the second translation vector into second spatial transformation parameters.

[0065] The second source coordinate set is the coordinate value of the common control point in the second spatial reference system. Similar to step S2332, the second source coordinate set is obtained by measuring the coordinates of the common control point in the second spatial reference system. The reference coordinate set is matched with the second source coordinate set to generate a second coordinate pair set. Based on the second coordinate pair set, a method such as the least squares method is used to calculate a second rotation matrix and a second translation vector, which are combined to form the second spatial transformation parameters.

[0066] Step S2334: For the third spatial reference system, obtain the third source coordinate set of the common control point in the third spatial reference system, perform coordinate matching between the reference coordinate set and the third source coordinate set to generate a third coordinate pair set, calculate the third rotation matrix and third translation vector from the third spatial reference system to the alignment reference coordinate system based on the third coordinate pair set, and combine the third rotation matrix and the third translation vector into third spatial transformation parameters.

[0067] The third source coordinate set is the coordinate values ​​of the common control points in the third spatial reference system. Similarly, the third source coordinate set is obtained by measuring the coordinates of the common control points in the third spatial reference system. The reference coordinate set is matched with the third source coordinate set to generate a third coordinate pair set. Based on the third coordinate pair set, a third rotation matrix and a third translation vector are calculated using a method such as the least squares method, and these are combined to form the third spatial transformation parameters.

[0068] Step S234: Based on the spatial conversion parameters, the bridge structure response monitoring data after the timestamp alignment processing is converted from the first spatial reference system to the registration reference coordinate system to obtain first registered monitoring data.

[0069] Based on the first spatial transformation parameters, the coordinates of each data point in the bridge structure response monitoring data, which has undergone timestamp alignment, are transformed from the first spatial reference system to the registration reference coordinate system. Specifically, the coordinates of each data point in the bridge structure response monitoring data are first multiplied by the first rotation matrix and then added to the first translation vector to obtain the new coordinates of the data point in the registration reference coordinate system. After the coordinates of all data points are transformed, the first registered monitoring data is obtained.

[0070] Step S235: Based on the spatial conversion parameters, the environmental impact factor monitoring data after the timestamp alignment process is converted from the second spatial reference system to the registration reference coordinate system to obtain second registration monitoring data.

[0071] In a method similar to step S234 , according to the second spatial conversion parameters, the environmental impact factor monitoring data after the timestamp alignment process is converted from the second spatial reference system to the registration reference coordinate system to obtain second registered monitoring data.

[0072] Step S236: Based on the spatial conversion parameters, the rotating actuator operation status monitoring data after the timestamp alignment processing is converted from the third spatial reference system to the alignment reference coordinate system to obtain the third alignment monitoring data; the first alignment monitoring data, the second alignment monitoring data and the third alignment monitoring data together constitute the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data after the spatial position alignment processing.

[0073] Similarly, the rotary actuator operating status monitoring data, which has undergone timestamp alignment, is transformed from the third spatial reference system to the registration reference coordinate system based on the third spatial transformation parameters, generating the third registered monitoring data. The first, second, and third registered monitoring data are spatially unified in the registration reference coordinate system, and together they constitute the various monitoring data after spatial position registration.

[0074] Step S240: After completing the timestamp alignment processing and the spatial position registration processing, the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data are input into a preset data fusion model, and the data fusion model is used to perform feature-level fusion operations on the input types of monitoring data, and output the fused monitoring data; the fused monitoring data is a data set that eliminates time and space differences and comprehensively reflects the coordinated change characteristics of the bridge structure status, environmental impact and actuator status.

[0075] The data fusion model is used to integrate different types of monitoring data. Feature-level fusion extracts and fuses the features of the input monitoring data to produce a fusion feature that comprehensively reflects the information from all of the data. By performing feature-level fusion operations on the data fusion model, different types of monitoring data can be organically combined, eliminating temporal and spatial differences. This yields a data set that comprehensively and accurately reflects the coordinated changes in the bridge structure, environmental impacts, and actuator status.

[0076] In one embodiment, the data fusion model is a feature fusion network based on an attention mechanism. Based on this, in step S240, the data fusion model is used to perform a feature-level fusion operation on the input monitoring data of various types, and the fused monitoring data is output. Specifically, the following steps S241 to S246 may be included:

[0077] Step S241: inputting the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotation actuator operation status monitoring data after spatial position registration processing into different feature extraction branches of the feature fusion network based on the attention mechanism respectively.

[0078] The attention-based feature fusion network is a neural network model that uses the attention mechanism to perform feature fusion. The network consists of multiple feature extraction branches, each of which extracts features for a corresponding type of monitoring data. After spatial registration, each type of monitoring data is fed into a different feature extraction branch, allowing for independent feature extraction for each type of data.

[0079] For example, a feature fusion network based on an attention mechanism can use a convolutional neural network (CNN) as the basic structure of its feature extraction branch. Bridge structure response monitoring data is input into a dedicated CNN feature extraction branch, which extracts deep features from this data through operations such as convolutional and pooling layers. Environmental impact factor monitoring data and rotary actuator operating status monitoring data are input into two separate CNN feature extraction branches for corresponding feature extraction.

[0080] Step S242: extract the deep feature expression of the corresponding type of monitoring data through each feature extraction branch.

[0081] Deep feature representations are high-level features that reflect the essential characteristics of the corresponding type of monitoring data. Each feature extraction branch processes the input monitoring data through a series of neural network layers, gradually extracting the data's deep features. Taking the CNN feature extraction branch as an example, the convolution layer convolves the input data with a convolution kernel to extract local features of the data. The pooling layer downsamples the feature map output by the convolution layer, reducing its size while retaining important feature information. Through the combination of multiple convolutional and pooling layers, the deep feature representation of the data is gradually extracted.

[0082] Step S243: Input the deep feature expressions extracted by each feature extraction branch into the attention allocation module.

[0083] The attention allocation module is a key module in the feature fusion network based on the attention mechanism. It is used to allocate attention to the deep feature expressions of different types of monitoring data. The deep feature expressions extracted by each feature extraction branch are input into the attention allocation module so that different features can be weighted according to their importance.

[0084] Step S244: In the attention allocation module, based on the intrinsic correlation between the state of the current bridge rotation stage and the deep feature expressions of the various types of monitoring data, the attention weight coefficients to be allocated to the deep feature expressions of the bridge structure response monitoring data, the deep feature expressions of the environmental impact factor monitoring data, and the deep feature expressions of the rotation actuator operation status monitoring data are calculated.

[0085] The attention weight coefficient is a coefficient used to measure the importance of the deep feature representations of different types of monitoring data in the fusion process. In the attention allocation module, an attention mechanism algorithm is used to calculate the attention weight coefficient to be assigned to each deep feature representation based on the inherent correlation between the current bridge rotation phase state and the deep feature representations of each type of monitoring data. The attention mechanism algorithm can calculate the attention weight coefficient based on factors such as the similarity and correlation between features. For example, the dot product attention mechanism can be used to calculate the dot product between the deep feature representations to obtain the similarity between them, and then calculate the attention weight coefficient based on the similarity.

[0086] Step S245: Based on the calculated attention weight coefficient, the deep feature expression of the bridge structure response monitoring data, the deep feature expression of the environmental impact factor monitoring data, and the deep feature expression of the rotating actuator operation status monitoring data are weightedly fused.

[0087] Weighted fusion is a process that combines the deep feature representations of various types of monitoring data with the calculated attention weight coefficients. This process can highlight the role of important features while weakening the influence of unimportant ones, resulting in a more comprehensive and accurate feature representation.

[0088] For example, assuming the deep feature expression of bridge structure response monitoring data is F1, the deep feature expression of environmental impact factor monitoring data is F2, and the deep feature expression of rotary actuator operation status monitoring data is F3, and their corresponding attention weight coefficients are w1, w2, and w3, respectively, the weighted fusion feature expression F = w1 × F1 + w2 × F2 + w3 × F3.

[0089] Step S246: performing feature dimension reduction and integration operations on the weighted fused feature expressions to generate the fused monitoring data; the fused monitoring data is a unified feature vector sequence that integrates structural response, environmental impact, and mechanism status information.

[0090] Feature dimensionality reduction and integration operations further process the weighted fused feature representation, reducing the feature dimensionality and integrating different features to generate a unified sequence of feature vectors. Feature dimensionality reduction can be achieved using methods such as principal component analysis (PCA). PCA converts high-dimensional data into low-dimensional data through linear transformations. Specifically, the covariance matrix of the weighted fused feature representation is calculated, and the eigenvalues ​​and eigenvectors of the covariance matrix are then calculated. The first few eigenvectors with the largest eigenvalues ​​are selected as principal components. The original feature representation is then projected onto these principal components to obtain the reduced-dimensional feature representation.

[0091] The reduced-dimensional feature representations are then integrated to generate a unified feature vector sequence, which is the fused monitoring data. This fused monitoring data eliminates temporal and spatial differences and comprehensively reflects the coordinated changes in the bridge structure, environmental impacts, and actuator status.

[0092] Step S300: Based on the fused monitoring data and a preset target posture parameter set, identifying posture deviation features between the bridge rotation posture and the target posture.

[0093] The target attitude parameter set is the preset ideal attitude parameters for the bridge rotation. The attitude deviation feature describes the difference between the actual bridge rotation attitude and the target attitude. By comparing and fusing monitoring data with the target attitude parameter set, the attitude deviation feature between the bridge rotation attitude and the target attitude can be identified, providing a basis for subsequent attitude adjustments.

[0094] As an embodiment, the step S300, based on the fused monitoring data and a preset target posture parameter set, identifying the posture deviation characteristics between the bridge rotation posture and the target posture, may specifically include the following steps S310 to S360:

[0095] Step S310: extracting a current posture feature set representing the current spatial position and posture of the bridge from the fused monitoring data; the current posture feature set at least includes the three-dimensional coordinate offset of the key control points of the bridge, the overall torsion angle of the bridge, and the deformation degree of the target section of the bridge.

[0096] The current posture feature set is a set of features used to describe the current spatial position and posture of the bridge. The 3D coordinate offset of the bridge's key control points refers to the difference between the actual 3D coordinates of the bridge's key control points and the initial 3D coordinates. It reflects the changes in the bridge's spatial position. The overall torsion angle of the bridge refers to the torsion angle of the bridge as a whole relative to its initial state, which reflects the torsional deformation of the bridge. The deformation degree of the bridge's target section refers to the degree of deformation of the target section during the rotation process, such as changes in cross-sectional area and shape. It reflects the deformation of the bridge's local structure.

[0097] Feature extraction algorithms can be used to extract the current posture feature set from the fused monitoring data. For example, the 3D coordinate offset of a bridge's key control point can be calculated based on the 3D coordinate information of the bridge's key control point in the fused monitoring data, and the difference between the initial 3D coordinates and the offset can be calculated. The overall torsion angle of the bridge can be calculated using geometric calculation methods by analyzing the relative displacement relationships between different parts of the bridge in the fused monitoring data. The degree of deformation of a target bridge section can be determined by calculating the area change rate, shape change parameters, and other parameters based on the geometric information of the target section in the fused monitoring data.

[0098] Step S320: Obtain a target posture feature set corresponding to the current posture feature set from the preset target posture parameter set; the target posture feature set includes at least the target three-dimensional coordinate offset of the key control point of the bridge, the overall target torsion angle of the bridge, and the target deformation degree of the target section of the bridge.

[0099] The target pose feature set is the pre-defined ideal pose feature set corresponding to the current pose feature set. The target 3D coordinate offset of the bridge's key control point is the difference between the 3D coordinates of the bridge's key control point at the target pose and the initial 3D coordinates. The target overall torsion angle of the bridge is the overall torsion angle of the bridge at the target pose. The target deformation degree of the target bridge section is the degree of deformation of the target bridge section at the target pose.

[0100] Obtaining a target pose feature set from a preset target pose parameter set involves extracting corresponding target feature values ​​from the target pose parameter set based on the feature type of the current pose feature set. For example, if the current pose feature set contains the 3D coordinate offset of a bridge's key control point, then the corresponding target 3D coordinate offset of the bridge's key control point is found from the target pose parameter set.

[0101] Step S330: performing a difference comparison process on the three-dimensional coordinate offsets of the key control points of the bridge in the current posture feature set and the target three-dimensional coordinate offsets of the key control points of the bridge in the target posture feature set to generate a position deviation feature.

[0102] The difference comparison process compares the 3D coordinate offsets of the bridge's key control points in the current pose feature set with the corresponding target 3D coordinate offsets in the target pose feature set, and calculates the difference between them. The position deviation feature reflects the deviation of the bridge's key control points from their target position in space.

[0103] For example, suppose the 3D coordinate offset of the bridge's key control point in the current pose feature set is (x1, y1, z1), and the target 3D coordinate offset of the corresponding bridge's key control point in the target pose feature set is (x2, y2, z2). Then the position deviation feature is (x1-x2, y1-y2, z1-z2).

[0104] Step S340: performing a difference comparison process on the overall torsion angle of the bridge in the current posture feature set and the overall target torsion angle of the bridge in the target posture feature set to generate a torsion deviation feature.

[0105] Similarly, the bridge's overall torsion angle in the current pose feature set is compared with the bridge's overall target torsion angle in the target pose feature set, and the difference between them is calculated to obtain a torsion deviation feature. The torsion deviation feature reflects the deviation between the bridge's overall torsion angle and the target torsion angle.

[0106] For example, assuming that the overall torsion angle of the bridge in the current posture feature set is θ1, and the overall target torsion angle of the bridge in the target posture feature set is θ2, then the torsion deviation feature is θ1-θ2.

[0107] Step S350: performing a difference comparison process on the deformation degree of the target bridge section in the current posture feature set and the target deformation degree of the target bridge section in the target posture feature set to generate a deformation deviation feature.

[0108] The deformation degree of the target bridge section in the current pose feature set is compared with the target deformation degree of the target bridge section in the target pose feature set, and the difference between them is calculated to obtain a deformation deviation feature. The deformation deviation feature reflects the deviation between the deformation degree of the target bridge section and the target deformation degree.

[0109] For example, assuming that the deformation degree of the target bridge section in the current posture feature set is d1, and the target deformation degree of the target bridge section in the target posture feature set is d2, then the deformation deviation feature is d1-d2.

[0110] Step S360: The position deviation feature, the torsion deviation feature and the deformation deviation feature are integrated to form the posture deviation feature; the posture deviation feature is used to quantitatively describe the comprehensive deviation degree between the bridge rotation posture and the target posture in terms of spatial position, overall torsion and local deformation.

[0111] The position deviation features, torsion deviation features, and deformation deviation features can be fused using a weighted summation method. For example, let the position deviation feature be P, the torsion deviation feature be T, and the deformation deviation feature be D, and their corresponding weight coefficients be wP, wT, and wD, respectively. Then, the posture deviation feature A = wP × P + wT × T + wD × D.

[0112] The posture deviation feature can accurately and quantitatively describe the comprehensive deviation degree between the bridge rotation posture and the target posture by comprehensively considering the deviation of the bridge in spatial position, overall torsion and local deformation, providing an important basis for subsequent posture adjustment.

[0113] Step S400: generating a posture adjustment strategy for the bridge rotation actuator according to the posture deviation characteristics.

[0114] The attitude adjustment strategy is a set of instructions for adjusting the bridge rotation attitude. It determines the specific actions of the bridge rotation actuator according to the attitude deviation characteristics to reduce the deviation between the bridge rotation attitude and the target attitude.

[0115] As an embodiment, the step S400 generates a posture adjustment strategy for the bridge rotation actuator according to the posture deviation characteristics, which may specifically include the following steps S410 to S440:

[0116] Step S410: Inputting the posture deviation feature into a preset posture adjustment strategy generation model, and analyzing the deviation degree and change trend of the position deviation feature, torsion deviation feature and deformation deviation feature in the posture deviation feature through the posture adjustment strategy generation model.

[0117] The posture adjustment strategy generation model is used to generate posture adjustment strategies based on posture deviation features. By inputting posture deviation features into the posture adjustment strategy generation model, the model analyzes and processes the posture deviation features, determining the degree of deviation and change trends of position deviation features, torsion deviation features, and deformation deviation features.

[0118] The posture adjustment strategy generation model can adopt a decision-making model that combines rules and learning. This model makes decisions based on pre-set rules and learning from historical data. When analyzing posture deviation features, the model can evaluate position deviation features, torsional deviation features, and deformation deviation features according to pre-set rules to determine whether their deviations are large or small, and whether their trends are gradually increasing or decreasing. For example, for position deviation features, if its value exceeds a preset threshold, the deviation is considered large; if its value continues to increase over a period of time, the trend is considered to be gradually increasing.

[0119] Step S420: Based on the built-in bridge structure mechanics constraint rules and rotation actuator operation constraint rules of the posture adjustment strategy generation model, evaluate the effect prediction of different adjustment action combinations on eliminating the position deviation characteristics, the torsion deviation characteristics and the deformation deviation characteristics.

[0120] Bridge structural mechanical constraint rules govern the mechanical performance and safety requirements of bridge structures. Rotational actuator operational constraint rules govern the operational capabilities and safety limits of rotational actuators. Different adjustment action combinations involve different combinations of actions for different drive units within a bridge's rotational actuator. Effect prediction estimates the effectiveness of different adjustment action combinations in eliminating positional, torsional, and deformation deviation characteristics.

[0121] In one embodiment, the posture adjustment strategy generation model is a decision-making model based on a combination of rules and learning; multiple candidate adjustment action combinations are preset in the posture adjustment strategy generation model. Based on this, step S420, based on the bridge structure mechanics constraints and the rotation actuator operation constraints built into the posture adjustment strategy generation model, evaluates the effects of different adjustment action combinations on eliminating the position deviation characteristics, the torsional deviation characteristics, and the deformation deviation characteristics. Specifically, the following steps S421-S424 may be included:

[0122] Step S421: For each candidate adjustment action combination, based on the current values ​​of the position deviation feature, torsion deviation feature and deformation deviation feature in the posture deviation feature, combined with the bridge structure mechanics constraint rules, predict the stress change distribution and potential risk areas of the bridge structure caused by executing the candidate adjustment action combination.

[0123] The stress change distribution of a bridge structure refers to the changes in the magnitude and distribution of internal stress in the bridge structure after the candidate adjustment action combination is executed. Potential risk areas refer to areas of the bridge structure that may pose safety hazards after the candidate adjustment action combination is executed.

[0124] Based on the current values ​​of the position deviation feature, torsional deviation feature, and deformation deviation feature in the posture deviation feature, combined with the mechanical constraint rules of the bridge structure, the finite element analysis method can be used to predict the stress change distribution and potential risk areas of the bridge structure. The finite element analysis method is a method that discretizes the bridge structure into multiple finite elements, establishes the mechanical equations of the elements, and then solves these equations to obtain the stress distribution and deformation of the bridge structure. Specifically, based on the current posture deviation feature and the candidate adjustment action combination, the force condition of the bridge structure is determined, and then the bridge structure is mechanically analyzed using finite element software to obtain the stress change distribution and potential risk areas of the bridge structure.

[0125] Step S422: Based on the current values ​​of the position deviation feature, torsion deviation feature and deformation deviation feature in the posture deviation feature, combined with the operation constraint rules of the rotation actuator, predict the output power, operating speed and duration required by each drive unit when executing the candidate adjustment action combination, and evaluate whether it exceeds the safe operation threshold of the actuator.

[0126] The required output power, operating speed, and duration of each drive unit refer to the power output, operating speed, and operating duration required by each drive unit of the rotary actuator when executing the candidate adjustment action combination. The actuator's safe operating threshold refers to the maximum output power, maximum operating speed, and maximum duration that each drive unit can withstand while ensuring safe operation of the rotary actuator.

[0127] Based on the current values ​​of the position deviation characteristics, torsional deviation characteristics, and deformation deviation characteristics in the posture deviation characteristics, combined with the operating constraints of the rotary actuator, a dynamic analysis method can be used to predict the output power, operating speed, and duration required by each drive unit. The dynamic analysis method is a method that predicts the operating parameters of each drive unit by establishing a dynamic model of the rotary actuator and analyzing its motion and force conditions under different working conditions. Specifically, based on the current posture deviation characteristics and the candidate adjustment action combination, the motion state and force conditions of the rotary actuator are determined. The rotary actuator is then analyzed using dynamic software to obtain the output power, operating speed, and duration required by each drive unit. These predicted values ​​are compared with the safe operation threshold of the actuator to assess whether the safe operation threshold is exceeded.

[0128] Step S423: Based on the predicted stress change distribution of the bridge structure, the potential risk area assessment results, and the output power, operating speed, and duration assessment results of each drive unit, calculate the comprehensive effect score of the candidate adjustment action combination on reducing the position deviation characteristics, the torsional deviation characteristics, and the deformation deviation characteristics.

[0129] The comprehensive effectiveness score is used to measure the effectiveness of candidate adjustment action combinations in reducing posture deviation characteristics. This score is calculated using a comprehensive evaluation index system, combining the predicted stress distribution of the bridge structure, the assessment of potential risk areas, and the output power, operating speed, and duration of each drive unit.

[0130] The comprehensive evaluation index system can include multiple evaluation indicators, such as the degree of stress change in the bridge structure, the size of the potential risk area, the rationality of the drive unit output power, and the stability of the operating speed. A corresponding weight coefficient is assigned to each evaluation indicator. A score for each evaluation indicator is then calculated based on the prediction results. Finally, the scores for each evaluation indicator are weighted and summed according to the weight coefficient to obtain a comprehensive performance score. For example, let the weight coefficient for the degree of stress change in the bridge structure be w1, the weight coefficient for the size of the potential risk area be w2, the weight coefficient for the rationality of the drive unit output power be w3, and the weight coefficient for the stability of the operating speed be w4. The score for the degree of stress change in the bridge structure is S1, the score for the size of the potential risk area is S2, the score for the rationality of the drive unit output power is S3, and the score for the stability of the operating speed is S4. The comprehensive performance score S = w1 × S1 + w2 × S2 + w3 × S3 + w4 × S4.

[0131] Step S424: Based on the historical adjustment effect feedback data obtained by learning the posture adjustment strategy generation model, the comprehensive effect score is corrected; the adjustment instruction combination that promotes the optimal coordinated reduction of the position deviation characteristics, the torsion deviation characteristics and the deformation deviation characteristics is generated to select the candidate adjustment action combination with the highest corrected comprehensive effect score.

[0132] Historical adjustment effect feedback data is the adjustment effect data accumulated by the posture adjustment strategy generation model during past adjustment processes. By learning from this historical adjustment effect feedback data, the model can understand the effectiveness of different adjustment action combinations in real-world applications and thus revise the overall effect score.

[0133] The comprehensive effect score can be modified using a machine learning algorithm, such as a neural network algorithm. Historical adjustment effect feedback data is used as training data to train a neural network model. The predicted comprehensive effect score is used as input and modified using the trained neural network model to obtain the modified comprehensive effect score.

[0134] The candidate adjustment action combination with the highest corrected comprehensive effect score is selected as the adjustment instruction combination that promotes the optimal coordinated reduction of the position deviation feature, the torsion deviation feature, and the deformation deviation feature. This ensures that the selected adjustment action combination has the best effect in reducing the posture deviation feature.

[0135] Step S430: Based on the effect prediction, and on the premise of satisfying the bridge structure mechanics constraint rules and the rotation actuator operation constraint rules, generate an adjustment instruction combination that promotes the optimal coordinated reduction of the position deviation characteristics, the torsional deviation characteristics and the deformation deviation characteristics.

[0136] Based on the prediction results, the optimal adjustment action combination is selected from multiple candidate adjustment action combinations, ensuring that both the structural mechanics constraints of the bridge and the operational constraints of the rotation actuator are met. This adjustment instruction combination specifies the specific actions that each drive unit of the rotation actuator must perform to synergistically reduce positional, torsional, and deformation deviations.

[0137] Step S440: Encapsulating the adjustment instruction combination into the posture adjustment strategy; the posture adjustment strategy includes an instruction set for the action type, action direction, action amplitude and action timing of different driving units in the rotation actuator.

[0138] Adjustment instructions are combined and packaged to form a complete attitude adjustment strategy. This strategy contains detailed instructions for the different drive units in the rotation actuator, including the action type (e.g., rotation, translation), direction (e.g., clockwise, counterclockwise), amplitude (e.g., rotation angle, translation distance), and timing (e.g., action sequence and duration). By encapsulating these adjustment instructions into an attitude adjustment strategy, the instructions can be easily transmitted to the rotation actuator, enabling adjustments to the bridge's rotational attitude.

[0139] Step S500: Execute the posture adjustment strategy to perform real-time posture adjustment operations on the bridge rotation process.

[0140] The real-time attitude adjustment operation is to control the bridge rotation actuator in real time according to the instructions of the attitude adjustment strategy to adjust the rotation attitude of the bridge so that it gradually approaches the target attitude.

[0141] As an embodiment, the step S500, executing the posture adjustment strategy to perform real-time posture adjustment operations on the bridge rotation process, may specifically include the following steps S510 to S550:

[0142] Step S510: parsing the adjustment instruction set in the posture adjustment strategy into a control instruction sequence recognizable by the rotation actuator.

[0143] The adjustment instruction set is a set of instructions for the rotation actuator included in the posture adjustment strategy. The control instruction sequence recognizable by the rotation actuator is an instruction sequence that the rotation actuator can understand and execute after the adjustment instruction set is converted into a format.

[0144] An instruction parsing algorithm can be used to parse the adjustment instruction set. Based on the communication protocol and control requirements of the rotary actuator, the instruction information in the adjustment instruction set is formatted to generate a control instruction sequence recognizable by the rotary actuator. For example, if the rotary actuator utilizes a specific communication protocol, the instruction information in the adjustment instruction set must be encoded according to the format of that communication protocol to generate a control instruction sequence.

[0145] Step S520: sending the control instruction sequence to a plurality of drive controllers corresponding to the rotary actuators.

[0146] The multiple drive controllers corresponding to the rotary actuator are devices responsible for controlling the various drive units of the rotary actuator. A control instruction sequence is sent to the drive controller so that the drive controller controls the operation of the drive unit according to the instruction sequence.

[0147] The control command sequence can be sent to the drive controller via a communication interface. This can be a wired communication interface (such as a serial port or Ethernet) or a wireless communication interface (such as Bluetooth or WiFi). The specific communication interface selected depends on the actual situation and communication requirements of the rotary actuator.

[0148] Step S530: Controlling the corresponding driving units respectively through the plurality of driving controllers to execute the action type, action direction, action amplitude and action timing specified by the control instruction sequence.

[0149] Multiple drive controllers control corresponding drive units to perform corresponding actions based on the control instruction sequence they receive. The drive units operate according to the action type, direction, amplitude, and timing specified in the control instruction sequence, thereby adjusting the bridge's rotation posture.

[0150] For example, if a control instruction sequence specifies that a drive unit needs to rotate clockwise, at a 30-degree angle, and for 10 seconds, the drive controller will control the drive unit to rotate in the specified direction, amplitude, and timing according to this instruction.

[0151] Step S540: During the process of the driving unit performing the action, the multi-source monitoring data set is continuously acquired, and the steps of performing spatiotemporal alignment and fusion processing on the multi-source monitoring data set to generate fused monitoring data are cyclically executed; the step of identifying the posture deviation characteristics between the bridge rotation posture and the target posture based on the fused monitoring data and the preset target posture parameter set; the step of generating a posture adjustment strategy for the bridge rotation actuator based on the posture deviation characteristics; and the step of executing the posture adjustment strategy to perform real-time posture adjustment operations on the bridge rotation process.

[0152] As the drive unit executes its motion, continuously acquiring multi-source monitoring data provides real-time insights into the bridge's rotational state. By looping through subsequent steps, the bridge's rotational posture can be continuously evaluated and adjusted, forming a closed-loop control process that ensures the bridge's rotational posture gradually approaches the target position.

[0153] For example, while the drive unit is executing its motion, sensors continuously collect monitoring data on the bridge structure's response, environmental factors, and the operating status of the rotating actuator. This data is then spatially aligned and fused to generate fused monitoring data. Based on this fused monitoring data and a preset set of target posture parameters, posture deviation characteristics are identified. Based on these deviation characteristics, a new posture adjustment strategy is generated. Finally, the new posture adjustment strategy is implemented to further adjust the bridge's rotation posture.

[0154] Step S550: When the identified posture deviation feature meets the preset posture adjustment completion condition, the execution of the posture adjustment strategy is terminated.

[0155] The attitude adjustment completion condition is a pre-defined criterion used to determine whether the bridge's rotational posture has been adjusted close to the target posture. When the identified posture deviation characteristics meet the attitude adjustment completion condition, the bridge's rotation posture has achieved the desired effect, and the attitude adjustment strategy can be terminated.

[0156] As an embodiment, the step S550, when the identified posture deviation feature satisfies a preset posture adjustment completion condition, terminates the execution of the posture adjustment strategy, which may specifically include the following steps S551 to S555:

[0157] Step S551: After the posture deviation feature is identified each time, the modulus of the position deviation feature, the torsion deviation feature and the deformation deviation feature in the posture deviation feature is calculated.

[0158] The modulus of the position deviation feature, torsional deviation feature, and deformation deviation feature is a numerical value used to measure their magnitude. The modulus can be calculated using the Euclidean distance formula. For example, for the position deviation feature P = (Px, Py, Pz), its modulus |P| = √(Px² + Py² + Pz²). A similar method is used to calculate the modulus of the torsional deviation feature T and the deformation deviation feature D. By calculating the modulus, the magnitude of the position deviation feature, torsional deviation feature, and deformation deviation feature can be quantified, facilitating subsequent judgment and comparison.

[0159] Step S552: Determine whether the module length is less than a preset posture deviation tolerance threshold.

[0160] The preset attitude deviation tolerance threshold is a pre-set value used to measure the acceptable range of attitude deviation characteristics. If the modulus lengths of the position deviation characteristic, torsional deviation characteristic, and deformation deviation characteristic are all less than the attitude deviation tolerance threshold, it means that the current attitude deviation is already within the acceptable range. The setting of the attitude deviation tolerance threshold requires comprehensive consideration of factors such as the design requirements of the bridge, the accuracy requirements of the rotation, and actual engineering experience. For example, according to the design standards for bridges, the modulus length tolerance threshold for the position deviation characteristic is 5mm, the modulus length tolerance threshold for the torsional deviation characteristic is 0.5°, and the modulus length tolerance threshold for the deformation deviation characteristic is 2%. After calculating the modulus lengths of the position deviation characteristic, torsional deviation characteristic, and deformation deviation characteristic, they are compared with the corresponding tolerance thresholds respectively.

[0161] Step S553: ​​If the mold length is less than the posture deviation tolerance threshold, it is determined whether the change trend of the mold length is stable and less than the posture deviation tolerance threshold in the control process for a consecutive preset number of times.

[0162] Even if the modulus length of the current posture deviation feature is less than the posture deviation tolerance threshold, posture adjustment cannot be immediately considered complete. It is necessary to determine whether the trend of modulus length changes has stabilized over a predetermined number of consecutive control flows. The predetermined number of times is a fixed number determined based on actual conditions, such as 5 or 10. If the modulus lengths of the position deviation feature, torsion deviation feature, and deformation deviation feature remain below the posture deviation tolerance threshold over a predetermined number of consecutive control flows, and their fluctuations are minimal, it indicates that the posture deviation has stabilized and no longer experiences significant fluctuations.

[0163] Statistical analysis can be used to determine whether the mold length trend is stable. For example, the mean and standard deviation of the mold length over a predetermined number of consecutive control flows can be calculated. A small standard deviation indicates minimal fluctuations in the mold length and a stable trend. Furthermore, the mold length can be checked to see if it is consistently below the attitude deviation tolerance threshold during each control flow. If these two conditions are met, the mold length trend is considered stable and below the attitude deviation tolerance threshold.

[0164] Step S554: If the mold length is less than the posture deviation tolerance threshold and the change trend of the mold length tends to be stable and less than the posture deviation tolerance threshold in the control process for a consecutive preset number of times, it is confirmed that the posture deviation feature meets the posture adjustment completion condition.

[0165] When the moduli of the position deviation, torsion deviation, and deformation deviation are all less than the attitude deviation tolerance threshold, and the moduli remain stable and less than the attitude deviation tolerance threshold for a predetermined number of consecutive control flows, the bridge's rotation posture has achieved the desired effect and the attitude deviation characteristics meet the posture adjustment completion conditions. At this point, the bridge's rotation posture can be considered to be close to the target posture.

[0166] Step S555: In response to confirming that the posture adjustment completion condition is met, a stop instruction is sent to the multiple drive controllers to terminate the execution of the current posture adjustment strategy and maintain the current state of the rotation actuator until a new rotation control instruction is received.

[0167] Once the posture deviation characteristics are confirmed to meet the posture adjustment completion conditions, the current posture adjustment strategy must be terminated promptly. A stop command is sent to multiple drive controllers, causing them to cease controlling the drive units, thereby terminating the posture adjustment strategy. Simultaneously, the current state of the rotation actuators is maintained to ensure the bridge's rotational posture remains stable. This process continues until a new rotation control command is received, at which point further adjustments or operations to the bridge's rotational posture are made based on the new command.

[0168] Throughout the implementation of the intelligent bridge rotation posture adjustment method based on multi-source data fusion, each step is interconnected and synergistic. By acquiring a multi-source monitoring data set, a comprehensive and accurate understanding of various information related to the bridge rotation process is achieved. Temporal and spatial alignment and fusion of the multi-source monitoring data eliminates temporal and spatial differences, resulting in fused monitoring data that comprehensively reflects the coordinated changes in the bridge structure, environmental impacts, and actuator states. Attitude deviation characteristics are identified based on the fused monitoring data and the target attitude parameter set, providing a clear target for attitude adjustment. A posture adjustment strategy is generated based on the posture deviation characteristics and executed to perform real-time posture adjustment operations, forming a closed-loop control process that continuously optimizes and adjusts the bridge rotation posture. Ultimately, when the posture deviation characteristics meet the posture adjustment completion conditions, the posture adjustment strategy is terminated, ensuring that the bridge rotation posture achieves the desired effect.

[0169] It is understandable that the various algorithms involved in the above-mentioned introductions of the embodiments of the present invention, such as the attention mechanism algorithm, the feature extraction algorithm, etc., can all be learned from the relevant content in the prior art. In order to save space, they will not be expanded too much in the embodiments of the present invention. In addition, when implementing the scheme of the present invention, those skilled in the art can supplement the details according to the common knowledge in this field. For example, according to the common knowledge in this field, normalization can be used to eliminate the dimension conflict before feature fusion, interpolation can be used to eliminate the dimension difference, and the threshold value can be reasonably set based on historical data, experience or business scenario requirements. The model can be trained based on a general model training method, and the number of layers in the model structure can be set based on actual needs, the activation function can be selected, etc. The present invention will no longer provide redundant introductions to the overly detailed implementation process.

[0170] See also Figure 2 , Figure 2This is a schematic diagram of the structure of a system for intelligently adjusting the rotational posture of a bridge based on multi-source data fusion, provided in an embodiment of the present invention. The system is a computer system for implementing a method for intelligently adjusting the rotational posture of a bridge based on multi-source data fusion. The computer system includes at least a processor 101, a communication interface 102, and a memory 103. The processor 101, communication interface 102, and memory 103 may be connected via a bus or other means. The processor 101 (also known as the Central Processing Unit (CPU)) is the computing and control core of the computer system, capable of parsing various instructions within the computer system and processing various data within the computer system. The communication interface 102 may optionally include a standard wired interface or a wireless interface (such as Wi-Fi or a mobile communication interface), and may be used to send and receive data under the control of the processor 101. The communication interface 102 may also be used for data transmission and interaction within the computer system. The memory 103 is a storage device within the computer system, used to store programs and data. It is understood that the memory 103 herein may include both the computer system's built-in memory and, of course, the computer system's supported extended memory. The memory 103 provides a storage space for storing the operating system of the computer system, which may include but is not limited to: Android system, iOS system, Windows Phone system, etc., and the present invention is not limited to this.

[0171] In one embodiment, the processor 101 executes the intelligent adjustment method for bridge rotation posture based on multi-source data fusion provided in the above embodiment of the present invention by running the computer program in the memory 103 .

Claims

1. A method for intelligently adjusting the rotation posture of a bridge based on multi-source data fusion, characterized in that: include: Acquire a multi-source monitoring data set generated during the bridge rotation process, wherein the multi-source monitoring data set includes bridge structure response monitoring data, environmental impact factor monitoring data, and rotation actuator operation status monitoring data; Performing spatiotemporal alignment and fusion processing on the multi-source monitoring data set to generate fused monitoring data; Based on the fused monitoring data and a preset target posture parameter set, identifying posture deviation characteristics between the bridge rotation posture and the target posture; generating a posture adjustment strategy for the bridge rotation actuator according to the posture deviation characteristics; parsing the adjustment instruction set in the posture adjustment strategy into a control instruction sequence recognizable by the rotation actuator; Sending the control instruction sequence to multiple drive controllers corresponding to the rotary actuators; Controlling the corresponding drive units respectively by the plurality of drive controllers to execute the action type, action direction, action amplitude and action timing specified by the control instruction sequence; During the process of the driving unit performing the action, the multi-source monitoring data set is continuously acquired, and the steps of performing spatiotemporal alignment and fusion processing on the multi-source monitoring data set to generate fused monitoring data, identifying the posture deviation characteristics between the bridge rotation posture and the target posture based on the fused monitoring data and a preset target posture parameter set, generating a posture adjustment strategy for the bridge rotation actuator according to the posture deviation characteristics, and executing the posture adjustment strategy to perform real-time posture adjustment operations on the bridge rotation process are cyclically performed; After the posture deviation feature is identified each time, the modulus of the position deviation feature, the torsion deviation feature and the deformation deviation feature in the posture deviation feature is calculated; Determine whether the mold length is less than a preset posture deviation tolerance threshold; if the mold length is less than the posture deviation tolerance threshold, determine whether the change trend of the mold length is stable and less than the posture deviation tolerance threshold in the control process for a consecutive preset number of times; If the module length is less than the posture deviation tolerance threshold and the change trend of the module length tends to be stable and less than the posture deviation tolerance threshold in the control process for a consecutive preset number of times, it is determined that the posture deviation feature meets the posture adjustment completion condition; In response to confirming that the posture adjustment completion condition is met, a stop instruction is sent to the multiple drive controllers to terminate the execution of the current posture adjustment strategy and maintain the current state of the rotation actuator until a new rotation control instruction is received.

2. The method according to claim 1, characterized in that The performing spatiotemporal alignment and fusion processing on the multi-source monitoring data set to generate fused monitoring data includes: Performing data validity verification on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotation actuator operation status monitoring data, respectively, eliminating invalid data items to obtain a valid monitoring data set; Performing timestamp alignment processing on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotation actuator operation status monitoring data in the valid monitoring data set, respectively, so that the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotation actuator operation status monitoring data have a synchronous corresponding relationship in the time dimension; Performing spatial position registration processing on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotation actuator operation status monitoring data in the valid monitoring data set after the timestamp alignment processing, so that the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotation actuator operation status monitoring data have a unified spatial reference benchmark in the spatial dimension; On the basis of completing the timestamp alignment processing and the spatial position registration processing, the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data are input into a preset data fusion model, and the data fusion model is used to perform feature-level fusion operations on the input types of monitoring data to output the fused monitoring data; the fused monitoring data is a data set that eliminates time and space differences and comprehensively reflects the coordinated change characteristics of the bridge structure status, environmental impact and actuator status.

3. The method according to claim 2, characterized in that The identifying, based on the fused monitoring data and a preset target posture parameter set, a posture deviation feature between the bridge rotation posture and the target posture, includes: Extracting a current posture feature set representing the current spatial position and posture of the bridge from the fused monitoring data; the current posture feature set at least includes a three-dimensional coordinate offset of a key control point of the bridge, an overall torsion angle of the bridge, and a deformation degree of a target cross section of the bridge; Acquire a target posture feature set corresponding to the current posture feature set from the preset target posture parameter set; the target posture feature set includes at least a target three-dimensional coordinate offset of a key control point of the bridge, a target torsion angle of the entire bridge, and a target deformation degree of a target cross section of the bridge; Performing a difference comparison process on the three-dimensional coordinate offset of the key control point of the bridge in the current posture feature set and the target three-dimensional coordinate offset of the key control point of the bridge in the target posture feature set to generate a position deviation feature; Performing a difference comparison process on the overall torsion angle of the bridge in the current posture feature set and the overall target torsion angle of the bridge in the target posture feature set to generate a torsion deviation feature; Performing a difference comparison process on the deformation degree of the target bridge section in the current posture feature set and the target deformation degree of the target bridge section in the target posture feature set to generate a deformation deviation feature; The position deviation feature, the torsion deviation feature and the deformation deviation feature are fused to form the posture deviation feature; the posture deviation feature is used to quantitatively describe the comprehensive deviation degree between the bridge rotation posture and the target posture in terms of spatial position, overall torsion and local deformation.

4. The method according to claim 3, characterized in that Generating a posture adjustment strategy for the bridge rotation actuator according to the posture deviation characteristics includes: Inputting the posture deviation feature into a preset posture adjustment strategy generation model, and analyzing the deviation degree and change trend of the position deviation feature, the torsion deviation feature, and the deformation deviation feature in the posture deviation feature through the posture adjustment strategy generation model; Based on the bridge structure mechanics constraint rules and the rotation actuator operation constraint rules built into the posture adjustment strategy generation model, the effect prediction of different adjustment action combinations on eliminating the position deviation characteristics, the torsion deviation characteristics, and the deformation deviation characteristics is evaluated; Based on the effect prediction, and under the premise of satisfying the bridge structure mechanics constraint rules and the rotation actuator operation constraint rules, generating an adjustment instruction combination that promotes optimal coordinated reduction of the position deviation characteristic, the torsional deviation characteristic, and the deformation deviation characteristic; The adjustment instructions are combined and encapsulated into the posture adjustment strategy; the posture adjustment strategy includes an instruction set for the action type, action direction, action amplitude and action timing of different driving units in the rotation actuator.

5. The method according to claim 2, characterized in that The performing timestamp alignment processing on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotation actuator operation status monitoring data in the effective monitoring data set respectively includes: Obtaining, from the valid monitoring data set, a first timestamp sequence corresponding to the bridge structure response monitoring data, a second timestamp sequence corresponding to the environmental impact factor monitoring data, and a third timestamp sequence corresponding to the rotary actuator operation status monitoring data; performing time resampling processing on the first timestamp sequence, the second timestamp sequence, and the third timestamp sequence at a preset reference time interval to generate a first resampled data sequence of the bridge structure response monitoring data, a second resampled data sequence of the environmental impact factor monitoring data, and a third resampled data sequence of the swivel actuator operation status monitoring data, respectively; performing time synchronization processing on the first resampled data sequence, the second resampled data sequence, and the third resampled data sequence to ensure that corresponding valid data points exist in the first resampled data sequence, the second resampled data sequence, and the third resampled data sequence at the same reference time point; The first resampled data sequence, the second resampled data sequence and the third resampled data sequence that have completed time synchronization processing are respectively used as the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotating actuator operation status monitoring data after time stamp alignment processing.

6. The method according to claim 2, characterized in that The performing of spatial position registration processing on the bridge structure response monitoring data, the environmental impact factor monitoring data, and the rotation actuator operation status monitoring data in the valid monitoring data set after the timestamp alignment processing includes: Establish a unified spatial coordinate system as the registration reference coordinate system; Identifying a first spatial reference system corresponding to the bridge structure response monitoring data, a second spatial reference system corresponding to the environmental impact factor monitoring data, and a third spatial reference system corresponding to the rotation actuator operation status monitoring data; respectively calculating spatial transformation parameters between the first spatial reference system, the second spatial reference system, the third spatial reference system, and the registration reference coordinate system; Based on the spatial conversion parameters, the bridge structure response monitoring data after the timestamp alignment processing is converted from the first spatial reference system to the registration reference coordinate system to obtain first registered monitoring data; Based on the spatial conversion parameters, the environmental influencing factor monitoring data after the timestamp alignment processing is converted from the second spatial reference system to the registration reference coordinate system to obtain second registered monitoring data; Based on the spatial conversion parameters, the operating state monitoring data of the rotary actuator after the timestamp alignment processing is converted from the third spatial reference system to the registration reference coordinate system to obtain third registration monitoring data; The first registration monitoring data, the second registration monitoring data and the third registration monitoring data together constitute the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotation actuator operation status monitoring data after spatial position registration processing.

7. The method according to claim 2, characterized in that The data fusion model is a feature fusion network based on an attention mechanism; the data fusion model performs a feature-level fusion operation on each type of input monitoring data and outputs the fused monitoring data, including: Inputting the bridge structure response monitoring data, the environmental impact factor monitoring data and the rotation actuator operation status monitoring data after spatial position registration processing into different feature extraction branches of the feature fusion network based on the attention mechanism respectively; Through each feature extraction branch, the deep feature expression of the corresponding type of monitoring data is extracted respectively; The deep feature expressions extracted by each feature extraction branch are input into the attention allocation module; In the attention allocation module, based on the inherent correlation between the state of the current bridge rotation stage and the deep feature expressions of the various types of monitoring data, the attention weight coefficients to be allocated to the deep feature expressions of the bridge structure response monitoring data, the deep feature expressions of the environmental impact factor monitoring data, and the deep feature expressions of the rotation actuator operation status monitoring data are calculated; Based on the calculated attention weight coefficient, a weighted fusion process is performed on the deep feature expression of the bridge structure response monitoring data, the deep feature expression of the environmental impact factor monitoring data, and the deep feature expression of the rotary actuator operation status monitoring data; The feature expression after weighted fusion is subjected to feature dimensionality reduction and integration operations to generate the fused monitoring data; the fused monitoring data is a unified feature vector sequence of comprehensive structural response, environmental impact and mechanism status information.

8. The method according to claim 4, characterized in that The posture adjustment strategy generation model is a decision-making model based on a combination of rules and learning. In the posture adjustment strategy generation model, multiple candidate adjustment action combinations are preset. Based on the built-in bridge structure mechanics constraint rules and rotation actuator operation constraint rules of the posture adjustment strategy generation model, the effect prediction of different adjustment action combinations on eliminating the position deviation characteristics, the torsion deviation characteristics, and the deformation deviation characteristics is evaluated, including: For each candidate combination of adjustment actions, based on the current values ​​of the position deviation feature, the torsion deviation feature, and the deformation deviation feature in the posture deviation feature, combined with the bridge structure mechanics constraint rules, predict the stress change distribution and potential risk areas of the bridge structure caused by executing the candidate combination of adjustment actions; Based on the current values ​​of the position deviation feature, the torsion deviation feature, and the deformation deviation feature in the posture deviation feature, combined with the operation constraint rules of the rotation actuator, the output power, operating speed, and duration required by each drive unit when executing the candidate adjustment action combination are predicted, and whether the required output power, operating speed, and duration are exceeded. The safe operation threshold of the actuator is evaluated; Calculate the comprehensive effect score of the candidate adjustment action combination on reducing the position deviation characteristic, the torsional deviation characteristic, and the deformation deviation characteristic by combining the predicted stress change distribution of the bridge structure, the potential risk area assessment results, and the output power, operating speed, and duration assessment results of each drive unit; Based on the historical adjustment effect feedback data learned by the posture adjustment strategy generation model, the comprehensive effect score is revised; The generating of the adjustment instruction combination that promotes the optimal coordinated reduction of the position deviation feature, the torsion deviation feature, and the deformation deviation feature is to select the candidate adjustment action combination with the highest corrected comprehensive effect score.

9. A computer system, characterized in that: include: a memory storing a computer program; A processor is used to load the computer program to implement the intelligent adjustment method of bridge rotation posture based on multi-source data fusion as described in any one of claims 1 to 8.

Images