Multi-task cooperative intelligent progress management method based on artificial intelligence

Abstract

The invention discloses a multi-task cooperative intelligent progress management method based on artificial intelligence, and relates to the technical field of bridge engineering data management and analysis. The method comprises the following steps: carrying out the first adjustment of the bridge measurement radar; performing bridge engineering intelligent progress first management; carrying out second adjustment on the bridge measurement radar; performing bridge engineering intelligent progress second management; and self-updating the intelligent progress management model of the bridge engineering. According to the invention, single radar parameter evaluation is carried out and then error coefficient calculation and first adjustment are carried out; performing radar composite influence evaluation, model correction and second adjustment; and finally, model self-updating is retrained through actual and target time error data, so that the effect of improving the robustness of collaborative progress management in bridge construction is achieved, and the problem of insufficient robustness of collaborative progress management in bridge construction in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of bridge engineering data management and analysis technology, and in particular to a multi-task collaborative intelligent progress management method based on artificial intelligence. Background Technology

[0002] As infrastructure construction becomes increasingly complex and regionalized, multi-bridge coordinated construction has become the norm. These projects are characterized by large spatial spans, overlapping construction phases, and intensive resource sharing. Their progress management requires simultaneous coordination of equipment scheduling, environmental adaptation, and work process connections across multiple bridges, making traditional management models inadequate to meet the demands.

[0003] The mainstream technologies in the field of current project schedule management include map remote sensing monitoring, Internet of Things monitoring, and cloud computing collaborative platforms. They have been applied to some extent in single bridge or simple multi-task scenarios, but their limitations are significant in multi-bridge linkage scenarios. They mainly rely on static planning driven by human experience and single-factor monitoring. However, in multi-bridge linkage scenarios, factors such as regional weather differences, equipment sharing conflicts, and differences in sensitivity of construction stages are intertwined, leading to schedule misalignment.

[0004] For example, a project construction progress management system and method with publication number CN119477200A includes a computer terminal and multiple mobile terminals. The computer terminal includes a construction plan preparation module, a first construction record module, and a progress dashboard module. The mobile terminals include a second construction record module. The construction plan preparation module allows online editing of task plans, generating a list of planned tasks and a Gantt chart. The second construction record module is used for closed-loop updating of construction tasks: construction workers select construction tasks from the task plan, update the task progress multiple times by multiple people, and enter construction records to generate supporting materials. Unfinished tasks are connected sequentially at their actual progress points in the Gantt chart to form a leading line, indicating the progress deviation.

[0005] For example, the announcement number CN117649129B describes a multi-agent collaborative system and strategy method applicable to industrial digitalization. This includes a large model system for implementing central management functions; a tool library for providing basic functions; and a functional model library for providing analytical operations. Based on the input task, the large model system acquires and understands the user request corresponding to that task, thereby decomposing the task into multiple objectives and several sub-tasks corresponding to each objective. It then schedules the tools corresponding to each sub-task from the tool library and the functional models corresponding to each sub-task from the functional model library to complete the combined calculation or decision-making of data and models, and forms a strategy scheme corresponding to the input task.

[0006] There is an urgent need for an intelligent progress management method that can dynamically adapt to the sensitivity of equipment at multiple stages and handle the coupling effect of multiple factors in order to solve the problem of progress loss of control in multi-bridge linkage scenarios.

[0007] In existing technologies, IoT monitoring technology, relying on single-factor sensing, struggles to handle the coupling effects of multiple factors. It can only collect data on a single factor and cannot analyze the combined impact of multiple factors on equipment. Furthermore, sensors only output raw data and lack integrated fault prediction models for multi-factor coupling, preventing artificial intelligence from providing early warnings of equipment failures and allowing only passive responses after data interruptions. This results in insufficient robustness for collaborative progress management in bridge construction. Summary of the Invention

[0008] This application provides an AI-based multi-task collaborative intelligent progress management method, which solves the problem of insufficient robustness of collaborative progress management in bridge construction in the prior art, and achieves the effect of improving the robustness of collaborative progress management in bridge construction.

[0009] This application provides an artificial intelligence-based multi-task collaborative intelligent progress management method, comprising the following steps: evaluating the impact parameters output by a bridge measurement radar; performing a first adjustment of the bridge measurement radar based on the impact evaluation; performing a first intelligent progress management of the bridge project based on the first adjustment result of the bridge measurement radar to obtain an intelligent progress management model for the bridge project; evaluating the combined influence of the bridge measurement radars; performing a second adjustment of the bridge measurement radar based on the combined influence evaluation to adjust the intelligent progress management model for the bridge project; performing a second intelligent progress management of the bridge project based on the second adjustment result of the bridge measurement radar; and triggering a self-update of the intelligent progress management model for the bridge project after a predefined model update time.

[0010] Furthermore, the impact parameters of the bridge measurement radar output are evaluated, specifically including: acquiring the bridge radar signal-to-noise ratio (SNR) through the radar's built-in digital signal processing module; measuring the radar's transmit and receive power using a power meter and calculating the bridge radar rain attenuation loss using free-space path loss; analyzing the peak power of clutter in the echo signal spectrum using the radar's built-in fast Fourier transform module to obtain the bridge radar echo clutter power; fitting the linear relationship between QLC and bridge radar SNR, bridge radar rain attenuation loss, and bridge radar echo clutter power using the least squares method in multivariate regression analysis under different rainfall intensities; fitting the linear relationship between the final measurement error coefficient of the bridge radar and the reciprocal of the bridge radar SNR, bridge radar rain attenuation loss, and bridge radar echo clutter power using the least squares method under different rainfall intensities to obtain the corresponding weighting factors; and performing a comprehensive coupling analysis of the bridge radar SNR, bridge radar rain attenuation loss, and bridge radar echo clutter power with the corresponding weighting factors to obtain the final measurement error coefficient of the bridge radar.

[0011] Furthermore, based on the impact assessment of the bridge measurement radar output, the first adjustment of the bridge measurement radar is carried out, specifically including: if the final measurement error coefficient of the bridge radar is less than the final measurement error threshold of the bridge radar, no adjustment is made; if the final measurement error coefficient of the bridge radar is equal to or greater than the final measurement error threshold of the bridge radar, the bridge section is switched to the low frequency band, the equalizer weight is adjusted by the least mean square algorithm, and the over-threshold duration when the corresponding output frequency error coefficient of the bridge radar is equal to or greater than the final measurement error threshold of the bridge radar is recorded as the final measurement over-threshold duration of the bridge radar; if the final measurement over-threshold duration of the bridge radar is less than the maximum value of the final measurement over-threshold duration of the bridge radar, no adjustment is made; if the final measurement over-threshold duration of the bridge radar is equal to or greater than the first value of the final measurement over-threshold duration of the bridge radar, the receiving power of the corresponding bridge radar is increased, and intelligent progress management of the bridge project is carried out; if the final measurement over-threshold duration of the bridge radar is equal to or greater than the second value of the final measurement over-threshold duration of the bridge radar, the receiving power of the corresponding bridge radar is increased, intelligent progress management of the bridge project is carried out, and regional early warnings are issued to relevant personnel.

[0012] Furthermore, based on the initial adjustment results of the bridge measurement radar, intelligent progress management of the bridge project is implemented. The specific process is as follows: Input parameters for the intelligent progress management model of the bridge project are extracted from historical bridge construction data, and the corresponding progress delay times are labeled. The input parameters of the intelligent progress management model include the final measurement error coefficient of the bridge radar, the magnitude of the error deviation threshold, the duration of the final measurement exceeding the threshold, the degree of increase in the receiving power of the corresponding bridge radar, the radar frequency band adjustment indicator, the priority indicator of the bridge construction task, and the redundancy of bridge construction scheduling resources. The input parameters of the intelligent progress management model are standardized, and the relationship between the input parameters and the delay time is trained using a predefined intelligent progress management model labeled with the progress delay time, resulting in the intelligent progress management model of the bridge project. The predicted delay time is output through the intelligent progress management model of the bridge project, and the predicted delay time is notified to relevant personnel.

[0013] Furthermore, an assessment of the mutual composite influence of bridge measurement radars is conducted, specifically including: defining the proportion of the frequency range overlap between bridge measurement radars to the total frequency band as the bridge measurement radar frequency band overlap; measuring the interference power at the target radar receiver using a spectrum analyzer, and calculating the bridge measurement radar interference power intensity coefficient by combining the transmit power and distance; measuring the interference power at the target bridge measurement radar receiver using a spectrum analyzer, and obtaining the bridge measurement radar interference coefficient by comparing the bridge measurement radar power with the original noise power and then performing logarithmic analysis; assessing the risk of bridge measurement radar delays due to conflicts based on the average probability of bridge measurement radars simultaneously requesting resources according to historical bridge construction data, and combining the resource buffer capacity with the total capacity of shared resources, thus obtaining the radar mission resource conflict probability coefficient; and coupling the bridge measurement radar frequency band overlap, bridge measurement radar interference power intensity coefficient, and bridge measurement radar interference coefficient with the corresponding weighting factors, and then coupling them with the radar mission resource conflict probability coefficient to obtain the bridge measurement radar mutual composite coefficient.

[0014] Furthermore, based on the composite impact assessment of bridge measurement radar, a second adjustment to the bridge measurement radar is performed. Specifically, if the cross-composite coefficient of the bridge measurement radar is less than or equal to the threshold of the cross-composite coefficient of the bridge measurement radar, then no adjustment is made to the intelligent progress management model of the bridge project; if the cross-composite coefficient of the bridge measurement radar is greater than the threshold of the cross-composite coefficient of the bridge measurement radar, then the intelligent progress management model of the bridge project is adjusted.

[0015] Furthermore, the intelligent progress management model for bridge engineering is adjusted, specifically including: comparing and analyzing the cross-composite coefficient of bridge measurement radar with the threshold value of the cross-composite coefficient of bridge measurement radar to obtain the cross-composite difference coefficient of bridge measurement radar; performing a coupling analysis of the cross-composite difference coefficient of bridge measurement radar with the priority indicator value of bridge construction tasks and rounding it down to obtain the corrected task priority indicator value; updating the priority indicator value of bridge construction tasks in the intelligent progress management model of bridge engineering with the corrected task priority indicator value, and outputting a new predicted delay time through the intelligent progress management model of bridge engineering, which is recorded as the first predicted delay time; and performing a coupling analysis of the cross-composite difference coefficient of bridge measurement radar with the first predicted delay time and rounding it down to obtain the second predicted delay time.

[0016] Furthermore, the second adjustment of the bridge measurement radar based on the composite impact assessment of the bridge measurement radar also includes: comparing and analyzing the number of overlapping bridge construction tasks with the total number of bridge construction tasks, then performing coupling analysis and summing the results with the corresponding composite difference coefficients of the bridge measurement radar to obtain the bridge task interference coupling degree; if the bridge task interference coupling degree is less than the first threshold, no multi-task allocation is required; if the bridge task interference coupling degree is equal to or greater than the first threshold and less than the second threshold, the resource redundancy corresponding to the shared resource tasks will be used; if the bridge task interference coupling degree is equal to or greater than the second threshold, relevant personnel will be notified to increase the resource redundancy separately according to the corrected task priority indicator value.

[0017] Furthermore, based on the second adjustment results of the bridge measurement radar, a second intelligent progress management system for the bridge project is implemented. Specifically, this includes: if the second predicted delay time is less than the predicted delay time threshold, then the second intelligent progress management system for the bridge project is not implemented; if the second predicted delay time is equal to or greater than the predicted delay time threshold, then the area corresponding to the second predicted delay time is located, and the bridge radar frequency band within the area is gradually reduced to the predefined minimum bridge radar operating frequency band; if the second predicted delay time is less than the predicted delay time threshold during the gradual reduction of the bridge radar frequency band, then the reduction of the bridge radar frequency band is stopped; if the bridge radar frequency band within the area is reduced to the predefined minimum bridge radar operating frequency band and the second predicted delay time is equal to or greater than the predicted delay time threshold, then the duration of the composite interference is recorded; if the duration of the composite interference exceeds the composite interference duration threshold, then an area warning is issued to relevant personnel.

[0018] Furthermore, the bridge engineering intelligent progress management model is self-updated after the predefined model update time is triggered. Specifically, this includes: acquiring the actual completion time data and target completion time data of the bridge construction task, analyzing the error data between the actual completion time data and the target completion time data, recording it as the actual progress delay time data, merging the actual progress delay time data and the second predicted delay time with the historical progress delay time to form a new training set, and retraining and updating the bridge engineering intelligent progress management model with the new training set.

[0019] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0020] 1. By evaluating single radar parameters, calculating error coefficients and performing the first adjustment; evaluating radar composite effects, correcting the model and performing the second adjustment; and finally retraining the model for self-updating using actual and target time error data, the robustness of collaborative progress management in bridge construction has been improved, solving the problem of insufficient robustness of collaborative progress management in bridge construction in existing technologies.

[0021] 2. By comprehensively considering three core parameters—bridge radar signal quality, signal attenuation due to rainfall, and electromagnetic interference intensity—a model trained on historical data quantifies the impact of multiple factors on measurement errors. This accurately identifies the combined effects of high interference and heavy rain attenuation, resolving interference blind spots that cannot be identified by single-factor monitoring. Based on the evaluation results, a three-tiered adjustment strategy is automatically triggered: frequency reduction to combat rain attenuation, compensation for rain attenuation distortion to reduce clutter overlap, and enhancement of the signal-to-noise ratio to reduce signal quality impact. This forms an "evaluation-adjustment-feedback" closed loop, directly reducing the impact of interference on measurements and shortening schedule delays caused by errors. Furthermore, by combining task priority, resource redundancy, and interference coupling, the schedule prediction model is dynamically corrected and resources are allocated, thereby improving the accuracy of resource allocation and management in bridge construction under complex influences.

[0022] 3. By reflecting frequency resource competition, interference energy intensity, signal quality degradation, and assessing resource competition risks, and coupling the effects of these four factors through a linear regression model, the superimposed impact of multiple radar electromagnetic interference and resource competition on measurement errors is accurately quantified. Based on the comparison results of thresholds, the intelligent progress management model for bridge engineering is dynamically adjusted. This is achieved by correcting task priorities and predicting delay times, prioritizing resources for critical tasks, reducing progress delays caused by cross-task interference, and further combining the number of overlapping tasks with the probability of resource conflicts to increase the redundancy of shared resources or allocate dedicated resources separately for high-priority tasks. This ensures the controllability of resource allocation management under the combined effects.

[0023] 4. When the second predicted delay time exceeds the threshold, the system automatically locates the delay area and gradually reduces the operating frequency of the bridge radar in that area. This gradual adjustment avoids signal instability caused by a large one-time change. At the same time, the system monitors the changes in delay time in real time and dynamically judges the adjustment effect. If the threshold cannot be met even after the frequency band is reduced to the lowest level, the system records the duration of the composite interference and issues an area warning, prompting manual intervention. Furthermore, the system combines the actual construction results, model prediction output, and historical data to form a new training set covering "low-risk accuracy - high-risk deviation". Using the progress delay time as a label, the system updates the training model through the XGBoost regressor, optimizes the mapping relationship between input features and delay time, generates new feature weights and hyperparameters, and stores them in the configuration library. This allows subsequent predictions to use the latest parameters by default, reuse historical experience, continuously correct model deviations, improve the adaptability to complex interference in multi-task collaborative scenarios, and ultimately enhance the accuracy of progress prediction and the effectiveness of adjustment. Attached Figure Description

[0024] Figure 1 A flowchart of an AI-based multi-task collaborative intelligent progress management method provided in this application embodiment. Detailed Implementation

[0025] This application provides an AI-based multi-task collaborative intelligent progress management method, which solves the problem of insufficient robustness in collaborative progress management during bridge construction in existing technologies. The method improves the robustness of collaborative progress management during bridge construction by providing a multi-task collaborative intelligent progress management method based on artificial intelligence. This method involves single radar parameter evaluation followed by error coefficient calculation and first adjustment; radar composite impact evaluation followed by model correction and second adjustment; and finally, model self-updating through retraining using actual and target time error data.

[0026] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0027] like Figure 1 The diagram shown is a flowchart of a multi-task collaborative intelligent progress management method based on artificial intelligence provided in this application embodiment. The method includes the following steps: evaluating the impact parameters output by a bridge measurement radar; performing a first adjustment of the bridge measurement radar based on the impact evaluation; performing first intelligent progress management of the bridge project based on the first adjustment result of the bridge measurement radar to obtain an intelligent progress management model for the bridge project; evaluating the combined influence of the bridge measurement radars; performing a second adjustment of the bridge measurement radar based on the combined influence evaluation to adjust the intelligent progress management model for the bridge project; performing second intelligent progress management of the bridge project based on the second adjustment result of the bridge measurement radar; and triggering a self-update of the intelligent progress management model for the bridge project after a predefined model update time.

[0028] In this embodiment, existing IoT monitoring technologies, relying on single-factor sensing, struggle to handle multi-factor coupling effects. They can only collect data on a single factor and cannot analyze the combined impact of multiple factors on equipment. For example, in a project where multiple adjacent bridges are under construction simultaneously, heavy rain causes radar signal attenuation, and the simultaneous operation of multiple bridge radars triggers electromagnetic interference. The platform mistakenly identifies radar data from bridge A as data from bridge B, leading to misjudgment of progress. Furthermore, sensors can only output raw data (such as strain value fluctuations) and lack integrated fault prediction models for multi-factor coupling (e.g., strain gauge zero-bias drift caused by "rainfall + vibration + temperature difference"). This prevents AI systems from providing early warnings of equipment failure, allowing only passive responses after data interruption (e.g., manual drying takes 2 hours per instance).

[0029] Millimeter-wave radar and ground-penetrating radar, with their characteristics of non-contact measurement, high precision, and all-weather operation, are mainly used in bridge construction for three major scenarios: structural deformation monitoring, position deviation calibration, and hidden engineering detection.

[0030] The configuration of radar needs to be decided comprehensively based on the bridge scale, accuracy requirements, and budget. For long-span bridges, such as cable-stayed bridges and suspension bridges with a main span of more than 500m, the structure is complex and the accuracy requirements are high. Therefore, each bridge needs to be equipped with a dedicated millimeter-wave radar to cover the entire bridge towers, beams, and cables.

[0031] Bridges with special geological conditions, such as those crossing earthquake fault zones or soft soil areas, are at high risk of foundation settlement. They need to be equipped with ground-penetrating radar embedded in the pile foundation or abutment to monitor deep soil displacement in real time and prevent sudden collapse.

[0032] Furthermore, the impact parameters of the bridge measurement radar output are evaluated, specifically including: calculating the ratio of received signal power (Pr) to noise power (N) using the radar's built-in digital signal processing module to obtain the bridge measurement radar signal-to-noise ratio (SNR = Pr / N); measuring the transmit power (Pt) and receive power (Pr) with a power meter, and calculating the bridge radar rain attenuation loss (Lrain = PLmeasured - PLfs, where PLmeasured = 20log10(d) + 20log10(f) + 32.45) using the radar's built-in fast Fourier transform module to analyze the peak power of clutter in the echo signal spectrum to obtain the bridge radar echo clutter power (Pclutter); and measuring QLC, SNR, Rain, and P under different rainfall intensities R. The value of clutter was obtained by fitting the linear relationship between QLC and SNR, Lrain, and Pclutter values ​​using the least squares method in multiple regression analysis, yielding coefficients α, β, and γ. The final measurement error coefficient of the bridge radar was obtained by comprehensively coupling the bridge radar signal-to-noise ratio, bridge radar rain attenuation loss, and bridge radar echo clutter power with their respective weighting factors. QLC represents the final measurement error coefficient of the bridge radar, QLC=α*(1 / SNR)+β*Lrain+γ*Pclutter; where α represents the weighting factor of the reciprocal of the bridge radar signal-to-noise ratio on the final measurement error, β represents the weighting factor of the bridge radar rain attenuation loss on the final measurement error, and γ represents the weighting factor of the bridge radar echo clutter power on the final measurement error. These weighting factors reflect the degree of influence of each parameter on the error.

[0033] In this embodiment, the impact on the tower verticality monitoring radar varies under different rainfall conditions. During light rain, electromagnetic waves penetrate the rain curtain, resulting in a verticality measurement accuracy of ±0.1 mm / m. During moderate rain, raindrop scattering increases radar echo clutter, leading to misjudgments of verticality deviation. During heavy rain, water accumulates on the radome, causing electromagnetic wave refraction and distortion, rendering it unusable for high-precision verticality control.

[0034] Implement multiple linear regression using Python's scikit-learn library and fit an error model. Example code is shown below:

[0035] Bridge measurement radar includes tower verticality monitoring radar, millimeter-wave radar, and ground-penetrating radar.

[0036] import numpy as np

[0037] from sklearn.linear_model import LinearRegression

[0038] #Construct input features (including intercept term)

[0039] X=np.hstack([np.ones_like(R),R,SNR.reshape(-1,1),L_rain.reshape(-1,1),P_clutter.reshape(-1,1)])

[0040] y = delta

[0041] #Training a linear regression model

[0042] model = LinearRegression()

[0043] model.fit(X, y)

[0044] # Output weighting factors (intercept term + coefficients of each parameter)

[0045] print("Weight factors (intercept, R, SNR, L_rain, P_clutter):", model.coef_)

[0046] Example of the output result of the above code:

[0047] Weighting factors (intercept, R, SNR, L_rain, P_clutter): [0.1, 0.02, -0.05, 0.03, 0.04]

[0048] The above output results represent the contribution weight of each parameter to the error.

[0049] Furthermore, based on the impact assessment of the bridge measurement radar output, the first adjustment of the bridge measurement radar is performed, specifically including: if the final measurement error coefficient of the bridge radar is less than the final measurement error threshold of the bridge radar, no adjustment is made; if the final measurement error coefficient of the bridge radar is equal to or greater than the final measurement error threshold of the bridge radar, the radar will be switched to a low-frequency band, such as the S-band, to reduce rain by utilizing the low-frequency anti-attenuation characteristics; the equalizer weights are adjusted using the least mean square algorithm to compensate for frequency response distortion caused by rain attenuation and suppress the overlap of clutter and target signals; the final measurement error coefficient of the bridge radar is recorded as equal to or greater than the bridge radar output frequency error threshold of the bridge radar. The duration of exceeding the final measurement error threshold is denoted as the final measurement over-threshold duration of the bridge radar. If the final measurement over-threshold duration of the bridge radar is less than the maximum value of the final measurement over-threshold duration of the bridge radar, no adjustment is made. If the final measurement over-threshold duration of the bridge radar is equal to or greater than the maximum value of the final measurement over-threshold duration of the bridge radar, the receiving power of the corresponding bridge radar is increased to improve the SNR; intelligent progress management of the bridge project is implemented. If the final measurement over-threshold duration of the bridge radar is equal to or greater than the second value of the final measurement over-threshold duration of the bridge radar, the receiving power of the corresponding bridge radar is increased, intelligent progress management of the bridge project is implemented, and regional early warnings are issued to relevant personnel.

[0050] In this embodiment, the core device for performing adaptive equalization filtering in the signal processing link of the bridge radar is a digital filter in the digital signal processing module or a field-programmable gate array, such as a finite impulse response (FIR) filter. Example code demonstrating the least mean square algorithm for adjusting equalizer weights using the `adaptive_equalization` function is as follows:

[0051] def adaptive_equalization(self, signal): #LMS adaptive equilibration

[0052] step_size = 0.01 # step size μ

[0053] delayed_signal = np.concatenate([np.zeros(4), signal]) # Tap delay

[0054] equalized=np.convolve(delayed_signal, self.eq_weights, mode='valid')#Filter output

[0055] error = np.mean(signal) - np.mean(equalized) # Error calculation (simplified version)

[0056] self.eq_weights += step_size * error * np.flip(delayed_signal[:5]) # Weight update

[0057] return equalized, error

[0058] The second value of the final measurement over-threshold duration of the bridge radar is greater than the first value of the final measurement over-threshold duration of the bridge radar.

[0059] Furthermore, by extracting the bridge radar final measurement error coefficient, error deviation threshold amplitude, bridge radar final measurement over-threshold duration, increase in receiving power of the corresponding bridge radar, radar frequency band adjustment indicator, bridge construction task priority indicator, and redundancy of bridge construction scheduling resources from historical bridge construction data, the corresponding schedule delay time ΔT is marked, and the schedule delay time is the target variable; the constraints of the input vector X are as follows:

[0060] X = [QLC, QLC-YZ, Tover, Padjust, fswitch, R, Ptask, Aavail], where QLC represents the final measurement error coefficient of the bridge radar, QLC-YZ represents the magnitude of the error deviation from the threshold, QLC-YZ represents the difference between the first value and the second value when the final measurement over-threshold duration is equal to or greater than the first value, and Tover represents the final measurement over-threshold duration, reflecting the persistence of the error, e.g., Tover = 90. Minutes represent the duration of the error over 1.5 hours. Padjust represents the increase in the receiving power of the corresponding bridge radar, indicating the radar's active compensation measures. For example, Padjust = 10dBm means the receiving power is increased by 10dB. fswitch represents the radar frequency band adjustment indicator, reflecting the radar's frequency band adjustment strategy. For example, fswitch = 1 means switching to a lower frequency band. R reflects the maximum intensity of local rainfall in the corresponding area. For example, R = 50mm / h indicates heavy rainfall. Ptask represents the priority indicator of the bridge construction task, indicating the importance of the bridge construction task. For example, Ptask = 10 indicates the highest priority task. Aavail represents the redundancy of bridge construction scheduling resources. For example, Aavail = 10 indicates sufficient construction scheduling resources. Ptask and Aavail are obtained from expert prior knowledge.

[0061] Standardize the model input parameters, eliminate units and scale the data to 1-10; standardize continuous parameters (such as R), eliminate units and scale the data to 1-10; and perform one-hot encoding on categorical parameters (such as fswitch) (e.g., [0, 1] represents switching).

[0062] The bridge engineering intelligent progress management model outputs the predicted delay time and notifies relevant personnel of the predicted delay time.

[0063] In this embodiment, an XGBoost regressor is used to train the relationship between input parameters and delay time, using the schedule delay time as the label.

[0064] Model validation: Evaluate the model accuracy through cross-validation.

[0065] The model output is a quantified value of the schedule impact ΔT^(predicted delay time, in hours) or P^delay(predicted delay probability, in range [0, 1]).

[0066] Here is a code example using Python's XGBoost:

[0067] import pandas as pd

[0068] import numpy as np

[0069] from xgboost import XGBRegressor

[0070] from sklearn.model_selection import train_test_split

[0071] from sklearn.metrics import mean_squared_error

[0072] #Simulated Data Generation: Assuming a historical dataset (example): Input parameters and corresponding delay times ΔT

[0073] data = {

[0074] 'E': [1.2, 1.5, 2.0, 1.8, 1.3], # Error coefficient (E≥1 indicates exceeding the threshold)

[0075] 'E-E_th': [0.2, 0.5, 1.0, 0.8, 0.3], # Error threshold (assuming E_th = 1)

[0076] 'Tover': [30, 60, 90, 45, 20], # Duration of overthreshold (minutes)

[0077] 'Padjust': [10, 15, 20, 12, 8], # Receive power adjustment (dBm)

[0078] 'fswitch':[1, 1, 0, 1, 0], # Band switching flag (1 = switch)

[0079] 'R': [20, 50, 35, 40, 15], #Rainfall intensity (mm / h)

[0080] 'Ptask': [8, 10, 7, 9, 6], #Task priority (0-10)

[0081] 'Aavail': [5, 3, 6, 4, 7], #Available resources (units)

[0082] 'ΔT':[2.5, 5.0, 8.0, 6.0, 3.0]#Actual delay time (hours, tags)

[0083] }

[0084] df = pd.DataFrame(data)

[0085] #Feature and Label Separation

[0086] X = df.drop('ΔT', axis = 1)

[0087] y = df['ΔT']

[0088] #Dataset partitioning

[0089] X_train, X_test, y_train, y_test=train_test_split(X, y, test_size=0.2, random_state=42)

[0090] #Model Training

[0091] model = XGBRegressor(

[0092] n_estimators = 100, # Number of trees

[0093] learning_rate = 0.1, #learning rate

[0094] max_depth = 3, # Maximum depth of the tree

[0095] random_state = 42 )

[0097] model.fit(X_train, y_train)

[0098] #Model Validation

[0099] y_pred=model.predict(X_test)

[0100] mse=mean_squared_error(y_test,y_pred)

[0101] print(f"Model MSE:{mse:.2f}") # Output prediction error

[0102] #Prediction Example: Input new scene parameters (assuming a bridge radar exceeds the threshold)

[0103] new_data = pd.DataFrame({

[0104] 'E': [1.6], # Error coefficient 1.6 (60% over the threshold)

[0105] 'E-E_th':[0.6], # Error exceeding threshold level 0.6 (threshold 1, actual error 1.6)

[0106] 'Tover':

[75] , #Tover duration 75 minutes (1.25 hours)

[0107] 'Padjust':

[18] , #Receive power adjustment 18dBm

[0108] 'fswitch':[1], # Switch to low frequency band

[0109] 'R':

[45] , #Rainfall intensity 45mm / h (heavy rain)

[0110] 'Ptask':[9], #task priority 9 (high priority)

[0111] 'Aavail':[4]# Available resources: 4 units (relatively limited)

[0112] })

[0113] predicted_delay=model.predict(new_data)

[0114] print(f"Predicted delay time:{predicted_delay[0]:.2f} hours") # Output the predicted result (e.g., 3.8 hours).

[0115] Furthermore, an assessment is conducted on the combined influence of bridge measurement radars. Specifically, this includes calculating the proportion of the frequency range overlap between the two radars to the total frequency band (FT = Δf / ftotal). For example, millimeter-wave radar (76-78 GHz) and ground-penetrating radar (90-110 MHz) have no overlap, while the overlap between two millimeter-wave radars (76-78 GHz and 77-79 GHz) is (78-77) / (79-76) = 33%. FT represents the frequency band overlap of the bridge measurement radars. The interference power intensity coefficient (GRD) is calculated by combining the transmitted power (Pt) and distance (d), (GRD = Pt*Gt / Gr*(λ / 4πd)). 2 Where GRD represents the interference power intensity coefficient, which is dimensionless and only the numerical value is retained in subsequent calculations; Gt / Gr is the transceiver antenna gain; and λ is the wavelength. The interference power (Pf) at the target radar receiver is measured using a spectrum analyzer, and the SNR drop caused by the interference is calculated as ΔSNR = 10lg(Pinterf / N0), where N0 is the original noise power. ΔSNR represents the interference coefficient of the bridge measurement radar. Based on historical bridge construction data, the probability (Pconflict) of multiple radars simultaneously requesting resources (such as communication bandwidth) is statistically analyzed. Combined with the resource buffer (B) and the total capacity of shared resources (C), the delay risk caused by the conflict is assessed (ΔTconflict = k*Pconflict*B / C, where k is the delay coefficient). The radar mission resource conflict probability coefficient is represented by EEM = ΔTconflict * (Q1 * FT + Q2 * GRD + Q3 * ΔSNR), where EEM represents the mutual composite coefficient of the bridge measurement radar, Q1 represents the weighting factor of the bridge measurement radar frequency band overlap with the bridge measurement radar mutual composite coefficient, Q2 represents the weighting factor of the interference power intensity coefficient with the bridge measurement radar mutual composite coefficient, and Q3 represents the weighting factor of the bridge measurement radar interference coefficient with the bridge measurement radar mutual composite coefficient. The bridge measurement radar mutual composite coefficient is obtained by coupling the bridge measurement radar frequency band overlap, the bridge measurement radar interference power intensity coefficient, the bridge measurement radar interference coefficient, and the corresponding weighting factors, and then coupling them with the radar mission resource conflict probability coefficient.

[0116] In this embodiment, in single-bridge or multi-bridge construction scenarios, deployed tower verticality monitoring radar, millimeter-wave radar, and ground-penetrating radar may interact due to electromagnetic interference, resource competition, and the combined effects of environmental factors. The operating frequency bands of different radars may overlap. For example, although the 77GHz millimeter-wave radar and the 100MHz ground-penetrating radar operate in different frequency bands, energy leakage or harmonics in their transmitted signals may cover adjacent frequency bands, leading to signal crosstalk. The high-frequency signal (77GHz) of the millimeter-wave radar may be received as noise by the low-frequency receiver (100MHz) of the ground-penetrating radar, causing a decrease in the signal-to-noise ratio of the ground-penetrating radar. When the antenna patterns of the tower verticality monitoring radar (e.g., 24GHz) and the millimeter-wave radar (77GHz) overlap, sidelobe interference may cause misjudgment of the target echo.

[0117] Sharing construction resources, such as concrete pump trucks, tower cranes, and communication bandwidth, across multiple bridges or radar systems can lead to resource allocation conflicts. For example, when multiple radars are deployed on a single bridge—with millimeter-wave radar on the main bridge and ground-penetrating radar on the ramps—the shared communication link may experience transmission delays due to excessive data volume, affecting the real-time performance of the radar data. Similarly, during the construction of multiple bridges, the scheduling of concrete pump trucks may also lead to resource allocation imbalances due to deviations in radar measurement progress.

[0118] In bridge construction, the combined influence of bridge measurement radar frequency band overlap, bridge measurement radar interference power intensity coefficient, and bridge measurement radar interference coefficient stems from the superposition effect of the complex electromagnetic environment.

[0119] Frequency band overlap reflects the probability of two radar signals overlapping, essentially representing competition for frequency resources in the electromagnetic environment. Interference power intensity is calculated from factors such as transmit power and distance, reflecting the actual intensity of interference energy (affected by electromagnetic phenomena such as reflection and scattering in the environment). The interference coefficient measures the degree to which interference degrades signal quality (directly related to the superposition effect of noise and interference in the environment). All three are driven by the complexity of the electromagnetic environment—in complex environments, multipath effects and reflections increase FT (increased frequency overlap probability), enhance GRD (increased interference energy due to reflection superposition), and worsen ΔSNR (intensified noise and interference mixing). Therefore, the complexity of the electromagnetic environment is a "common regulating factor" for all three: the more complex the environment, the more significant the combined effect of FT, GRD, and ΔSNR, and the greater their contribution to the error of bridge measurement radar; conversely, in simple electromagnetic environments, the weight of their influence decreases, and the sum of their weighting factors is not a fixed value.

[0120] The specific fitting process is as follows: By collecting historical data on environmental electromagnetic intensity (input parameter) and FT, GRD, and ΔSNR (output parameters) from multiple bridge construction scenarios, a linear regression model is established. During model training, Q1, Q2, and Q3 are automatically adjusted to minimize the error between the predicted value and FT, GRD, and ΔSNR (output parameters). Ultimately, the values ​​of Q1, Q2, and Q3 directly reflect the relative contribution of the three to the error and are controlled by the complexity of the electromagnetic environment—the more complex the environment, the larger the overall weight factor, and vice versa. This yields a specific linear regression model; the environmental electromagnetic intensity is input to obtain the corresponding weights of the three parameters.

[0121] Furthermore, based on the composite impact assessment of bridge measurement radar, a second adjustment to the bridge measurement radar is performed. Specifically, if the mutual composite coefficient of the bridge measurement radar is less than or equal to the threshold of the mutual composite coefficient of the bridge measurement radar, the intelligent progress management model for bridge engineering is not adjusted; if the mutual composite coefficient of the bridge measurement radar is greater than the threshold of the mutual composite coefficient of the bridge measurement radar, the intelligent progress management model for bridge engineering is adjusted.

[0122] Furthermore, the intelligent progress management model for bridge engineering is adjusted, specifically including: comparing and analyzing the cross-composite coefficient of bridge measurement radar with the threshold value of the cross-composite coefficient of bridge measurement radar to obtain the cross-composite difference coefficient of bridge measurement radar; ETX = (EEM - Eth) / Eth, where ETX represents the cross-composite difference coefficient of bridge measurement radar and Eth represents the threshold value of the cross-composite coefficient of bridge measurement radar; after coupling analysis of the cross-composite difference coefficient of bridge measurement radar with the priority indicator value of bridge construction tasks and rounding down, the corrected task priority indicator value is obtained. Ptask represents the priority indicator value for bridge construction tasks, and PtaskP is the modified task priority indicator value (the larger the value, the higher the priority of resource allocation). This indicates rounding down; SCT represents the first predicted delay time from the output of the intelligent schedule management model for bridge engineering. SCTP is the second predicted delay time.

[0123] In this embodiment, the mutual interference of multiple radars (quantified by EEM) is essentially the superposition of environmental uncertainties. Its core impact is reflected in two aspects: error propagation: the larger the EEM, the more likely the final error of the bridge measurement radar is to exceed the threshold, resulting in unreliable measurement data, which in turn prolongs the error correction time (such as re-measurement and adjustment of construction parameters), indirectly increasing the risk of schedule delays; in order to deal with the errors caused by the mutual interference of multiple radars, additional resources need to be invested (such as increasing the number of surveyors and extending the equipment calibration time), which leads to a decrease in resource redundancy, intensifies resource competition, and further increases the probability of schedule delays.

[0124] When the relationship between the cross-composite coefficient of bridge measurement radar and the threshold of the cross-composite coefficient of bridge measurement radar changes, it is necessary to adjust the error prediction logic or resource allocation strategy of the intelligent progress management model to reflect the actual impact of multi-radar interference on the construction progress.

[0125] Furthermore, the second adjustment of the bridge measurement radar based on the composite impact assessment of the bridge measurement radar also includes: comparing and analyzing the number of overlapping bridge construction tasks with the total number of bridge construction tasks, then performing coupling analysis and summing the results with the corresponding composite difference coefficients of the bridge measurement radar to obtain the bridge task interference coupling degree, where COH represents the bridge task interference coupling degree. Among them, EEM i RW represents the composite difference coefficient between the bridge measurement radars for the i-th bridge construction task. i This represents the number of overlapping tasks for the i-th bridge construction task, specifically the number of overlapping tasks between the i-th bridge construction task and any other task in terms of resource requirements (e.g., concrete usage, equipment shifts) and shared resource buffers (e.g., communication bandwidth, number of pump trucks). n represents the total number of bridge construction tasks, which is directly extracted from the production logs. If the bridge task interference coupling degree is less than the first threshold, it is considered low coupling, and no multi-task allocation is required. If the bridge task interference coupling degree is equal to or greater than the first threshold and less than the second threshold, it is considered medium coupling, and the resource redundancy corresponding to the shared resource tasks is increased; if shared resources are needed, redundancy is added. If the bridge task interference coupling degree is equal to or greater than the second threshold, it is considered high coupling, and relevant personnel are notified to increase the resource redundancy separately according to the corrected task priority indicator value; dedicated resources need to be allocated separately for high-priority tasks.

[0126] Furthermore, based on the second adjustment results of the bridge measurement radar, a second intelligent progress management system for the bridge project is implemented. Specifically, this includes: if the second predicted delay time is less than the predicted delay time threshold, then the second intelligent progress management system for the bridge project is not implemented; if the second predicted delay time is equal to or greater than the predicted delay time threshold, then the area corresponding to the second predicted delay time is located, and the bridge radar frequency band within the area is gradually reduced to the predefined minimum bridge radar operating frequency band; if the second predicted delay time is less than the predicted delay time threshold during the gradual reduction of the bridge radar frequency band, then the reduction of the bridge radar frequency band is stopped; if the bridge radar frequency band within the area is reduced to the predefined minimum bridge radar operating frequency band and the second predicted delay time is equal to or greater than the predicted delay time threshold, then the duration of the composite interference is recorded; if the duration of the composite interference exceeds the composite interference duration threshold, then an area warning is issued to relevant personnel.

[0127] In this embodiment, the impact of electromagnetic interference on the measurement is reduced by lowering the operating frequency band of the interference source, thereby reducing the progress delay caused by interference.

[0128] High-frequency radar is susceptible to environmental interference. Lowering the frequency can reduce interference power, improve measurement accuracy, and indirectly shorten the delay time caused by errors.

[0129] The frequency bands are gradually reduced to avoid signal instability caused by a large-scale adjustment at once. At the same time, the delay time is monitored in real time to dynamically judge the effect of the adjustment.

[0130] If the frequency band is reduced to the lowest level and still cannot meet the delay threshold, the duration of the composite interference is recorded and an early warning is issued, indicating that manual intervention is required to ensure that the construction progress is controllable.

[0131] Furthermore, the intelligent progress management model for bridge engineering is self-updated after a predefined model update time. This process includes: acquiring actual and target completion time data for bridge construction tasks; analyzing the error between the actual and target completion time data, recording this as actual progress delay time data; merging the actual progress delay time data and the second predicted delay time with historical progress delay times to form a new training set covering all scenarios of "accurate prediction in low-risk situations" and "biased prediction in high-risk situations"; using a predefined XGBoost regressor structure, loading historical training parameters such as feature weights and tree structures as initial values; and minimizing the predicted delay time as a label. The mean squared error between measured and actual values ​​is used to optimize the mapping relationship between input features and delay time. A new training set is input into the model, and the node weights of the tree are iteratively updated using a gradient boosting algorithm. After training, new model parameters (such as feature importance ranking and hyperparameters, e.g., max_depth = 6, learning_rate = 0.08) are generated to replace the old parameters of the original model. The optimal parameters of the new model (such as feature weights and hyperparameters) are stored in a configuration library (such as the database table xgb_params), recording the parameter version number and timestamp. In subsequent prediction tasks, the model defaults to reading the latest parameter version from the configuration library, replacing the initial default parameters, thus reusing historical experience.

[0132] In this embodiment, the actual progress delay time data, progress delay time, first predicted delay time, and second predicted delay time all represent the corresponding parameter meanings of the bridge construction delay time. However, the actual progress delay time data is actually obtained, the first predicted delay time and the second predicted delay time are predicted values ​​output by the bridge engineering intelligent progress management model, and the progress delay time is the label value of the bridge engineering intelligent progress management model, which includes values ​​from the dataset of historical actual progress delay time data.

[0133] Through a closed-loop process of "collecting actual-target time data, calculating prediction error, merging multi-source data, retraining the XGBoost model, and reusing and storing parameters", the prediction error data is used to correct model bias, enabling the model to continuously adapt to complex interference in multi-task collaborative scenarios, ultimately improving the accuracy of progress prediction and the effectiveness of adjustment.

[0134] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0135] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0136] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0137] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0138] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0139] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A multi-task collaborative intelligent progress management method based on artificial intelligence, characterized in that, The following steps are involved: The impact parameters of the bridge measurement radar output are evaluated, and the first adjustment of the bridge measurement radar is carried out based on the impact evaluation of the bridge measurement radar output. Based on the first adjustment results of the bridge measurement radar, the first intelligent progress management of the bridge project is carried out to obtain the intelligent progress management model of the bridge project. The combined influence of bridge measurement radar is evaluated, and the bridge measurement radar is adjusted in the second adjustment based on the combined influence evaluation, thereby adjusting the intelligent progress management model for bridge engineering. Intelligent progress management of bridge engineering is carried out based on the second adjustment results of bridge measurement radar. The bridge engineering intelligent progress management model is automatically updated after a predefined model update time.

2. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 1, characterized in that, The evaluation using the influence parameters output by the bridge measurement radar specifically includes: The signal-to-noise ratio of the bridge radar is obtained by acquiring data through the radar's built-in digital signal processing module; The radar's transmit and receive power are measured by a power meter, and the rain attenuation loss of the bridge radar is calculated by combining the free space path loss. By analyzing the peak power of clutter in the echo signal spectrum using the radar's built-in Fast Fourier Transform module, the clutter power of the bridge radar echo is obtained. By fitting the linear relationship between QLC and bridge radar signal-to-noise ratio, bridge radar rain attenuation loss and bridge radar echo clutter power using the least squares method in multiple regression analysis under different rainfall intensities; By fitting the linear relationship between the final measurement error coefficient of the bridge radar and the reciprocal of the bridge radar signal-to-noise ratio, the bridge radar rain attenuation loss, and the bridge radar echo clutter power under different rainfall intensities using the least squares method, the corresponding weighting factors are obtained. By comprehensively coupling the bridge radar signal-to-noise ratio, bridge radar rain attenuation loss, and bridge radar echo clutter power with their respective weighting factors, the final measurement error coefficient of the bridge radar is obtained.

3. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 1, characterized in that, The first adjustment of the bridge measurement radar based on the impact assessment of the bridge measurement radar output specifically includes: If the final measurement error coefficient of the bridge radar is less than the final measurement error threshold of the bridge radar, no adjustment will be made; If the final measurement error coefficient of the bridge radar is equal to or greater than the final measurement error threshold of the bridge radar, the bridge section is switched to the low frequency band, the equalizer weight is adjusted by the least mean square algorithm, and the over-threshold duration when the output frequency error coefficient of the bridge radar is equal to or greater than the final measurement error threshold of the bridge radar is recorded as the over-threshold duration of the final measurement of the bridge radar. If the final measured over-threshold duration of the bridge radar is less than the maximum value of the final measured over-threshold duration of the bridge radar, then no adjustment is made; If the final measurement over-threshold duration of the bridge radar is equal to or greater than the first value of the final measurement over-threshold duration of the bridge radar, then the receiving power of the corresponding bridge radar will be increased to carry out intelligent progress management of the bridge project. If the final measurement of the bridge radar exceeds the threshold duration is equal to or greater than the second value of the final measurement of the bridge radar exceeding the threshold duration, the receiving power of the corresponding bridge radar will be increased to carry out intelligent progress management of the bridge project and issue regional early warnings to relevant personnel.

4. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 3, characterized in that, The process of implementing intelligent progress management for bridge engineering based on the first adjustment results of bridge measurement radar is as follows: The input parameters of the intelligent progress management model for bridge engineering are extracted from historical bridge construction data, and the progress delay time corresponding to the input parameters of the intelligent progress management model for bridge engineering is marked. The input parameters of the intelligent progress management model for bridge engineering include the final measurement error coefficient of bridge radar, the magnitude of the error deviation threshold, the duration of the final measurement exceeding the threshold of bridge radar, the degree of increase in the receiving power of the corresponding bridge radar, the radar frequency band adjustment indicator, the priority indicator of bridge construction tasks, and the redundancy of bridge construction scheduling resources. The input parameters of the intelligent progress management model for bridge engineering are standardized and normalized. The relationship between the input parameters and the delay time is trained by using the predefined intelligent progress management model for bridge engineering with the delay time as the label, so as to obtain the intelligent progress management model for bridge engineering. The bridge engineering intelligent progress management model outputs the predicted delay time and notifies relevant personnel of the predicted delay time.

5. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 1, characterized in that, The assessment based on the combined effects of bridge measurement radar specifically includes: The proportion of the frequency range overlap between bridge measurement radars to the total frequency band is denoted as the bridge measurement radar frequency band overlap. The interference power at the target radar receiver is measured using a spectrum analyzer. Combined with the transmit power and distance, the interference power intensity coefficient of the bridge measurement radar is calculated. The interference power of the target bridge measurement radar receiver is measured using a spectrum analyzer. The interference coefficient of the bridge measurement radar is obtained by comparing the bridge measurement radar with the original noise power and then performing logarithmic analysis. Based on historical bridge construction data, the average probability of bridge measurement radars simultaneously requesting resources, combined with the comparative analysis of resource buffer capacity and total capacity of shared resources, the risk of bridge measurement radar delays caused by conflicts is assessed, and the probability coefficient of radar mission resource conflicts is obtained. By coupling analysis of the frequency band overlap of bridge measurement radar, the interference power intensity coefficient of bridge measurement radar, the interference coefficient of bridge measurement radar and the corresponding weighting factors, and then coupling with the radar mission resource conflict probability coefficient, the mutual composite coefficient of bridge measurement radar is obtained.

6. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 1, characterized in that, The second adjustment of the bridge measurement radar based on the composite impact assessment of the bridge measurement radar specifically includes: If the cross-composite coefficient of bridge measurement radar is less than or equal to the threshold of the cross-composite coefficient of bridge measurement radar, then the intelligent progress management model for bridge engineering will not be adjusted. If the composite coefficient of bridge measurement radar exceeds the threshold of the composite coefficient of bridge measurement radar, then the intelligent progress management model for bridge engineering will be adjusted.

7. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 6, characterized in that, The intelligent progress management model for bridge engineering specifically includes: By comparing and analyzing the mutual composite coefficient of bridge measurement radar with the threshold of the mutual composite coefficient of bridge measurement radar, the mutual composite difference coefficient of bridge measurement radar is obtained. After coupling the composite difference coefficient of the bridge measurement radar with the bridge construction task priority indicator value and then rounding down, the corrected task priority indicator value is obtained. The corrected task priority flag value is updated to the bridge construction task priority flag value in the bridge engineering intelligent progress management model, and the new predicted delay time is output through the bridge engineering intelligent progress management model, which is recorded as the first predicted delay time. The first predicted delay time is obtained by coupling the composite difference coefficient of the bridge measurement radar with the first predicted delay time and then rounding it down.

8. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 7, characterized in that, The second adjustment of the bridge measurement radar based on the composite impact assessment of the bridge measurement radar also includes: By comparing and analyzing the number of overlapping bridge construction tasks with the total number of bridge construction tasks, and then performing coupling analysis and summation with the corresponding bridge measurement radar composite difference coefficient, the bridge task interference coupling degree is obtained. If the interference coupling degree of the bridge task is less than the first threshold of the interference coupling degree of the bridge task, then there is no need to perform multi-task allocation. If the interference coupling degree of the bridge task is equal to or greater than the first threshold of the interference coupling degree of the bridge task and less than the second threshold of the interference coupling degree of the bridge task, then the resource quantity redundancy corresponding to the shared resource task will be reduced. If the interference coupling degree of the bridge task is equal to or greater than the second threshold of the interference coupling degree of the bridge task, then relevant personnel shall be notified to increase the resource redundancy separately according to the corrected task priority indicator value.

9. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 7, characterized in that, The second intelligent progress management of bridge engineering based on the second adjustment results of bridge measurement radar specifically includes: If the second predicted delay time is less than the predicted delay time threshold, then the second intelligent progress management of the bridge project will not be implemented. If the second predicted delay time is equal to or greater than the predicted delay time threshold, then locate the area corresponding to the second predicted delay time, and gradually reduce the bridge radar frequency band within the area to the predefined minimum bridge radar operating frequency band. If the second predicted delay time is less than the predicted delay time threshold during the gradual reduction of the bridge radar frequency band, then stop reducing the bridge radar frequency band. If the frequency band of the bridge radar in the area is reduced to the predefined minimum operating frequency band of the bridge radar and the second predicted delay time is equal to or greater than the predicted delay time threshold, then the duration of the composite interference is recorded. If the duration of the combined interference exceeds the threshold, an area warning will be issued to relevant personnel.

10. The multi-task collaborative intelligent progress management method based on artificial intelligence as described in claim 7, characterized in that, The self-updating of the bridge engineering intelligent progress management model triggered by the predefined model update time specifically includes: Obtain the actual completion time data and target completion time data of the bridge construction task, analyze the error data between the actual completion time data and the target completion time data, and record it as the actual progress delay time data. Combine the actual progress delay time data and the second predicted delay time with the historical progress delay time to form a new training set. Retrain and update the intelligent progress management model of bridge engineering with the new training set.

Images