Bridge hanging basket installation precise positioning automation operation method

Abstract

The application relates to the technical field of building engineering, and discloses a bridge hanging basket installation accurate positioning automation operation method, which comprises the following steps: through installation of a high-precision sensor and a visual system, real-time collection of position, speed, posture and image data of a construction environment of a bridge hanging basket; analysis of the collected image data by using a deep learning model, identification and estimation of the spatial position and angle of the hanging basket; the system comprises: a high-precision sensor and an image collection system, which are used for real-time acquisition of the position, posture and environment data of the hanging basket; a deep learning module, which comprises a convolutional neural network and a recurrent neural network, and is used for real-time identification of the position of the hanging basket and prediction of future deviation; and a nonlinear optimization module, which is used for real-time calculation and correction of optimal control input according to an optimization objective function. Through the technical scheme of the high-precision sensor and the image collection system, real-time and accurate monitoring of the position, posture and construction environment of the hanging basket is achieved.

Description

Technical Field

[0001] This invention relates to the field of building engineering technology, specifically to an automated operation method for precise positioning of bridge hanging basket installation. Background Technology

[0002] In existing technologies, the accuracy of positioning and dynamic adjustment are crucial during the installation of bridge formwork. However, many traditional technologies still rely on single positioning sensors or low-precision equipment, which makes it impossible to guarantee accurate measurement of the formwork's position and angle in complex construction environments. For example, while traditional total stations and laser rangefinders can provide spatial position data for the formwork, these devices often exhibit instability in dynamic environments, especially under conditions of changing lighting, weather, or construction interference, resulting in significantly reduced accuracy and delayed correction processes, making it difficult to adjust the formwork's position in a timely manner.

[0003] Another significant problem arises during the correction of the hanging basket position. Existing correction schemes typically employ fixed control strategies, neglecting the variability and dynamism inherent in actual construction. This strategy often leads to over-correction or energy waste. In some cases, because the energy consumption of control inputs during correction is not considered, the system may inefficiently utilize excessive resources, resulting in low operational efficiency and increased unnecessary energy consumption. Moreover, when the hanging basket correction deviation is large, traditional control methods often require multiple adjustments, increasing construction time and costs.

[0004] Meanwhile, traditional feedback systems largely rely on static error calculations and linear correction methods, which can easily lead to significant error accumulation in dynamic and complex construction environments. For example, external conditions at the construction site (such as wind speed and weather) have a significant impact on the stability of the hanging basket, but existing technologies rarely respond quickly to these changes and make adjustments. This limitation results in a high error rate in the correction process and adversely affects the safety and stability of construction.

[0005] Furthermore, existing technologies remain insufficient in predicting future deviation trends and making dynamic adjustments. Many systems rely on manual intervention to adjust the hanging basket position, which not only reduces operational efficiency but also increases the risk of human error. Systems with low levels of automation cannot respond promptly to real-time changes at the construction site, making it difficult for them to maintain high-precision operation under variable construction conditions.

[0006] Therefore, this invention proposes an automated operation method for precise positioning of bridge hanging basket installation to address the shortcomings of existing technologies. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides an automated operation method for precise positioning during bridge hanging basket installation, which solves the problems of insufficient positional accuracy, delayed dynamic adjustment, and energy waste during bridge hanging basket installation in complex construction environments.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an automated operation method for precise positioning of bridge hanging basket installation, comprising the following steps:

[0009] By installing high-precision sensors and vision systems, image data of the bridge hanging basket's position, speed, attitude, and construction environment are collected in real time.

[0010] The collected image data is analyzed using a deep learning model to identify and estimate the spatial position and angle of the hanging basket;

[0011] Based on the output of the deep learning model, a nonlinear optimization objective function is constructed, and the optimal correction strategy is calculated by minimizing the position deviation and the energy consumption of the control input.

[0012] The control input is adjusted in real time using a nonlinear optimization algorithm to correct the deviation of the hanging basket;

[0013] The automated execution system executes correction commands in real time to adjust the position and angle of the hanging basket;

[0014] The corrected results are compared with the target location, and the control strategy is adjusted in real time based on the feedback.

[0015] Preferably, the high-precision sensor includes the following devices:

[0016] LiDAR is used to measure the position of the hanging basket in three-dimensional space in real time. By emitting lasers and receiving reflected signals, it can accurately obtain the relative position of the hanging basket with respect to the surrounding environment.

[0017] Total station is used to accurately measure the angle and attitude of hanging baskets, combining laser ranging and angle measurement to provide high-precision position and attitude data.

[0018] Accelerometer and gyroscope: The accelerometer monitors the acceleration of the basket, and the gyroscope is used to measure angular velocity.

[0019] The vision system collects environmental data around the hanging basket through cameras or other imaging devices, including at least lighting and surrounding obstacles, providing visual information related to the environment and supporting analysis by deep learning modules.

[0020] Preferably, the optimization objective function obtains the optimal control strategy by minimizing the position deviation and the power consumption of the correction control input;

[0021] The optimization objective function is:

[0022] ;

[0023] in, This represents the objective function to be optimized. This represents the total number of time steps in the optimization process; For time steps, from arrive ; It is hanging basket at all times The position vector; For the predetermined target location; For control input; This represents the squared error between the basket's position and the target position. It is a regularization parameter used to control the balance between the energy consumption of the control input and the position error; This represents the squared energy consumption of the control input.

[0024] Preferably, the deep learning model includes a convolutional neural network and a recurrent neural network, used for basket position recognition and deviation trend prediction, respectively.

[0025] Preferably, the steps of the nonlinear optimization algorithm include:

[0026] Based on the hanging basket position deviation data output by the deep learning model, a nonlinear optimization objective function is constructed.

[0027] Gradient descent and Newton's method were selected as optimization algorithms to calculate the optimal corrected control input.

[0028] By calculating the gradient of the objective function, the control input is adjusted in real time to minimize position deviation and control energy consumption;

[0029] During the iterative optimization process, the control input is continuously updated based on real-time feedback;

[0030] The optimized control inputs are used in the automated execution system to adjust the actual position and angle of the hanging basket.

[0031] Preferably, the steps of the automated execution system include:

[0032] Based on the optimized control input, generate corresponding correction instructions;

[0033] Use automated equipment to execute correction commands and adjust the position and angle of the hanging basket;

[0034] The corrective effect of the hanging basket was verified through a real-time monitoring system;

[0035] Based on real-time feedback data, further adjust the execution strategy;

[0036] After the basket correction is completed, the automated execution system enters standby mode, ready to perform the next round of correction operations.

[0037] Preferably, the step of providing real-time feedback and adjusting the control strategy includes:

[0038] The current position and attitude of the hanging basket are monitored in real time through sensors and vision systems;

[0039] The actual position is compared with the target position, the deviation is calculated, and a feedback signal is generated.

[0040] The feedback signal is input to the control system to adjust the control input and optimize the correction path of the hanging basket;

[0041] The control strategy is adjusted in real time based on feedback results, and the position and angle of the basket are continuously corrected.

[0042] This is used to stop adjusting and complete the correction process after the hanging basket reaches the predetermined target position.

[0043] Preferably, the deep learning model includes enhancing the system's adaptability to uncertain environments through generative adversarial networks, simulating data under different lighting and wind speed conditions.

[0044] Preferably, the correction strategy uses real-time feedback and adjustment for closed-loop control, ensuring that the hanging basket remains in the predetermined target position throughout the installation process. The real-time feedback and adjustment include:

[0045] The current position and attitude of the hanging basket are monitored in real time through sensors and vision systems;

[0046] The actual position is compared with the target position, the deviation is calculated, and a feedback signal is generated.

[0047] The feedback signal is input to the control system to adjust the control input and optimize the correction path of the hanging basket;

[0048] The control strategy is adjusted in real time based on feedback results, and the position and angle of the basket are continuously corrected.

[0049] This is used to stop adjusting and complete the correction process after the hanging basket reaches the predetermined target position.

[0050] This invention also provides an automated operating system for precise positioning of bridge hanging basket installation, including:

[0051] High-precision sensors and image acquisition systems are used to acquire real-time data on the position, attitude, and environment of the hanging basket;

[0052] The deep learning module, which includes convolutional neural networks and recurrent neural networks, is used to identify the hanging basket position in real time and predict future deviations.

[0053] The nonlinear optimization module is used to calculate and correct the optimal control input in real time based on the optimization objective function.

[0054] An automated execution system is used to automatically adjust the position and angle of the hanging basket according to optimization instructions;

[0055] A real-time feedback system is used to compare the target position with the correction results and adjust the control strategy in real time.

[0056] This invention provides an automated method for precise positioning during the installation of bridge hanging baskets. It offers the following advantages:

[0057] 1. This invention achieves real-time and accurate monitoring of the position, attitude, and construction environment of the hanging basket by adopting a high-precision sensor and image acquisition system. Compared with the single sensor or low-precision equipment commonly used in the prior art, this invention solves the problem of the hanging basket being difficult to accurately measure the position and angle in complex construction environments, and significantly improves the accuracy and real-time performance of correction operations.

[0058] 2. This invention introduces a deep learning module, combining convolutional neural networks and recurrent neural networks, to achieve real-time identification and prediction of the hanging basket position and future deviations. Compared with traditional rule-based control systems, this invention effectively solves the problem of correction lag caused by environmental changes and the dynamics of the hanging basket, and provides a more intelligent and accurate dynamic adjustment scheme.

[0059] 3. This invention employs a nonlinear optimization module to calculate the optimal correction control input in real time based on the output data of the deep learning module, achieving a balance between optimized position correction and energy consumption. Compared with the traditional fixed control strategy, this invention solves the problems of energy waste and over-adjustment in the basket correction process, and realizes a more efficient correction process.

[0060] 4. This invention achieves the ability to calibrate the position of the hanging basket in real time during construction through continuous monitoring and dynamic adjustment of the real-time feedback system. Compared with the feedback mechanism with low correction accuracy in the prior art, this invention effectively solves the problems of cumulative correction error and instability caused by external environmental interference, and significantly improves construction safety and stability. Attached Figure Description

[0061] Figure 1 For the present invention picture;

[0062] Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0063] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] Please see Figure 1 This invention provides an automated operation method for precise positioning during bridge hanging basket installation, comprising the following steps:

[0065] S1. By installing high-precision sensors and vision systems, real-time image data of the position, speed, attitude and construction environment of the bridge hanging basket are collected.

[0066] S2. Analyze the acquired image data using a deep learning model to identify and estimate the spatial position and angle of the hanging basket;

[0067] S3. Based on the output of the deep learning model, construct a nonlinear optimization objective function, and calculate the optimal correction strategy by minimizing the position deviation and the energy consumption of the control input.

[0068] S4. Use a nonlinear optimization algorithm to adjust the control input in real time and correct the deviation of the hanging basket;

[0069] S5. Through the automated execution system, correction commands are executed in real time to adjust the position and angle of the hanging basket;

[0070] S6. Compare the correction results with the target position, provide real-time feedback, and adjust the control strategy.

[0071] For step S1, real-time acquisition of the formwork's positioning data is crucial during the bridge formwork installation process, especially given the increasing complexity of hoisting and moving operations in bridge engineering. To ensure the accuracy of the formwork installation, this invention utilizes a series of high-precision sensors and a vision system to collect real-time data on the formwork's position, speed, attitude, and the construction environment. This data provides fundamental support for subsequent deep learning analysis and corrective control.

[0072] In this embodiment, multiple high-precision sensors and vision systems are first installed at the bridge construction site to collect comprehensive data on the hanging basket. These include, but are not limited to, lidar, total station, accelerometer, gyroscope, and image acquisition equipment. Through precise measurement and capture, these sensors and devices obtain real-time data on the hanging basket's position, velocity, attitude, and the environmental data of the construction site.

[0073] In one possible implementation, lidar is used to measure the position of the hanging basket in three-dimensional space in real time. By emitting a laser beam and receiving its reflected signal, lidar can accurately provide data on the relative position of the hanging basket to its surrounding environment, especially useful in bridge construction sites with complex structures. This position data is transmitted in real time to a data processing module for subsequent analysis by a deep learning module.

[0074] Alternatively, a total station is installed at a suitable angle and position to measure the angle and attitude of the hanging basket. The total station not only provides high-precision angle measurements but also monitors the attitude changes of the hanging basket in real time through an automatic tracking system. This is particularly important for accurately controlling the installation process of the hanging basket, especially when the hanging basket needs to be precisely adjusted in multi-dimensional space.

[0075] In some embodiments, accelerometers and gyroscopes are used to monitor the dynamic changes of the hanging basket. These sensors can acquire acceleration and angular velocity data of the hanging basket in real time. Accelerometers provide acceleration information of the hanging basket in various directions, while gyroscopes provide angular velocity data of the basket's rotation. This dynamic data is crucial for tracking the real-time motion changes of the hanging basket during installation, helping the system predict and adjust the future position of the basket during processing.

[0076] Specifically, the vision system is used to collect environmental data at the construction site, mainly including information on environmental changes such as light, temperature, and humidity around the hanging basket. The vision system not only provides direct visual data of the hanging basket but also assists in assessing changes in the construction environment through image recognition technology. This is extremely useful for dealing with unforeseen environmental changes during construction, such as sudden weather or lighting changes.

[0077] In this embodiment, real-time data from all sensors and vision systems is transmitted to the deep learning module for further processing. The deep learning model can extract valuable information from this multimodal data, providing data support for subsequent optimization algorithms.

[0078] In another implementation, to ensure data accuracy and real-time performance, the sensor system of this invention can be connected to the data processing system via a wireless communication network. This allows data collected by all sensors and vision devices to be transmitted to the central processing system in real time and processed through data fusion algorithms to improve data reliability and accuracy. Real-time data acquisition and processing provide a solid foundation for subsequent analysis, especially for the precise positioning and correction of the hanging basket in complex environments.

[0079] In summary, step S1, through multi-sensor data acquisition, ensured the accuracy of the hanging basket installation. The collaborative work of the sensors and the vision system not only provides precise location information but also monitors the dynamic behavior of the hanging basket and its surrounding environment in real time, thus providing rich data support for subsequent deep learning model analysis, optimization algorithm calculation, and execution system adjustments.

[0080] Step S2 follows the aforementioned data acquisition step and mainly uses a deep learning model to analyze the acquired image data, thereby recognizing the spatial position and angle of the hanging basket. The connection between step S2 and step S1 is crucial. Step S1 provides high-precision sensor and visual data support, while step S2 uses this data to accurately estimate the position of the hanging basket.

[0081] In this embodiment, in step S1, spatial position information, attitude data, and environmental data of the hanging basket are collected using devices such as lidar, total station, accelerometer, and gyroscope. Based on this, the deep learning module is responsible for processing this multimodal data. Specifically, convolutional neural networks (CNNs) and recurrent neural networks (RNNs) are used for position and dynamic state analysis.

[0082] Specifically, in the implementation of the deep learning model, the first step is to extract features from the image data using a convolutional neural network (CNN). CNNs have powerful capabilities in image recognition, able to extract the spatial location features of the hanging basket from complex image data, including key information such as the shape, position, and size of the basket. Specifically, the multi-layered structure of the convolutional neural network can progressively extract features from simple to complex, ultimately identifying the accurate spatial location and angle of the hanging basket.

[0083] In some embodiments, CNNs not only process static image data but also handle image changes in dynamic scenes. Therefore, even in complex and dynamic construction environments, the system can still accurately identify the position and orientation of the hanging basket using CNNs. This capability is particularly important when the construction environment changes frequently. For example, construction sites may experience rapidly changing factors such as lighting and weather; CNNs can adapt to these changes, ensuring the robustness and accuracy of data analysis.

[0084] As an alternative, recurrent neural networks (RNNs) were introduced to model the dynamic state of the hanging basket. Unlike traditional neural networks, RNNs have the advantage of processing sequential data, capturing temporal information within the time series. During the hanging basket position adjustment process, the RNN predicts the deviation trend in future moments based on the current and historical states of the hanging basket, making position corrections in advance. This approach significantly improves the response speed to position changes in dynamic environments. Through the predictions of the RNN model, the system can identify and respond to potential deviations of the hanging basket in advance, reducing the time required for real-time correction.

[0085] In one possible implementation, by combining CNNs and RNNs, the system can simultaneously perform spatial location identification and dynamic trend prediction. Specifically, the CNN handles static data processing, identifying the real-time position and orientation of the hanging basket; while the RNN handles time-series data, predicting possible deviations of the hanging basket over a future period. This combination not only improves the accuracy of location identification but also provides more comprehensive support for subsequent nonlinear optimization and control strategies.

[0086] To ensure accuracy, the training process of the deep learning module is crucial. Typically, deep learning models are trained using a large amount of labeled data. In the application of hanging basket positioning, the model needs to be trained using a large amount of construction site data (such as the hanging basket's position, orientation, and environmental conditions). Through deep learning algorithms, the model can learn positioning patterns under different environments and construction scenarios from this data, further improving the accuracy of recognition and prediction.

[0087] In some embodiments, generative adversarial networks (GANs) can be incorporated into the training process to optimize model performance. GANs allow the system to learn independently under simulated complex environmental conditions, enhancing its adaptability to changes in real-world construction environments. By simulating environmental changes (such as lighting and weather), GANs enable the model to be more robust to real-world construction scenarios.

[0088] Through the deep learning analysis in step S2, not only can the spatial position and angle of the hanging basket be accurately identified, but its future deviations can also be predicted. The core advantage of this process is that, through deep learning inference and prediction, the system can dynamically adjust the position of the hanging basket and anticipate potential deviations, thereby significantly improving the accuracy of hanging basket installation.

[0089] In step S2, the output of the deep learning model becomes a crucial input to the subsequent nonlinear optimization process. The objective function is adjusted based on these outputs to ensure the precise positioning of the hanging basket throughout the installation process. Therefore, step S2 provides essential data support for subsequent nonlinear optimization and corrective control.

[0090] In summary, step S2 uses deep learning technology to process multimodal sensor data, accurately identifying the hanging basket's position and predicting deviation trends. The combination of CNN and RNN not only improves the accuracy of hanging basket positioning but also enhances the system's adaptability to dynamic environments. This combination of technologies provides strong technical support for precise hanging basket positioning, effectively improving construction efficiency and accuracy while reducing errors and risks.

[0091] In step S3, following step S2, data from various sensors and the vision system were processed using a deep learning model to accurately identify the spatial position, attitude, and dynamic state of the hanging basket, and to predict future deviation trends. Based on this information, the task of step S3 is to construct a nonlinear optimization objective function. This objective function calculates the optimal correction strategy by minimizing the position deviation and the energy consumption of the control input, thereby achieving precise positioning and dynamic correction of the hanging basket.

[0092] In this embodiment, step S3 constructs an optimization objective function using the current position information and future deviation predictions of the hanging basket provided by the deep learning model. The core of this objective function lies in balancing the relationship between position deviation and control input energy consumption to ensure accurate positioning of the hanging basket. During this process, the design of the optimization objective function must consider the deviation between the hanging basket and the target position, as well as the energy required for correction operations, to avoid unnecessary energy consumption.

[0093] Specifically, the objective function can typically be expressed in the following form:

[0094] ;

[0095] in, This represents the objective function to be optimized. This represents the total number of time steps in the optimization process; For time steps, from arrive ; It is hanging basket at all times The position vector; For the predetermined target location; For control input; This represents the squared error between the basket's position and the target position. It is a regularization parameter used to control the balance between the energy consumption of the control input and the position error; This represents the squared energy consumption of the control input.

[0096] The core of this objective function is to simultaneously minimize the position error and the energy consumption of the control input. In this way, the optimization algorithm can not only precisely adjust the position of the basket but also effectively control energy usage during the correction process, thereby improving the efficiency of the correction process.

[0097] In one possible implementation, the objective function is constructed by considering not only the position error but also the energy consumption of the control input. This is achieved by introducing a regularization parameter. The system can balance positional accuracy and control input energy consumption during the correction process. This can be achieved by adjusting the system's settings in different construction environments. This allows for the selection of a suitable balance between accuracy and efficiency. For example, if more precise positioning is required, the system can add... The value of allows for a moderate increase in energy consumption for control inputs while maintaining accuracy; conversely, if energy consumption needs to be minimized, it can be reduced. This reduces the consumption of control inputs.

[0098] Alternatively, the objective function can be further extended to consider the dynamic response of the basket. For example, control over the basket's speed or acceleration can be added to avoid overcorrection or instability caused by vigorous movement. In this case, the objective function can be extended to:

[0099] ;

[0100] in, It is the objective function, representing the overall optimization objective throughout the entire correction process; This represents the total number of time steps in the optimization process; Indicates a time step, from arrive , representing each time point in the optimization process; Indicates the time when the basket is hanging. The position vector, i.e., the three-dimensional coordinate information of the hanging basket; The predetermined target position is the final position that the hanging basket should reach; This indicates a correction control input, which is a command to adjust the position and attitude of the basket, and can be used to control the actions of equipment such as robotic arms and automated guided vehicles (AGHs); The squared error of the position deviation represents the magnitude of the deviation between the current hanging basket position and the target position, and the goal is to minimize this deviation; It is a regularization parameter used to balance the weight between position deviation and control input energy consumption; It controls the energy consumption of the input and measures the energy consumed during the correction process; It is a regularization parameter used to balance the influence of control speed among other terms; It is hanging basket at all times The speed indicates the speed during the basket correction process; The square of the hanging basket's velocity represents the speed of the hanging basket during the correction process.

[0101] This extended optimization objective function can more comprehensively control the dynamic response during the correction process, ensuring a smoother correction process and reducing unnecessary oscillations or vibrations.

[0102] In some embodiments, the design of the objective function may be further adjusted based on real-time feedback information. For example, in a construction environment, the position and attitude of the hanging basket are affected not only by the control input but also by external factors (such as wind speed, temperature, etc.). In this case, the system can dynamically update the parameters in the objective function based on real-time feedback signals to further optimize the correction process.

[0103] Specifically, once the objective function is constructed, the next step is to calculate the optimal correction strategy using a nonlinear optimization algorithm. The optimization algorithm utilizes the current state information of the basket, the target position, and the predicted deviation trend to calculate the best corrective control input. Typically, nonlinear optimization algorithms employ common methods such as gradient descent, Newton's method, or quasi-Newton methods for iterative optimization.

[0104] In one possible implementation, the system uses gradient descent to progressively calculate the gradient of the objective function relative to the corrected control input and adjusts the control input based on this gradient information. After each iteration, the system calculates the error between the current position and the target position and updates the control input until the predetermined accuracy is achieved. For complex construction environments, heuristic search or genetic algorithms may also be employed to further accelerate the optimization process and ensure the real-time performance and accuracy of the correction operation.

[0105] In summary, step S3 ensures that the formwork can be precisely positioned to the predetermined location during installation by constructing a nonlinear optimization objective function and adjusting the control input in real time using an optimization algorithm. The optimization objective function balances positional deviation and energy consumption, and is flexibly adjusted by introducing regularization parameters to adapt to different construction needs. Through this optimization control method, the present invention can maximize the efficiency of the correction process while ensuring high-precision positioning, thus ensuring the safety and accuracy of bridge formwork installation.

[0106] In step S4, based on the predicted position, attitude, and dynamic deviation of the hanging basket output by the deep learning model, a nonlinear optimization objective function is constructed. This objective function comprehensively considers position deviation, energy consumption for correction, and possible dynamic response, and calculates the optimal correction strategy through an optimization algorithm. Step S4 then applies this nonlinear optimization algorithm to adjust the control input in real time, thereby correcting the deviation of the hanging basket and ensuring that it remains in the target position. The connection between step S4 and the preceding steps is crucial; the optimization calculation results provide decision support for subsequent correction operations, ensuring the accuracy and efficiency of the entire correction process.

[0107] In this embodiment, step S4 applies a nonlinear optimization algorithm (such as gradient descent, Newton's method, or quasi-Newton method) to adjust the control input and correct the deviation of the hanging basket. The role of the nonlinear optimization algorithm here is to calculate the optimal corrected control input based on the objective function, enabling the hanging basket to gradually adjust and ultimately achieve precise positioning. The objective function, as defined in step S3, includes the error term for the hanging basket position and the control input energy consumption term. Based on this, the optimization algorithm calculates the correction command in real time.

[0108] Specifically, the optimization algorithm first needs to calculate the gradient of the objective function with respect to the control input and then adjust it based on this gradient information. Gradient descent is a commonly used optimization algorithm that updates the control input step by step by calculating the gradient of the objective function relative to the control input. After each optimization iteration, the system adjusts the control strategy based on the deviation between the current position and the target position to minimize the error, while also paying attention to the energy consumption of the control input to avoid unnecessary corrections.

[0109] Alternatively, if higher computational efficiency or faster convergence is required, the system can use second-order optimization methods such as Newton's method or quasi-Newton methods. Compared to gradient descent, these methods, by utilizing the second derivative of the objective function (i.e., the Hessian matrix), can more accurately estimate the optimal update direction of the corrected control input in each iteration. This is particularly effective when the correction error is large, significantly improving the efficiency of the correction process.

[0110] In some embodiments, to improve optimization efficiency, the system can also adjust the learning rate or step size of the algorithm based on real-time feedback. By dynamically adjusting the learning rate, the system can quickly adapt to more complex or uncertain environments and improve convergence speed. For example, during the initial correction, the learning rate can be set larger to speed up the correction process; while as the target position approaches, the learning rate can be reduced to avoid oscillations caused by over-adjustment.

[0111] In one possible implementation, the optimization algorithm in step S4 is progressively corrected based on real-time data. For example, each time the system performs a correction and updates the basket position, the real-time feedback system provides new error values ​​and correction results. This data is input into the optimization algorithm to recalculate the optimal control input. Based on this feedback information, the optimization algorithm adjusts the control input in real time, causing the basket to gradually converge towards the target position. This process forms a dynamic closed loop, ensuring that the correction operation is precisely adjusted at each time step.

[0112] Specifically, in some complex environments, the system can also utilize adaptive control algorithms to optimize parameter settings during the optimization process. Adaptive control can automatically adjust control parameters in the optimization algorithm, such as regularization parameters and learning rates, based on the actual effect of the hanging basket position correction. This adaptive mechanism ensures that the system maintains high-precision positioning under different construction environments while improving the robustness of the correction process.

[0113] For example, in situations with significant external disturbances (such as strong winds or sudden environmental changes), the control system can automatically adjust parameters to reduce the impact of external factors on the correction process. This allows the system to maintain accuracy and efficiency even in complex construction sites.

[0114] In step S5, the system has calculated and adjusted the corrected control input using a nonlinear optimization algorithm to ensure that the optimal strategy for the hanging basket correction is established. Step S5 follows step S4 and aims to use an automated execution system to execute correction commands in real time, adjusting the position and angle of the hanging basket to ensure precise positioning during construction. Step S5 is crucial in the system flow; it translates the results of the optimization calculation into actual control actions and ensures that the correction operation can be executed accurately.

[0115] In this embodiment, the automated execution system receives the corrected control input calculated in step S4 and transmits these control commands to the execution devices, such as robotic arms, automated guided vehicles (AGVs), and hydraulic control systems. The execution devices adjust the position and angle of the hanging basket according to the optimized commands. This process involves converting the control input obtained from the optimization algorithm into specific operating signals, guiding the execution system to precisely control the hanging basket.

[0116] Specifically, in an automated execution system, the execution of corrective commands begins with the control system dynamically adjusting the position and orientation of the basket. For example, if the optimization result indicates that the basket needs to move along the X-axis, the Automated Guided Vehicle (AGV) will receive the corresponding command and adjust the basket's position through the onboard control system. Similarly, if the command requires adjusting the basket's angle, the robotic arm will execute precise angle adjustments based on the control input command to ensure the basket reaches the target orientation.

[0117] Alternatively, in some embodiments, the execution system may also include multiple feedback loops to ensure the accuracy of the correction operation. Specifically, the execution system may be equipped with position sensors, angle sensors, etc., to monitor the movement and adjustment effect of the basket in real time. These sensors feed data back to the control system in real time, allowing the system to determine whether the current correction effect meets the expected accuracy. If the correction accuracy does not meet expectations, the system will recalculate the correction strategy and adjust the control input to ensure the accuracy of the correction process.

[0118] In some embodiments, to further improve the accuracy and safety of correction operations, the automated execution system can integrate an environmental perception system, such as lidar or a vision system, for real-time monitoring of the construction environment. The environmental perception system can identify surrounding obstacles, changes in construction conditions, etc., and adjust the execution path and correction strategy in a timely manner to avoid correction errors caused by environmental changes.

[0119] Specifically, there is a close collaborative relationship between the execution system and the real-time monitoring system. The real-time monitoring system not only oversees the completion of corrective operations but also provides immediate feedback to the execution system. For example, after the execution system completes the correction of the hanging basket's position, the monitoring system verifies the difference between the final position and the target position using sensors or cameras. If the difference is significant, the real-time feedback system immediately sends a warning signal to the control system, indicating potential undercorrection or error accumulation. Based on the real-time feedback information, the control system further adjusts the correction strategy and readjusts the control input for the next step of refined correction.

[0120] Alternatively, in some embodiments, the execution system may also use preset path planning and correction trajectories, combined with the precise control of automated equipment, to make the correction process of the hanging basket smoother and more efficient. For example, an automated guided vehicle (AGV) not only moves according to the target position, but also presets multiple possible correction trajectories in the path planning to ensure the continuity and smoothness of the execution process and avoid sudden errors caused by environmental changes.

[0121] In one possible implementation, to ensure both high efficiency and high precision in the correction operation, the execution system can improve control accuracy through multi-sensor fusion technology. Multi-sensor fusion technology can integrate data from different sensors (such as lidar, accelerometers, gyroscopes, etc.) in real time, thereby more accurately estimating the position and motion state of the basket. After the sensor data is fused, the automated execution system can obtain more precise correction commands, thus finely adjusting each movement step of the basket.

[0122] Specifically, when sensor data and optimization commands are combined, the execution system can dynamically adjust the position of the hanging basket with higher precision. For example, visual sensors can be used to monitor the relative position of the hanging basket and the construction target, providing high-resolution position information; lidar provides three-dimensional spatial data of the hanging basket, helping the execution system to accurately adjust the angle and attitude of the hanging basket. All of this data is processed through multi-sensor fusion and ultimately provided to the execution system to ensure the precise alignment of the hanging basket with the target position.

[0123] In some embodiments, the system may also include a safety warning function. During the correction process, if the system detects that the correction speed is too fast or exceeds a preset safety range, the system will automatically stop the current operation to prevent over-correction or equipment damage. This function ensures that system security is maintained while correcting efficiently through real-time data analysis during the correction process.

[0124] Step S5, through an automated execution system, transforms the optimal control input calculated in step S4 into actual formwork correction operations. The close coordination between the automated equipment and the real-time monitoring system ensures the accuracy and efficiency of the correction operation. Through precise control system integration and multi-sensor fusion, the correction process maintains high efficiency and accuracy even under changing construction environments. Simultaneously, the safety mechanisms and real-time feedback loops within the execution system further enhance the system's stability and reliability. Thus, through precise automated execution and real-time monitoring, the system can effectively complete the accurate installation and dynamic adjustment of the bridge formwork, ensuring the efficient and safe progress of the project.

[0125] In step S6, following step S5, the automated execution system successfully corrected the hanging basket by adjusting its position and angle based on the optimal control input calculated by the optimization algorithm in step S4. Step S6 then monitors and evaluates the correction results through a real-time feedback system, adjusting the control strategy based on real-time data to ensure the hanging basket remains precisely aligned with the predetermined target position. The main objective of step S6 is to dynamically adjust the control input through real-time feedback, ensuring high precision and efficiency in the correction process.

[0126] In this embodiment, the real-time feedback system monitors the error between the current position and the target position of the hanging basket, calculates and generates a feedback signal in real time. This feedback signal is transmitted to the control system to help adjust the current control strategy. If the correction result does not meet the predetermined accuracy requirements, the system will immediately initiate a new correction calculation and adjust the control input. Through this real-time feedback mechanism, the system can react quickly to the position deviation of the hanging basket, ensuring the accurate positioning of the hanging basket.

[0127] Specifically, the real-time feedback system includes multiple high-precision sensors and monitoring devices, such as lidar, vision sensors, accelerometers, and gyroscopes. Lidar provides precise distance data between the basket and the target position, while vision sensors capture the angular differences between them. Accelerometers and gyroscopes monitor the dynamic motion of the basket, including velocity, acceleration, and rotation angle. All this data is processed in real-time using sensor fusion technology to generate accurate correction results and provide feedback suggestions for the dynamic adjustment of the basket.

[0128] As an alternative, the real-time feedback system not only monitors the position but also adjusts the speed and acceleration components of the control input through integrated dynamic response analysis. For example, during correction, if a rapidly changing dynamic response (such as oscillation) is detected in the basket, the system automatically reduces the correction speed or changes the correction strategy to avoid deviations caused by over-correction. By introducing speed and acceleration control terms, the system can avoid excessively rapid dynamic changes during basket correction, ensuring a smooth and oscillating correction process.

[0129] In some embodiments, to further improve the accuracy and intelligence of the feedback, the real-time feedback system can be combined with a deep learning model to predict potential future errors. The deep learning model can predict future deviation trends of the hanging basket position based on historical data and real-time input, and dynamically adjust the control input to further optimize the correction path. This intelligent feedback mechanism can significantly improve the system's adaptability in complex environments and reduce errors during the correction process.

[0130] In one possible implementation, the real-time feedback mechanism in step S6 is combined with the nonlinear optimization objective function in steps S3 and S4. By dynamically updating the parameters in the objective function, the accuracy and robustness of the control strategy are further improved. For example, when the real-time feedback system detects a deviation between the hanging basket and the target position, the feedback signal will affect the position error term in the objective function, thereby adjusting the control input during the optimization process. This dynamic adjustment mechanism enables the system to better cope with the dynamic changes at the construction site and optimize the correction path in real time at each correction step.

[0131] By integrating dynamic control functionality into the real-time feedback system, the system can adjust its control strategy based on real-time feedback after each correction. Feedback signals play a crucial role in this process; by providing data such as position, attitude, velocity, and acceleration, the system can adjust various parameters in the objective function in real time, ensuring a balance between accuracy and efficiency in the correction process.

[0132] Specifically, real-time feedback signals affect the correction strategy through the following steps:

[0133] Calculate the error between the current position and the target position, and generate a feedback signal.

[0134] The feedback signal is input to the control system, and the magnitude of the error determines whether the correction strategy needs to be adjusted.

[0135] Update and optimize the corresponding terms in the objective function (such as the position error term, control energy consumption term, etc.) and dynamically optimize and correct the path.

[0136] If the current correction does not achieve the target accuracy, recalculate the optimal correction input and execute a new correction operation.

[0137] This closed-loop control mechanism ensures that the system can adjust the control input based on real-time data after each correction, avoiding error accumulation and over-correction, and maintaining the efficiency and accuracy of the correction process.

[0138] In further implementation, step S6 may involve extending the objective function to incorporate dynamic control with real-time feedback. This extended objective function will not be limited to minimizing position error but may also include control over velocity, acceleration, and environmental factors.

[0139] Step S6 ensures the accuracy and efficiency of the formwork correction operation through dynamic adjustments to the real-time feedback system and control strategy. The real-time feedback system continuously collects dynamic data of the formwork and adjusts the control input in real time based on this data, optimizing the correction path. By introducing feedback signals into the dynamic update of the objective function, the system can precisely adjust the position, velocity, and acceleration of the formwork in each correction step, ensuring that the formwork remains at the predetermined target position. Furthermore, the system uses environmentally adaptive control to ensure that the correction operation can be carried out stably and efficiently in complex construction environments. This process guarantees the high precision, automation, and safety of the bridge formwork installation.

[0140] Please see Figure 2 The present invention also provides an automated operating system for precise positioning of bridge hanging basket installation, including:

[0141] High-precision sensors and image acquisition systems are used to acquire real-time data on the position, attitude, and environment of the hanging basket;

[0142] The task of the high-precision sensor and image acquisition system is to collect the position information, attitude information and construction environment data of the hanging basket in real time, so as to ensure the accurate monitoring and dynamic adjustment of the hanging basket position throughout the construction process.

[0143] This module plays a crucial role in the entire automation system, providing essential input data for subsequent deep learning and optimization modules. The high-precision sensor and image acquisition system, through the collaborative work of sensors and imaging equipment, can provide the system with multi-dimensional information, including position, attitude, velocity, and acceleration. Accurate acquisition of this information is a prerequisite for ensuring the system can determine the hanging basket's state and dynamic deviations in real time and accurately.

[0144] LiDAR: LiDAR accurately measures the distance to a target object by emitting a laser beam and receiving the reflected signal, effectively calculating the position and trajectory of the hanging basket in three-dimensional space. By monitoring changes in the distance between the hanging basket and its surrounding environment in real time, LiDAR provides the system with accurate real-time positioning data.

[0145] Total station: A total station, through a combination of laser ranging and angle measurement, can accurately measure the attitude and angle of the hanging basket, thereby helping the system to finely adjust the posture of the hanging basket. This equipment is crucial for measuring the high-precision position of the hanging basket.

[0146] Accelerometers and gyroscopes: These sensors can monitor the dynamic state of the hanging basket in real time, providing measurement data on acceleration and angular velocity. This information is crucial for real-time assessment of the hanging basket's stability, motion trends, and adjustment strategies, especially during construction when the basket's position and attitude are constantly changing.

[0147] Visual sensors (such as cameras): Visual sensors are used to acquire image information of the hanging basket and its surrounding environment. Through image processing technology, the spatial position and posture information of the hanging basket can be obtained, providing data support for deep learning models. Especially in complex construction environments, vision systems can assist the system in recognizing the surrounding environment and making decisions.

[0148] The deep learning module, which includes convolutional neural networks and recurrent neural networks, is used to identify the hanging basket position in real time and predict future deviations.

[0149] The deep learning module is responsible for analyzing data provided by sensors and image acquisition systems to estimate the position and angle of the hanging basket in real time, and predicting possible future deviation trends, providing data support for subsequent optimization modules.

[0150] The deep learning module is the core of the entire system. It can predict and estimate the position and dynamic deviation of the hanging basket in real time by processing complex image and time-series data. By employing convolutional neural networks (CNNs) and recurrent neural networks (RNNs), this module achieves accurate recognition of the hanging basket's position and dynamic behavior. Its advantages lie in its adaptability to different environments and scenarios, and its high fault tolerance.

[0151] Convolutional Neural Networks (CNNs): CNNs, through multiple convolutional layers, pooling layers, and fully connected layers, can effectively extract spatial features from input images. For basket recognition, CNNs can accurately capture the shape, position, and pose information of baskets under various conditions such as complex backgrounds, changing lighting, and occlusion. They are particularly suitable for processing image data, providing accurate visual feedback to the system.

[0152] Recurrent Neural Networks (RNNs): RNNs are crucial for processing time series data, enabling the prediction of deviation trends over a future period based on the past state of the hanging basket. In practice, RNNs use historical time series data such as acceleration, velocity, and angle changes to predict future states and provide accurate predictive information for optimizing the objective function, reducing latency corrections.

[0153] The nonlinear optimization module is used to calculate and correct the optimal control input in real time based on the optimization objective function.

[0154] Based on the data output provided by the deep learning module, the nonlinear optimization module constructs an optimization objective function, calculates and outputs the optimal correction control input in real time, and ensures that the basket is accurately aligned with the target position.

[0155] The optimization module serves as a crucial bridge between the deep learning module and the execution system. It translates the basket state identified and predicted by the deep learning module into actual control commands. This module ensures a balance between accuracy and energy consumption during basket adjustment, preventing over-correction.

[0156] The objective function is optimized based on the position error term, the energy consumption term for correcting the input, and possible dynamic changes (such as the basket's acceleration and velocity). The goal is to minimize the error between the basket and the target position while ensuring reasonable energy consumption. The system typically uses optimization algorithms such as gradient descent, Newton's method, and quasi-Newton methods to calculate the optimal corrected input.

[0157] By adjusting the regularization coefficient in the optimization function, the system can flexibly control the balance between accuracy and energy consumption during the correction process. The optimization module adjusts the control input based on real-time data output to correct the basket deviation as accurately as possible without wasting excessive energy.

[0158] The execution system is used to automatically adjust the position and angle of the hanging basket according to optimization instructions;

[0159] The execution system is responsible for receiving the control commands calculated by the optimization module and precisely adjusting the position and angle of the basket through automated equipment to ensure that the basket is accurately aligned with the target.

[0160] The execution system is the final execution unit for achieving precise positioning of the hanging basket. It translates optimization instructions into physical actions through automated equipment. A precise execution system can maximize the efficiency of the correction process and ensure that each correction meets the target requirements.

[0161] Automated Guided Vehicles (AGVs): When optimization commands require adjusting the spatial position of the hanging basket, the AGV will precisely move the basket to the target position according to the control input commands. The AGV system uses precise positioning and path planning technology to ensure the rapid and accurate movement of the hanging basket near the target position.

[0162] Robotic arm: If the correction command requires adjustment of the basket's posture (such as angle), the robotic arm will perform precise posture adjustments under the guidance of optimized input. The robotic arm's flexibility and high precision make it an indispensable component in the installation process.

[0163] Hydraulic control system: For certain correction operations (such as adjustments requiring large torque), the hydraulic system can provide sufficient power support. The hydraulic control system, in conjunction with sensor feedback, ensures precise operation of the basket during the correction process.

[0164] A real-time feedback system is used to compare the target position with the correction results and adjust the control strategy in real time.

[0165] The real-time feedback system is used to compare the position, attitude, and effects of correction operations of the hanging basket with the target position, and adjust the control strategy in real time based on the correction results.

[0166] The real-time feedback system ensures that every correction operation is verified by providing real-time data support. Through rapid feedback on the correction results, the system can immediately adjust the correction strategy upon detecting errors, thereby avoiding error accumulation and ensuring high-precision correction.

[0167] Real-time data acquisition and monitoring: The real-time feedback system uses multiple sensors, such as LiDAR, cameras, and accelerometers, to monitor every detail of the hanging basket's correction process. The system continuously calculates the error between the hanging basket's position and the target position, and generates a real-time feedback signal when a deviation is detected.

[0168] Environmental Adaptation and Dynamic Adjustment: The real-time feedback system not only monitors the position of the hanging basket but also adjusts the correction strategy based on environmental changes (such as wind speed and temperature). The system monitors changes in the construction environment in real time and optimizes the correction strategy promptly based on the impact of the external environment to cope with different construction challenges.

[0169] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automated operation method for precise positioning during bridge hanging basket installation, characterized in that: Includes the following steps: By installing high-precision sensors and vision systems, image data of the bridge hanging basket's position, speed, attitude, and construction environment are collected in real time. By using deep learning models to analyze the collected image data, the spatial position and angle of the hanging basket can be identified and estimated, the position of the hanging basket can be accurately identified, and the trend of the hanging basket deviation can be predicted. Based on the output of the deep learning model, a nonlinear optimization objective function is constructed, and the optimal correction strategy is calculated by minimizing the position deviation and the energy consumption of the control input. The control input is adjusted in real time using a nonlinear optimization algorithm to correct the deviation of the hanging basket; The automated execution system executes correction commands in real time to adjust the position and angle of the hanging basket; The correction results are compared with the target location, and the control strategy is adjusted in real time based on feedback. The optimization objective function is: ； in, This represents the objective function to be optimized. This represents the total number of time steps in the optimization process; For time steps, from arrive ; It is hanging basket at all times The position vector; For the predetermined target location; For control input; This represents the squared error between the basket's position and the target position. It is a regularization parameter used to control the balance between the energy consumption of the control input and the position error; This represents the squared energy consumption of the control input; The deep learning model includes a convolutional neural network (CNN) and a recurrent neural network (RNN). The CNN can extract the spatial location features of the hanging basket from image data, including the shape, position, and size of the basket. During the basket position adjustment process, the RNN processes time-series data based on the current and historical states of the basket, predicts the deviation trend in future moments, and performs position correction in advance. By combining the CNN and the RNN, spatial location recognition and dynamic trend prediction are performed simultaneously, providing support for subsequent nonlinear optimization and control strategies. The deep learning model enhances the system's adaptability to uncertain environments by using generative adversarial networks, simulating data under different lighting and wind speed conditions.

2. The automated operation method for precise positioning of bridge hanging basket installation according to claim 1, characterized in that, The high-precision sensor includes the following devices: LiDAR is used to measure the position of the hanging basket in three-dimensional space in real time. By emitting lasers and receiving reflected signals, it can accurately obtain the relative position of the hanging basket with respect to the surrounding environment. Total station is used to accurately measure the angle and attitude of hanging basket, and combines laser rangefinding and angle measurement to provide high-precision position and attitude data; Accelerometer and gyroscope: The accelerometer monitors the acceleration of the basket, and the gyroscope is used to measure angular velocity. The vision system collects environmental data around the hanging basket through a camera, including at least lighting and surrounding obstacles, providing visual information related to the environment and supporting analysis by a deep learning module.

3. The automated operation method for precise positioning of bridge hanging basket installation according to claim 1, characterized in that, The steps of the nonlinear optimization algorithm include: Based on the hanging basket position deviation data output by the deep learning model, a nonlinear optimization objective function is constructed. Gradient descent or Newton's method is selected as the optimization algorithm to calculate the optimal corrected control input; By calculating the gradient of the objective function, the control input is adjusted in real time to minimize position deviation and control energy consumption; During the iterative optimization process, the control input is continuously updated based on real-time feedback; The optimized control inputs are used in the automated execution system to adjust the actual position and angle of the hanging basket.

4. The automated operation method for precise positioning of bridge hanging basket installation according to claim 1, characterized in that, The steps of the automated execution system include: Based on the optimized control input, generate corresponding correction instructions; Use automated equipment to execute correction commands and adjust the position and angle of the hanging basket; The corrective effect of the hanging basket was verified through a real-time monitoring system; Based on real-time feedback data, further adjust the execution strategy; After the basket correction is completed, the automated execution system enters standby mode, ready to perform the next round of correction operations.

5. The automated operation method for precise positioning of bridge hanging basket installation according to claim 1, characterized in that, The steps of providing real-time feedback and adjusting the control strategy include: The current position and attitude of the hanging basket are monitored in real time through sensors and vision systems; The actual position is compared with the target position, the deviation is calculated, and a feedback signal is generated. The feedback signal is input to the control system to adjust the control input and optimize the correction path of the hanging basket; The control strategy is adjusted in real time based on feedback results, and the position and angle of the basket are continuously corrected. Once the hanging basket has reached the predetermined target position, stop adjusting and complete the correction process.

6. An automated operating system for precise positioning of bridge hanging basket installation, applied to the automated operation method for precise positioning of bridge hanging basket installation as described in any one of claims 1-5, characterized in that, include: High-precision sensors and image acquisition systems are used to acquire real-time data on the position, attitude, and environment of the hanging basket; The deep learning module, which includes convolutional neural networks and recurrent neural networks, is used to identify the hanging basket position in real time and predict future deviations. The nonlinear optimization module is used to calculate and correct the optimal control input in real time based on the optimization objective function. An automated execution system is used to automatically adjust the position and angle of the hanging basket according to optimization instructions; A real-time feedback system is used to compare the target position with the correction results and adjust the control strategy in real time.

Images