High-precision surveying and setting-out intelligent positioning system and method based on multi-source data fusion

Abstract

The present invention relates to the technical field of building surveying, in particular to a high-precision surveying and setting-out intelligent positioning system and method based on multi-source data fusion, including data acquisition, fusion processing, intelligent positioning calculation, setting-out path generation, wireless communication, and visualization terminal modules. The data acquisition module synchronously acquires multi-source data; the data fusion processing module performs spatio-temporal alignment, feature-level fusion, and detects the sensor status; the intelligent positioning calculation module outputs three-dimensional coordinates using various algorithms; the setting-out path generation module generates an optimized path based on the BIM model; the wireless communication module realizes high-speed data transmission and device communication; the visualization terminal module provides an intuitive interaction interface. The present invention utilizes multi-source data fusion to improve the positioning accuracy, ensures the reliability of the system through sensor failure detection, adapts to complex environments, optimizes the setting-out path, enhances the overall performance of the system, and provides an efficient and accurate solution for building surveying and setting-out.

Description

Technical Field

[0001] The present invention relates to the technical field of building surveying, and specifically to a high-precision surveying and setting-out intelligent positioning system and method based on multi-source data fusion. Background Technique

[0002] In the technical field of building surveying, with the continuous expansion of the scale and increasing complexity of modern building projects, higher requirements are put forward for the accuracy, efficiency, and reliability of surveying and setting-out positioning. However, the existing positioning technologies and systems have many limitations and are difficult to meet the actual engineering needs. Specifically:

[0003] Limited accuracy of traditional positioning technologies: Common traditional positioning technologies, such as single GNSS (Global Navigation Satellite System) positioning, in complex environments (such as areas with high-rise buildings in cities, mountainous areas, etc.), satellite signals are easily blocked, reflected, and interfered with, resulting in a significant drop in positioning accuracy and being unable to meet the requirements of centimeter-level or even higher accuracy for building surveying and setting-out; when conducting surveys in densely built-up urban areas, GNSS signals will be blocked by surrounding high-rise buildings, resulting in signal loss or multipath effects, and the positioning error can reach several meters or even larger, seriously affecting the accuracy of surveying and setting-out;

[0004] Insufficiencies in data collection and processing: Traditional data collection methods often rely on a single type of sensor, obtaining limited information and unable to comprehensively reflect the complex environment of the construction site; moreover, the data collected by different sensors lack effective synchronization and fusion in time and space, and the data processing efficiency is low, making it difficult to provide real-time and accurate positioning data support; for example, only using laser scanning to obtain terrain data cannot obtain the absolute position information of the equipment in real time; at the same time, due to the lack of a unified data processing mechanism, the data of different sensors cannot work together, resulting in inaccurate and untimely positioning results;

[0005] Lack of intelligence in setting-out path planning: Most of the existing setting-out path planning methods are based on simple rules or experience and do not fully consider the complex factors of the construction site, such as the distribution of obstacles, construction safety requirements, and the movement characteristics of equipment, etc.; this makes the planned path may not be the optimal path, not only increasing the construction time and cost, but also potentially having safety hazards; when planning the setting-out path, the obstacles formed by temporarily stacked materials and equipment at the construction site are not considered, resulting in the setting-out equipment needing to detour frequently, reducing the work efficiency;

[0006] Poor system stability and reliability: In the actual construction environment, sensors are vulnerable to various factors, such as bad weather (heavy rain, sandstorms, etc.), electromagnetic interference (interference from electrical equipment at the construction site), etc., which can lead to abnormal data or sensor failure. When facing these problems, traditional systems lack effective fault tolerance mechanisms and adaptive adjustment capabilities, resulting in poor system stability and reliability, and unable to ensure the continuity and accuracy of surveying and setting-out work. When abnormal data from a certain sensor is caused by bad weather, the traditional system may have positioning errors or be unable to locate, seriously affecting the construction progress.

[0007] Therefore, in view of the above problems, a high-precision surveying and setting-out intelligent positioning system and method based on multi-source data fusion are proposed. Summary of the Invention

[0008] The purpose of the present invention is to provide a high-precision surveying and setting-out intelligent positioning system and method based on multi-source data fusion to solve the problems raised in the above background technology.

[0009] To achieve the above purpose, the present invention provides the following technical solutions:

[0010] A high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion, comprising:

[0011] Data acquisition module: It includes a laser scanning unit, a GNSS positioning unit, an inertial measurement unit, and a vision sensor unit, which synchronously collect three-dimensional laser point cloud data, satellite positioning data, inertial navigation data, and visual image data of the construction site, and output an original data stream with timestamp alignment.

[0012] Data fusion and processing module: Connected to the data acquisition module, it sequentially executes:

[0013] Multi-source data spatio-temporal alignment unit: Adopts the double-reference alignment algorithm of Beidou timestamp and ORB-SLAM3 time axis to establish a unified spatio-temporal coordinate system.

[0014] Feature-level fusion unit: Executes an improved hybrid Kalman filter-graph attention network model to generate a fused point cloud map.

[0015] Sensor failure detection unit: Real-time monitors the health status of each sensor and dynamically adjusts the fusion strategy.

[0016] Intelligent positioning and solution module: Receives the fused point cloud map and is connected to the inertial measurement unit of the data acquisition module, configured as:

[0017] Solves the absolute coordinates through the RTK / PPK dual-mode differential positioning algorithm.

[0018] Calculates the relative positioning coordinates through the VIO-SLAM algorithm assisted by laser point cloud.

[0019] Apply a robust estimation model to eliminate the influence of abnormal observations and output three-dimensional space coordinates;

[0020] Laying-out path generation module: Communicate with the intelligent positioning and calculation module, including:

[0021] BIM model parsing unit: Convert the engineering BIM model into three-dimensional path constraint conditions in the construction coordinate system;

[0022] Path planning unit: Generate a laying-out path based on a multi-objective optimization algorithm;

[0023] Wireless communication module: Connect to each module and configure as:

[0024] 5G communication unit: Upload data to the cloud platform at a rate of ≥1 Gbps;

[0025] LoRa self-organizing network unit: Establish a Mesh network in the signal occlusion area to achieve direct communication between devices;

[0026] Visualization terminal module: Receive the laying-out path and real-time positioning data, including:

[0027] AR display unit: Overlay and display the theoretical / measured position deviation vector;

[0028] Dynamic deviation correction unit: Execute a three-level feedback control to achieve real-time path correction.

[0029] As an optimal solution, the feature-level fusion unit adopts an improved hybrid Kalman filter-graph attention network model, and its state update formula is:

[0030] , where, is The fused state estimate value at time; 、 The state transition matrix calibrated online; The Kalman gain matrix dynamically adjusted based on sensor confidence; is the graph network coupling coefficient; is the cross-modal attention weight; is the trainable parameter matrix; is The state value at time; is The control input at time; is The measurement value at time; is the measurement matrix; is the set of neighbor nodes of sensor node i; is the neighbor node 's state value.

[0031] As a preferred solution, the calculation of cross-modal attention weights adopts a multi-sensor feature cross-validation algorithm:

[0032] , where and are query / key vector projection matrices; is the Euclidean distance of the sensor node; is the environmental interference attenuation factor; and are sensor type coding vectors; is the sensor node 's feature vector; is the dimension of the feature vector, ; is the normalized exponential function.

[0033] As a preferred solution, the multi-objective optimization algorithm executed by the path planning unit includes:

[0034] ;

[0035] Constraint conditions: ;

[0036] where , is the total path length; , is the maximum path curvature; , is the safety scoring function; , is the dynamic weight coefficient; is the path planning scheme; is the path 's positioning accuracy; is the path 's maximum slope; is the minimum value of the safety score; is the curvature of the path at time; is the safety-related distance of the th point on the path; is the number of points on the path.

[0037] As a preferred solution, when the sensor failure detection unit executes the fault tolerance control strategy:

[0038] When the th sensor failure is detected, the fusion algorithm is automatically adjusted to:

[0039] , where is the failure indication factor, and , represents the One sensor fails completely; is the set of neighbor nodes that exclude the failed sensor.

[0040] As a preferred solution, the dynamic deviation correction unit performs three-level control, including:

[0041] Primary control , The interface displays a yellow warning box;

[0042] Secondary control , generates a local path correction instruction and vibrates for warning;

[0043] Tertiary control , starts the system self-check and freezes the positioning output until manual confirmation;

[0044] Among them, is the deviation value between the theoretical position and the measured position.

[0045] As a preferred solution, when multiple devices work together, a distributed consensus algorithm is used to synchronize the positioning data:

[0046] , where is the convergence rate parameter, is the positioning data of device at time; is the positioning data of device at time; is the positioning data of the neighbor device of device at time; is the Kalman gain; is the measurement value of device ; is the measurement matrix.

[0047] The high-precision surveying and setting-out intelligent positioning method based on multi-source data fusion is carried out by using the high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion.

[0048] It can be seen from the technical solutions provided by the present invention above that the beneficial effects of the high-precision surveying and setting-out intelligent positioning system and method based on multi-source data fusion provided by the present invention are:

[0049] Achieve high-precision positioning: Through multi-source data fusion processing, the system integrates multiple units such as laser scanning, GNSS positioning, inertial measurement, and vision sensors to collect rich data and fuse them; using the dual-reference alignment algorithm of Beidou timestamp and ORB-SLAM3 timeline to establish a unified spatio-temporal coordinate system, laying the foundation for precise positioning; improving the hybrid Kalman filter-graph attention network model to deeply fuse data, mining the correlation and complementarity of data, generating an accurate fused point cloud map, and providing reliable environmental information for positioning; adopting the RTK / PPK dual-mode differential positioning algorithm and the VIO-SLAM algorithm assisted by laser point cloud, combined with a robust estimation model, effectively eliminating the influence of abnormal observations, and accurately calculating three-dimensional space coordinates even in complex environments, meeting the strict requirements of high-precision positioning for building surveying and setting out;

[0050] Enhance system reliability: The sensor failure detection unit monitors the health status of each sensor in real time. Once a sensor failure is detected, the fusion strategy is automatically adjusted immediately; for example, when the th sensor fails, the fusion algorithm is adjusted to reduce the impact of the failed sensor, and continue to fuse and process data relying on the data of other normal sensors, ensuring the continuity and reliability of data fusion, maintaining the stable operation of the system, avoiding positioning errors or system failures caused by sensor failures, and ensuring the smooth progress of surveying and setting out work;

[0051] Improve environmental adaptability: In complex building construction environments, in the face of problems such as signal occlusion and electromagnetic interference, the system responds through multi-source data fusion; when satellite signals are occluded, the inertial measurement unit and vision sensor data assist in positioning; when electromagnetic interference affects GNSS positioning, laser scanning data is not affected and can still provide reliable environmental information; the wireless communication module integrates 5G and LoRa self-organizing network technologies. 5G realizes high-speed data transmission in areas with good signals, and LoRa self-organizing network establishes a Mesh network in signal-occluded areas to ensure communication between devices, enabling the system to adapt to different construction environments and improving adaptability and robustness;

[0052] Optimize setting-out path planning: The setting-out path generation module uses BIM model parsing and multi-objective optimization algorithms to generate an efficient and safe setting-out path; the BIM model parsing unit converts the engineering BIM model into three-dimensional path constraint conditions in the construction coordinate system, including rich design and construction information; the path planning unit, based on these conditions, comprehensively considers factors such as path length, curvature, and safety, and through multi-objective optimization algorithms and dynamic adjustment of weight coefficients, finds the optimal path, reducing the moving distance and time of equipment, improving setting-out efficiency, and ensuring construction safety;

[0053] Provide an intuitive interaction experience: The AR display and dynamic deviation correction functions of the visualization terminal module provide users with an intuitive and convenient interaction experience; The AR display unit superimposes and displays the theoretical / measured position deviation vector, allowing users to intuitively understand the position deviation without complex calculation and analysis; The dynamic deviation correction unit executes a three-level feedback control based on the deviation, timely reminding users and providing correction instructions to ensure the lofting accuracy; This intuitive interaction method reduces the operation threshold, reduces the dependence on professional technicians, and improves work efficiency and quality. Brief Description of the Drawings

[0054] Figure 1 It is a schematic diagram of the overall structure of the high-precision surveying and lofting intelligent positioning system and method based on multi-source data fusion of the present invention. Detailed Embodiment

[0055] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0056] In order to better understand the above technical solutions, the above technical solutions will be described in detail below in conjunction with the drawings of the specification and the specific embodiments.

[0057] As Figure 1 shown, the embodiment of the present invention provides a high-precision surveying and lofting intelligent positioning system based on multi-source data fusion, including a data acquisition module, a data fusion processing module, an intelligent positioning solution module, a lofting path generation module, a wireless communication module and a visualization terminal module (wherein, the wireless communication module is a communication tool between the data acquisition module, the data fusion processing module, the intelligent positioning solution module, the lofting path generation module and the visualization terminal module, so it is not reflected in Figure 1 ).

[0058] In this embodiment, the data acquisition module includes a laser scanning unit, a GNSS positioning unit, an inertial measurement unit and a vision sensor unit, which synchronously collect three-dimensional laser point cloud data, satellite positioning data, inertial navigation data and visual image data of the construction site, and output a raw data stream with time stamp alignment;

[0059] Furthermore, the data acquisition module plays a fundamental and crucial role in the high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion. It is like the "sensing antenna" of the system, responsible for obtaining various key data from the construction site, providing accurate and rich information support for subsequent data processing, positioning calculation, and setting-out operations. By integrating a variety of advanced sensor units, this module realizes the synchronous acquisition of multi-source data and outputs the original data stream with timestamp alignment, ensuring the timeliness and consistency of the data, laying a solid foundation for the high-precision operation of the entire positioning system. Specifically:

[0060] I. Composition of Sensor Units:

[0061] Laser Scanning Unit: Adopting advanced laser scanning technology, it can quickly and accurately obtain the three-dimensional laser point cloud data of the construction site. Its working principle is to emit laser beams and measure the time when the laser is reflected back, calculate the distance between the scanning point and the device, and thus construct a three-dimensional point cloud model of the construction site scene. For example, at a large-scale building construction site, the laser scanning unit can conduct a detailed scan of the terrain, building contours, and various obstacles at the construction site to obtain high-precision three-dimensional spatial information. These point cloud data not only contain the position information of objects but also can reflect the shape and surface characteristics of objects, providing rich environmental details for subsequent positioning and path planning. This unit features high resolution and a large scanning range, capable of meeting the survey requirements of construction sites of different scales. Whether it is an open site or a complex indoor environment, it can accurately obtain the required three-dimensional data.

[0062] GNSS Positioning Unit: Integrating Global Navigation Satellite System (GNSS) technologies such as GPS and Beidou, it is used to collect satellite positioning data. It calculates the precise position information of the device on the earth, including longitude, latitude, and altitude, by receiving satellite signals. During the building surveying and setting-out process, the GNSS positioning unit provides an absolute positioning reference for the system, ensuring the accuracy and reliability of positioning. For example, in the initial positioning stage of a field construction site or a large building project, the GNSS positioning unit can quickly determine the approximate position of the surveying device, providing the starting coordinates for subsequent precise measurement and setting-out operations. This unit has high-precision positioning capabilities. Combining RTK (Real-Time Kinematic) / PPK (Post-Processing Kinematic) technologies, it can achieve centimeter-level or even millimeter-level positioning accuracy, meeting the requirements of high-precision surveying and setting-out. At the same time, it also has good anti-interference capabilities and can stably receive satellite signals in a complex electromagnetic environment, ensuring the continuity of positioning.

[0063] Inertial Measurement Unit: Composed of sensors such as accelerometers and gyroscopes, it is used to measure the acceleration and angular velocity information of the device, and then obtain inertial navigation data. The inertial measurement unit plays an important auxiliary role in the positioning system. Especially when satellite signals are blocked or lost, it can calculate the position and attitude changes of the device through inertial measurement data. For example, in areas with weak satellite signals such as inside buildings or canyons, the inertial measurement unit can record the motion state of the device in real time, and combine with previous positioning information to achieve continuous positioning and tracking. It can quickly respond to the motion changes of the device, provide high-frequency measurement data, provide real-time motion information for positioning solution, and help improve the accuracy and real-time performance of positioning. In addition, the inertial measurement unit also has the characteristics of small size, light weight, and low power consumption, which is convenient for integration into various measurement devices.

[0064] Vision Sensor Unit: Equipped with a high-resolution camera, it is used to collect visual image data of the construction site. The vision sensor unit can obtain rich texture information and scene images, providing an intuitive visual basis for positioning and environmental perception. For example, at a construction site, the vision sensor can take photos or videos of the surrounding environment, and through image recognition technology, identify objects such as buildings, roads, and signs to assist in positioning solution and path planning. It can also be used to monitor potential safety hazards during the construction process, such as personnel's illegal operations and equipment abnormalities. The vision sensor unit has real-time imaging and image transmission functions, and can quickly transmit the collected image data to the data fusion processing module for analysis and processing. At the same time, combined with computer vision algorithms, it can realize the feature extraction and recognition of objects in the image, providing more valuable information for the system.

[0065] II. Data Acquisition and Synchronization Mechanism:

[0066] Synchronous Acquisition: The laser scanning unit, GNSS positioning unit, inertial measurement unit, and vision sensor unit in the data acquisition module work together to achieve synchronous acquisition of multi-source data. Through precise clock synchronization technology, each sensor unit starts data acquisition at the same moment, ensuring the temporal consistency of the collected data. For example, during a survey and setting-out operation, while the laser scanning unit scans the construction site scene, the GNSS positioning unit obtains the satellite positioning data of the current position, the inertial measurement unit records the motion state of the device, and the vision sensor unit takes pictures of the scene. These data are all collected at the same time point, providing an accurate time reference for subsequent data fusion and analysis.

[0067] Timestamp Alignment: To further ensure the temporal consistency of data, the data acquisition module adds precise timestamps to each set of collected data. The timestamps use a high-precision timing system, such as Beidou timestamps, to ensure the accuracy and uniqueness of time. At the same time, combined with the ORB-SLAM3 time-axis dual-reference alignment algorithm, the data collected by different sensors are precisely aligned on the time axis to establish a unified spatio-temporal coordinate system. In this way, during subsequent data processing, different sensor data can be accurately matched and fused according to the timestamps, avoiding data errors and inconsistencies caused by time differences, and improving the accuracy and reliability of data fusion.

[0068] III. Advantages of the Data Acquisition Module:

[0069] Multi-source Data Complementation: By integrating multiple types of sensor units, the data acquisition module can obtain rich and diverse information to achieve multi-source data complementation. The laser scanning unit provides high-precision three-dimensional spatial information, the GNSS positioning unit determines the absolute position, the inertial measurement unit tracks the motion state, and the vision sensor unit obtains textures and scene images. These different types of data complement each other, comprehensively describing the construction site environment and equipment status, providing a richer and more accurate information basis for high-precision surveying and setting out and intelligent positioning. For example, in a complex construction site environment, when the GNSS signal is blocked, the inertial measurement unit and the vision sensor unit can assist in positioning to ensure the continuity and accuracy of positioning. The combination of the three-dimensional point cloud data obtained by the laser scanning unit and the visual image data can more clearly identify and analyze the objects and obstacles in the construction site, providing a more accurate environmental model for path planning.

[0070] Real-time and Accuracy: The data acquisition module has high-speed data acquisition capabilities and precise time synchronization mechanisms, enabling it to obtain construction site data in real time and accurately. Each sensor unit collects data at a high frequency, and through timestamp alignment and synchronous transmission, the timeliness and consistency of the data are ensured. This enables the system to promptly reflect changes in the construction site, providing accurate data support for real-time positioning and decision-making. For example, during the construction process, when changes occur in the construction site environment (such as new obstacles, equipment movement, etc.), the data acquisition module can quickly collect this change information and transmit it to the subsequent modules for processing in a timely manner, ensuring that the positioning system can adjust the positioning results and setting out paths in real time, improving the safety and efficiency of construction.

[0071] Adapt to complex environments: Each sensor unit in this module is carefully designed to have good environmental adaptability. The laser scanning unit can work under different lighting conditions. The GNSS positioning unit has a certain anti-interference ability. The inertial measurement unit is not affected by external light and electromagnetic interference. The vision sensor unit can still normally collect images under certain dust and bad weather conditions. This adaptability to complex environments enables the data acquisition module to operate stably at various construction sites. Whether indoors or outdoors, during the day or at night, in sunny or bad weather, it can reliably collect the required data, ensuring the versatility and reliability of the positioning system.

[0072] As the front-end data acquisition component of the high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion, the data acquisition module provides comprehensive, accurate, and real-time data support for the entire system through its advanced sensor unit composition, efficient data acquisition and synchronization mechanism, and significant advantages. It is a key link to achieve high-precision surveying and setting-out and intelligent positioning, and strongly promotes the intelligent development of building surveying technology.

[0073] In this embodiment, the data fusion and processing module is connected to the data acquisition module and sequentially executes:

[0074] Multi-source data spatio-temporal alignment unit: Adopt the double-reference alignment algorithm of Beidou timestamp and ORB-SLAM3 time axis to establish a unified spatio-temporal coordinate system.

[0075] Feature-level fusion unit: Execute the improved hybrid Kalman filter-graph attention network model to generate a fused point cloud map.

[0076] Sensor failure detection unit: Real-time monitor the health status of each sensor and dynamically adjust the fusion strategy.

[0077] Furthermore, the data fusion and processing module is the core hub of the high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion. It relies on various types of raw data collected by the data acquisition module and uses a series of advanced technologies and algorithms to achieve the deep fusion and intelligent processing of multi-source data, providing key support for subsequent positioning calculation, setting-out operations, etc., and playing a decisive role in improving the positioning accuracy and reliability of the entire system. Specifically:

[0078] I. Hardware and data support:

[0079] Multi-sensor data access device: This module has powerful data access capabilities and can stably connect to the laser scanning unit, GNSS positioning unit, inertial measurement unit, and vision sensor unit. These sensors transmit the collected 3D laser point cloud data, satellite positioning data, inertial navigation data, and visual image data to the data fusion and processing module in real-time, providing rich raw information for data fusion. At the same time, the data access device has a high-speed data transmission interface to ensure the stability and timeliness of data transmission, avoid data loss or delay, and guarantee the timeliness of data.

[0080] Data preprocessing and storage capabilities: After receiving multi-source data, the data fusion and processing module first performs data preprocessing. It performs operations such as denoising and filtering on the raw data to remove noise and outliers caused by sensor errors, environmental interference, etc., and improve data quality. In addition, the module has a large-capacity data storage capability and can temporarily store the preprocessed data for subsequent fusion calculations and analyses. These stored data not only support the current survey and setting-out tasks but also provide historical data for subsequent review and data analysis.

[0081] II. Composition of functional modules:

[0082] Multi-source data spatio-temporal alignment unit:

[0083] Implementation of dual-reference alignment algorithm: The Beidou timestamp and the ORB-SLAM3 time axis dual-reference alignment algorithm are adopted to establish a unified spatio-temporal coordinate system. The Beidou timestamp has the characteristics of high precision and high reliability and can provide an accurate time reference for all sensor data. The ORB-SLAM3 time axis performs excellently in the field of visual positioning and has unique advantages in the time processing of visual data. This unit combines the advantages of both. First, it uses the Beidou timestamp to mark the time of various sensor data, and then finely calibrates the visual image data according to the ORB-SLAM3 time axis. Through this process, it ensures that the data collected by different sensors are accurately matched in time and space, laying a solid spatio-temporal foundation for subsequent data fusion.

[0084] Spatio-temporal calibration and synchronization mechanism: To ensure the accuracy and stability of spatio-temporal alignment, the module establishes a strict spatio-temporal calibration and synchronization mechanism. Regularly calibrate the time and space parameters of each sensor to ensure the accuracy of the timestamp and the consistency of the space coordinate system. During the data acquisition process, monitor the time and space deviations of sensor data in real-time. Once the deviation exceeds the threshold, immediately adjust and synchronize to ensure the coherence and accuracy of the data in time and space, providing a reliable data basis for subsequent fusion processing.

[0085] Feature-level fusion unit:

[0086] Improved model for fusing data: Implement an improved hybrid Kalman filter - graph attention network model to deeply fuse multi-source data and generate a fused point cloud map. This model combines the advantages of Kalman filtering in dynamic system state estimation and the ability of the graph attention network to mine multi-modal data features. Through the state update formula , comprehensively considering the state value at the previous moment, the current control input, and the measurement value, dynamically adjust the Kalman gain matrix based on sensor confidence, mine the correlation and complementarity between different sensor data, and generate a fused point cloud map containing rich environmental information, where is the fused state estimation value at time , is the online calibrated state transition matrix; is the Kalman gain matrix dynamically adjusted based on sensor confidence; is the graph network coupling coefficient; is the cross-modal attention weight; is the trainable parameter matrix; is the state value at time is the control input at time is the measurement value at time is the measurement matrix; is the set of neighbor nodes of sensor node i; is the neighbor node 's state value;

[0087] Cross-modal attention weight calculation: The cross-modal attention weight is calculated using a multi-sensor feature cross-validation algorithm, and the formula is ; This algorithm comprehensively considers factors such as the Euclidean distance between sensor nodes, the dimension of the feature vector, the environmental interference attenuation factor, and the sensor type encoding vector, accurately calculates the cross-modal attention weight, enables the model to adaptively focus on the key information in different sensor data, and improves the fusion effect, where , are the query / key vector projection matrices; is the Euclidean distance of the sensor node; is the environmental interference attenuation factor; , are the sensor type encoding vectors; is the sensor node 's feature vector; is the dimension of the feature vector, ; is the normalized exponential function;

[0088] Sensor failure detection unit:

[0089] Real-time health monitoring: Real-time monitor the health status of each sensor. By analyzing the characteristics of the data collected by the sensor, the data fluctuation situation, and comparing with historical data, judge whether the sensor is working properly; Once it is found that the sensor data shows abnormal fluctuations, exceeds the reasonable range, or has an obvious conflict with the data of other sensors, immediately start the fault diagnosis process to determine whether the sensor fails and the degree of failure;

[0090] Fault tolerance control strategy: When the th sensor fails, automatically adjust the fusion algorithm to:

[0091] , where, is the failure indication factor, and , represents that the th sensor fails completely; is the set of neighbor nodes excluding the failed sensor. Through this adjustment, ensure that even if some sensors fail, the system can still rely on the data of other normal sensors to continue the fusion process and maintain the stable operation of the system;

[0092] III. Advantages of application scenarios:

[0093] Improve positioning accuracy: Through spatio-temporal alignment and feature-level fusion of multi-source data, fully integrate the advantageous information of different sensors to generate a more accurate fused point cloud map; In the intelligent positioning and solution module, based on the accurate fused data, adopt the RTK / PPK dual-mode differential positioning algorithm and the VIO-SLAM algorithm assisted by laser point cloud, which can calculate the absolute coordinates and relative positioning coordinates more accurately, effectively eliminate the influence of abnormal observations, and output high-precision three-dimensional space coordinates to meet the strict requirements of building surveying and setting out for positioning accuracy;

[0094] Enhance system reliability: The sensor failure detection unit monitors the sensor status in real time. When a sensor fails, adjust the fusion strategy in time to ensure the continuity and reliability of data fusion; This enables the system to still operate stably in a complex building construction environment in the face of sudden situations such as sensor failures, avoiding positioning errors or system failures caused by sensor problems and ensuring the smooth progress of the surveying and setting out work;

[0095] Adapting to complex environmental changes: At a construction site, the environment is complex and changeable, such as signal occlusion, electromagnetic interference, etc.; the data fusion processing module can integrate multi-source data and utilize the advantages of different sensors in different environments. For example, when satellite signals are occluded, it relies on data from inertial measurement units and vision sensors for positioning; in an electromagnetic interference environment, laser scan data is not affected and can still provide reliable information for positioning. This multi-source data fusion method enables the system to better adapt to complex environmental changes and improves the adaptability and robustness of the system.

[0096] Intelligent positioning and solution module: Receives the fused point cloud map and is connected to the inertial measurement unit of the data acquisition module, configured as:

[0097] Calculates the absolute coordinates through the RTK / PPK dual-mode differential positioning algorithm;

[0098] Calculates the relative positioning coordinates through the VIO-SLAM algorithm assisted by laser point cloud;

[0099] Applies a robust estimation model to eliminate the influence of abnormal observations and outputs three-dimensional space coordinates;

[0100] Furthermore, the intelligent positioning and solution module is one of the core components of the high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion. It undertakes the key task of converting the fused data into accurate positioning information in the entire system; relying on the fused point cloud map generated by the data fusion processing module and combining the data of the inertial measurement unit in the data acquisition module, it uses a variety of advanced positioning algorithms and models to achieve high-precision positioning of the target and provides an accurate position basis for subsequent setting-out operations. Specifically:

[0101] I. Foundation of data input and processing:

[0102] Fused data access: The intelligent positioning and solution module receives the fused point cloud map from the data fusion processing module. This map integrates data collected by the laser scanning unit, GNSS positioning unit, inertial measurement unit, and vision sensor unit, containing rich environmental information and features; the fused point cloud map provides a comprehensive and accurate environmental model for positioning and solution, enabling the module to more accurately determine the relationship between the target position and the surrounding environment; at the same time, the module is directly connected to the inertial measurement unit of the data acquisition module to obtain inertial navigation data in real time. These data record the acceleration and angular velocity information of the device and can reflect the changes in the motion state of the device, providing dynamic motion parameters for positioning and solution.

[0103] Data preprocessing mechanism: Before performing positioning calculation, the intelligent positioning calculation module preprocesses the input data. For the fused point cloud map, feature extraction and screening are carried out to highlight key features related to positioning, such as prominent terrain contours and feature points, reduce data redundancy, and improve calculation efficiency. For the data of the inertial measurement unit, filtering and denoising processing are carried out to remove noise signals generated by sensor errors or environmental interference, ensuring the accuracy and stability of the data. Through these preprocessing steps, a high-quality data basis is provided for the subsequent positioning algorithm, ensuring the accuracy and reliability of the positioning calculation.

[0104] II. Positioning algorithm and model:

[0105] RTK / PPK dual-mode differential positioning algorithm: The RTK (Real-Time Kinematic) / PPK (Post-Processing Kinematic) dual-mode differential positioning algorithm is used to calculate the absolute coordinates. The RTK technology uses the carrier phase observations between the reference station and the rover for real-time differential processing, and can obtain centimeter-level positioning accuracy in the open field environment in real time. In building surveying and setting out, when the satellite signals are good and there is a suitable reference station around, the RTK mode can quickly determine the absolute position of the measuring device, providing a real-time positioning reference for the subsequent measurement and setting out operations. The PPK technology, on the other hand, improves the positioning accuracy by post-processing the observation data collected by the reference station and the rover after the measurement is completed. In some areas where the signals are easily blocked or interfered, such as between high-rise buildings in the city or in mountainous areas, etc., the RTK may not be able to obtain high-precision positioning results in real time. At this time, the PPK technology can play its advantages. Through the fine processing of the data afterwards, the errors can be effectively eliminated, and high-precision absolute coordinates can be obtained to make up for the deficiencies of the RTK in complex environments. The intelligent positioning calculation module flexibly switches between the RTK and PPK modes according to the actual environment and positioning requirements, ensuring accurate absolute coordinates can be obtained in various scenarios.

[0106] VIO-SLAM Algorithm Aided by Laser Point Cloud: Calculate the relative positioning coordinates through the VIO-SLAM (Visual-Inertial Simultaneous Localization and Mapping) algorithm aided by laser point cloud; the VIO-SLAM algorithm combines the data of visual sensors and inertial measurement units, uses visual information for environmental perception and feature matching, and at the same time predicts the motion state of the device with the measurement data of the inertial measurement unit, so as to achieve simultaneous localization and mapping; in the building survey scene, when satellite signals are blocked or unavailable, the VIO-SLAM algorithm can play an important role; and the assistance of laser point cloud further enhances the positioning ability of the algorithm; the three-dimensional laser point cloud data obtained by the laser scanning unit provides more accurate environmental geometric information, complementing the visual image data; when performing feature matching and positioning calculations, the high-precision spatial information of the laser point cloud can help the algorithm more accurately determine the position and attitude changes of the device, improving the accuracy and stability of relative positioning; for example, in areas where satellite signals cannot cover, such as inside buildings or underground spaces, using the VIO-SLAM algorithm aided by laser point cloud can calculate the position and attitude of the measurement device relative to known environmental features in real time, achieving accurate relative positioning;

[0107] Robust Estimation Model: Apply the robust estimation model to eliminate the influence of abnormal observations; in the actual positioning measurement process, due to environmental interference, sensor failures and other reasons, abnormal observations may occur. If these abnormal values are not processed, they will seriously affect the positioning accuracy; the robust estimation model identifies abnormal observations through statistical analysis of the observed data and assigns them smaller weights, thereby reducing the impact of abnormal values on the positioning result; the model can adaptively adjust the weights and determine the contribution of each observation in the positioning solution according to the reliability of the data; for example, when the data of a certain sensor shows abnormal fluctuations, the robust estimation model will automatically reduce the weight of this data in the calculation, ensuring that the final positioning result is not interfered by it and outputting stable and reliable three-dimensional space coordinates;

[0108] III. Advantages of Application Scenarios:

[0109] Guarantee of High-Precision Positioning: The intelligent positioning solution module comprehensively uses a variety of advanced positioning algorithms and models, combines the advantages of multi-source data, and can achieve high-precision positioning; whether in open field sites, complex urban building environments or indoor spaces, it can provide accurate position information for surveying and setting out operations; in the surveying and setting out of large bridge construction, this module can accurately determine the position of bridge piers, with the error controlled within centimeters or even smaller, meeting the strict requirements of engineering construction for high-precision positioning and ensuring the accuracy and quality of engineering construction;

[0110] Strong adaptability to complex environments: In the construction environment, there are often complex situations such as signal occlusion and electromagnetic interference. Traditional single positioning methods often struggle to meet the requirements. The intelligent positioning calculation module can achieve reliable positioning in various complex environments by integrating multiple positioning technologies. For example, it uses RTK / PPK to handle open environments and VIO-SLAM assisted by laser point clouds to process signal occlusion areas. When conducting building surveys in densely populated urban high-rise areas, even when satellite signals are severely occluded, accurate measurement point positions can still be obtained through the fusion positioning of laser point clouds with vision and inertial data, ensuring the smooth progress of the survey work.

[0111] Balancing real-time performance and stability: When performing positioning calculations, the module can process data in real time and quickly output positioning results. Whether it is real-time RTK positioning or relative positioning based on VIO-SLAM assisted by laser point clouds, it can meet the real-time requirements of on-site operations. At the same time, the application of the robust estimation model ensures the stability of the positioning results. Even in the presence of abnormal observation values, reliable three-dimensional space coordinates can be continuously output, providing stable positioning support for surveying and setting-out operations, improving work efficiency and reliability.

[0112] In this embodiment, the setting-out path generation module communicates with the intelligent positioning calculation module and includes:

[0113] BIM model parsing unit: Converts the engineering BIM model into three-dimensional path constraint conditions in the construction coordinate system.

[0114] Path planning unit: Generates the setting-out path based on the multi-objective optimization algorithm.

[0115] Furthermore, the setting-out path generation module is a key component of the high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion. It is responsible for planning a reasonable setting-out path according to engineering requirements and positioning information, playing an important connecting role in the entire surveying and setting-out process and directly affecting the efficiency and accuracy of the setting-out work. Specifically:

[0116] I. Module composition and data interaction:

[0117] BIM Model Analysis Unit: This unit serves as a bridge connecting engineering design and actual lofting operations. It takes the engineering BIM (Building Information Modeling) model as the core input and, through professional analysis algorithms, converts the rich design information contained in the BIM model into three-dimensional path constraint conditions in the construction coordinate system. The BIM model contains detailed information such as the precise geometric shape of the building, spatial location, and the relationships between various components. The analysis unit extracts the coordinates of lofting points, lofting sequence, connection relationships between points, and various limiting conditions during the construction process, such as the location of obstacles and the boundaries of construction areas. These constraint conditions provide a solid foundation for subsequent path planning, ensuring that the generated lofting path meets the requirements of engineering design. For example, in large construction projects, the BIM model analysis unit can accurately obtain the design coordinates of the building's foundation pile positions and information about obstacles such as surrounding underground pipelines, thus providing precise constraint basis for lofting path planning.

[0118] Path Planning Unit: The path planning unit is based on the three-dimensional path constraint conditions output by the BIM model analysis unit. Combining with the actual situation of the construction site, it uses multi-objective optimization algorithms to generate the optimal lofting path. This unit works closely with the intelligent positioning and solution module to obtain real-time positioning data so as to consider the current equipment position and moving direction when planning the path. During the planning process, it comprehensively considers multiple factors, such as the total path length, maximum path curvature, safety score, etc., to achieve multi-objective optimization. Its objective function is , where , is the total path length, and a shorter path can reduce lofting time and cost; , is the maximum path curvature, and a smaller curvature can ensure smoother equipment movement and reduce equipment wear; , is the safety score function, ensuring that the path avoids dangerous areas and guarantees construction safety; , is the dynamic weight coefficient; is the path planning scheme; At the same time, strict constraint conditions are set, such as to ensure positioning accuracy, to limit the path slope and avoid equipment operating on paths with too large slopes, to ensure path safety, where is the positioning accuracy of path ; is the maximum slope of path ; is the minimum value of the safety score; is the curvature of the path at moment; is the safety-related distance of the -th point on the path; is the number of points on the path; by continuously optimizing these parameters, the optimal lofting path that meets the engineering requirements is generated;

[0119] II. Algorithm and Optimization Strategy:

[0120] Core of the multi-objective optimization algorithm: The multi-objective optimization algorithm is the core of the path planning unit; in practical applications, there are often conflicts among objectives such as path length, curvature, and safety. For example, the shortest path may pass through dangerous areas, while the safe path may be longer; this algorithm uses dynamic weight coefficients to balance the importance of different objectives; according to the actual engineering situation and different requirements at the construction stage, the weight coefficients are dynamically adjusted; in the initial stage of construction, the safety requirement is relatively high, and the weight of the safety scoring function can be appropriately increased; in the later stage of construction, if more attention is paid to efficiency, the path length weight can be increased; in this way, the optimal balance is found among multiple objectives, and the lofting path that meets the actual requirements is generated;

[0121] Iterative optimization process: When generating the lofting path, the path planning unit adopts the method of iterative optimization; first, an initial path is generated according to the initial constraint conditions and weight coefficients; then, the path is continuously adjusted and optimized by the algorithm, and the value of the objective function is gradually reduced; in each iteration, the influence of the position change of each point on the path on the objective function is considered, such as adjusting the order or position of a lofting point, and observing the changes in path length, curvature, and safety; after multiple iterations, the path gradually approaches the optimal solution; when encountering a complex construction environment or multiple conflicting constraint conditions, a reasonable lofting path that takes into account multiple factors can be found through repeated iterative optimization;

[0122] III. Advantages of Application Scenarios:

[0123] Improve lofting efficiency: By accurately analyzing the BIM model and optimizing the path planning, the lofting path generation module can plan the shortest and most convenient lofting path; reduce the moving distance and time of the equipment on the construction site, avoid unnecessary detours and repeated operations, and greatly improve the efficiency of the lofting work; in the survey and lofting of large sites, compared with the traditional manual path planning, the path generated by using this module can save a lot of time and speed up the construction progress;

[0124] Ensure construction safety: The lofting path generated by the module fully considers safety factors. Through the safety scoring function and relevant constraint conditions, it ensures that the path avoids dangerous areas on the construction site, such as deep foundation pits, high-voltage equipment, and high-altitude operation areas; this effectively reduces the safety risks during the construction process and ensures the safety of construction personnel and equipment; in some complex construction sites, this module can provide a safe and reliable path planning for the lofting operation and reduce the occurrence of safety accidents;

[0125] Ensure lofting accuracy: Based on the precise analysis of the BIM model and strict positioning accuracy constraints, the generated lofting path can ensure that the positioning accuracy of the lofting points is within the specified range; this enables the actual lofting results to highly coincide with the design requirements, avoiding construction deviations caused by lofting errors and improving the project quality; in construction projects with extremely high precision requirements, such as bridges, high-rise buildings, etc., this module can ensure the accurate position of each lofting point, providing a reliable foundation for subsequent construction.

[0126] In this embodiment, the wireless communication module is connected to each module and configured as:

[0127] 5G communication unit: Upload data to the cloud platform at a rate of ≥1 Gbps;

[0128] LoRa self-organizing network unit: Establish a Mesh network in the signal occlusion area to achieve direct communication between devices;

[0129] Furthermore, the wireless communication module is an important part of the high-precision surveying and lofting intelligent positioning system based on multi-source data fusion. It is like the "nervous network" of the system, responsible for transmitting data between each module and between the system and external devices, ensuring the smooth flow of information, and playing a crucial role in the efficient operation of the system; by integrating multiple communication technologies, this module can adapt to different construction environments and data transmission requirements, providing stable and efficient data communication guarantee for the surveying and lofting work. Specifically:

[0130] I. Communication technology integration:

[0131] 5G communication unit: The 5G communication unit is an important part of the wireless communication module. It has the communication advantages of high speed, low latency, and large capacity; in this system, the 5G communication unit uploads data to the cloud platform at a rate of ≥1 Gbps; this high-speed data transmission ability enables the system to quickly upload a large amount of collected data (such as 3D laser point cloud data, visual image data, etc.) to the cloud for storage, analysis, and processing; during the building survey process, a large amount of real-time data collected on-site can be quickly transmitted to the cloud server through the 5G network, and professional personnel can perform real-time analysis on this data in the cloud, discover problems in a timely manner, and make decisions; at the same time, the low-latency feature ensures the timeliness of data transmission. For operations that require real-time response (such as remotely controlling equipment to adjust the position, etc.), 5G communication can ensure that the instructions are conveyed quickly and accurately, improving the real-time performance and response speed of the system; in addition, the large-capacity communication ability of 5G can meet the needs of multi-device simultaneous online communication and adapt to the complex construction site environment;

[0132] LoRa Self-Organizing Network Unit: In construction sites, signal blockage is quite common. For example, inside buildings, underground spaces, or areas blocked by large construction equipment, traditional communication methods may be affected. The LoRa self-organizing network unit addresses this issue by establishing a Mesh network in signal-blocked areas to enable direct communication between devices. The LoRa technology features low power consumption and long-distance transmission, capable of achieving stable communication in complex environments. Through the self-organizing characteristics of the Mesh network, devices can automatically establish connections and transmit data. Even if some nodes fail or signals are blocked, the data can be relayed through other nodes to ensure the reliability of communication. For instance, during the survey inside a large building, multiple positioning devices can establish a Mesh network through the LoRa self-organizing network unit to achieve real-time data sharing and interaction, ensuring that each device can work collaboratively and improving the efficiency and accuracy of surveying and setting out.

[0133] II. Data Transmission and Network Management:

[0134] Data Transmission Optimization: The wireless communication module classifies and transmits data according to the type and urgency of the data. For data with high real-time requirements (such as real-time positioning data, device control instructions, etc.), it is preferentially transmitted through the 5G communication unit to ensure the rapid transfer of data and the real-time response of the system. For some non-real-time data (such as historical data backup, large-scale point cloud data storage, etc.), it can be transmitted through 5G or the LoRa self-organizing network unit when the network load is low, rationally utilizing network resources and avoiding network congestion. At the same time, the module adopts data compression and encryption technologies to reduce the data transmission volume, improve the transmission efficiency, and ensure the security of the data under the premise of ensuring data integrity, preventing the data from being stolen or tampered with during transmission.

[0135] Network Switching and Management: During the actual construction process, the device may move in different areas, resulting in changes in the communication environment. The wireless communication module has an intelligent network switching function and can automatically switch between 5G and LoRa self-organizing networks according to factors such as signal strength and network quality. When the device is in an area with good signal, it preferentially uses the 5G network for high-speed data transmission. When the device enters a signal-blocked area, it automatically switches to the LoRa self-organizing network to ensure the continuity of communication. In addition, the module also manages the network in real time, monitors parameters such as the network connection status, signal strength, and data transmission rate, discovers and resolves network failures in a timely manner, and ensures the stability of system communication.

[0136] III. Advantages of Application Scenarios:

[0137] Adapt to complex construction environments: The construction site environment is complex, and signal interference and occlusion problems are prominent. The wireless communication module can achieve reliable communication in different environments by integrating 5G and LoRa self-organizing network technologies. Whether it is using the high-speed communication advantage of 5G in an open area or leveraging the flexible networking ability of LoRa self-organizing network in areas with signal occlusion, it can ensure that data transmission between system modules is not affected, guaranteeing the smooth progress of surveying and setting out work. In urban construction projects where high-rise buildings are surrounded and signals are easily blocked, the LoRa self-organizing network unit can play an important role in ensuring smooth communication between devices. In open areas with good signals, the 5G communication unit can quickly transmit a large amount of data, improving work efficiency.

[0138] Support remote monitoring and collaboration: Relying on the high speed and low latency characteristics of 5G communication, the wireless communication module supports remote monitoring and collaboration functions. Managers and technical experts can remotely and real-time view the surveying and setting out situation at the construction site through the cloud platform, obtain real-time positioning data, equipment status information, etc., and conduct remote collaboration with on-site staff. When encountering complex problems, experts can remotely guide on-site operations and solve problems in a timely manner, improving work efficiency and quality. This way of remote monitoring and collaboration breaks geographical restrictions, makes full use of the professional resources of all parties, and enhances the overall efficiency of the system.

[0139] Improve system integration and flexibility: The integrated design of the wireless communication module integrates multiple communication technologies, reduces external connection devices of the system, and improves the system's integration. At the same time, its flexible network configuration and adaptive ability enable the system to be quickly adjusted and deployed according to different construction requirements and environments. In construction projects of different scales and types, this wireless communication module can be conveniently applied, providing convenient conditions for the wide application of the system.

[0140] In this embodiment, the visualization terminal module receives the setting out path and real-time positioning data, including:

[0141] AR display unit: Superimpose and display the theoretical / measured position deviation vector;

[0142] Dynamic deviation correction unit: Execute three-level feedback control to achieve real-time path correction;

[0143] Furthermore, the visualization terminal module is a key part directly facing users in the high-precision surveying and setting out intelligent positioning system based on multi-source data fusion. As the window for interaction between the system and users, it presents data and provides operation functions in an intuitive and convenient visual way, significantly improving the user experience and work efficiency, making the surveying and setting out work more intelligent and efficient. Specifically:

[0144] I. Hardware and data interaction foundation:

[0145] Data receiving interface: The visualization terminal module enables stable data interaction with the intelligent positioning and calculation module and the lofting path generation module. This interface can efficiently receive real-time positioning data and lofting path data, ensuring that the information obtained by the terminal is always synchronized with the actual on-site situation. Whether it is the real-time position change of the positioning device at the construction site or the dynamically adjusted lofting path according to engineering requirements, it can be transmitted to the visualization terminal module in a timely and accurate manner, providing users with the latest work information.

[0146] Display and interaction hardware: This module is equipped with high-performance display devices with high resolution and good color restoration, capable of clearly and delicately displaying various data information. At the same time, combined with interactive hardware such as touch screens, users can interact with the terminal through simple touch operations to view and analyze data and call system functions, greatly improving the convenience and efficiency of operation.

[0147] II. Core function analysis:

[0148] AR display unit: The AR display unit is one of the core functions of the visualization terminal module. It uses augmented reality technology to fuse virtual information with the real scene for display. In the survey and lofting work, this unit can superimpose and display the theoretical / measured position deviation vector. Through AR devices (such as AR glasses or tablets supporting AR functions), users can intuitively see the deviation between the actual measurement position and the theoretical lofting position. This intuitive visualization presentation method allows users to quickly judge whether the current position is accurate and the direction and distance that need to be adjusted without complex calculations and analyses. When lofting the foundation pile positions of a building, after construction workers wear AR devices, they can directly see the deviation vector between the current pile position measurement point and the designed pile position, clearly understand the deviation size and direction, and thus quickly make adjustments to improve the lofting accuracy and work efficiency.

[0149] Dynamic deviation correction unit: The dynamic deviation correction unit executes three-level feedback control based on the deviation information provided by the AR display unit to achieve real-time path correction. When the deviation value Δ between the theoretical position and the measured position is in different ranges, the system will adopt different control strategies.

[0150] First-level control: When , The interface displays a yellow warning box. At this time, the deviation is in a relatively small range, but still needs attention. The appearance of the yellow warning box reminds users that there is a certain deviation at the current position, which requires attention and timely adjustment to avoid further expansion of the deviation.

[0151] Second-level control: When When this occurs, the system not only displays deviation information in the AR interface, but also generates local path correction instructions and notifies the user through vibration alerts; vibration alerts can attract the user's high attention, while local path correction instructions provide specific operation guidance for the user, helping the user quickly adjust the device position and bring the measurement point back to the correct lofting path;

[0152] Tertiary control: When this occurs, the deviation is relatively large, which may have a greater impact on the project quality; at this time, the system starts system self-check and freezes the positioning output until manual confirmation; system self-check can timely detect possible problems, such as sensor failures, abnormal positioning algorithms, etc., and freezing the positioning output avoids the further transmission and use of incorrect data; the manual confirmation link ensures that the positioning and lofting work continues only after the problem is properly handled, guaranteeing the accuracy and safety of the project;

[0153] Among them, is the deviation value between the theoretical position and the measured position;

[0154] III. Advantages of application scenarios:

[0155] Improve construction accuracy and efficiency: The AR display and dynamic deviation correction functions of the visualization terminal module enable construction personnel to intuitively and in real time grasp the deviation between the measurement position and the theoretical position and make timely adjustments; this greatly reduces rework and correction work caused by position deviation, improves construction accuracy and efficiency; in large-scale construction projects, accurate lofting work can ensure that the positions and dimensions of all parts of the building meet the design requirements, avoid structural problems caused by lofting errors, shorten the construction period, and reduce construction costs;

[0156] Lower the operation threshold and training costs: Through the intuitive visualization interface and simple operation methods, even relatively inexperienced construction personnel can quickly get started; they do not need to deeply understand complex positioning data and algorithms, and only need to operate according to the deviation information displayed by AR and system prompts; this lowers the operation threshold, reduces the dependence on professional technical personnel, and also reduces training costs, enabling new employees to adapt to the work position faster;

[0157] Enhance construction safety and quality control: The tertiary control strategy of the dynamic deviation correction unit, especially the system self-check and freezing of positioning output functions activated when the deviation is relatively large, effectively avoids construction accidents caused by incorrect operations or equipment failures, ensuring construction safety; at the same time, real-time deviation monitoring and correction ensure that the construction process always meets the design requirements, improving the level of project quality control and providing strong guarantees for the quality and safety of construction projects.

[0158] In this embodiment, when multiple devices work together, a distributed consensus algorithm is used to synchronize positioning data:

[0159] , where is the convergence rate parameter, is the positioning data of the device at moment; is the positioning data of the device at moment; is the positioning data of the neighbor device of the device at moment; is the Kalman gain; is the measurement value of the device ; is the measurement matrix;

[0160] Furthermore, in the high-precision surveying and setting-out intelligent positioning system based on multi-source data fusion of the present invention, the collaborative work of multiple devices is the key link to improve the overall performance of the system and achieve efficient and accurate surveying and setting-out; the collaborative operation of multiple devices covers aspects such as data synchronization, task allocation and cooperation, and fault tolerance processing. Through these mechanisms, each device closely cooperates to ensure the stable operation of the positioning system. Specifically:

[0161] Data synchronization:

[0162] Principle of distributed consensus algorithm: When multiple devices work together, the distributed consensus algorithm is used to synchronize positioning data. The formula is , where the convergence rate parameter determines the speed of data synchronization, and it controls the frequency of information interaction between devices; when is larger, the information update between devices is faster, but it may increase the network burden; on the contrary, the information update speed is slow, but the network pressure is small; in practical applications, it needs to be reasonably adjusted according to the network conditions and positioning accuracy requirements;

[0163] Data synchronization process: In the actual scenario, each device continuously collects data and exchanges positioning information with adjacent devices through this algorithm;

[0164] Task allocation and cooperation:

[0165] Task Allocation Strategy: The system allocates tasks according to the functional characteristics and actual working status of each device. For example, the laser scanning device is good at acquiring high-precision three-dimensional point cloud data. When conducting surveys of complex terrains or building structures, it is mainly responsible for collecting detailed spatial data. The GNSS positioning device has high positioning accuracy and can obtain absolute coordinates, playing a key role in open areas or when determining the overall position reference. The inertial measurement unit and the vision sensor unit assist in positioning and supplement environmental information in cases where satellite signals are interfered or in indoor environments.

[0166] Collaborative Operation Process: Taking a large bridge construction project as an example, in the early site survey stage, the GNSS positioning device first determines the approximate position and provides the reference coordinates for other devices. Subsequently, the laser scanning device scans the bridge site area to obtain the three-dimensional point cloud data of the terrain and existing buildings. During the construction process, when satellite signals are blocked, the inertial measurement unit and the vision sensor unit start to play their roles to ensure the continuity and accuracy of positioning. The data collected by each device is transmitted to the data fusion processing module in real time. After fusion processing, it provides more accurate data for the intelligent positioning and calculation module, thus realizing high-precision surveying and setting-out operations.

[0167] Fault Tolerance and Handling:

[0168] Fault Detection and Notification: During collaborative work, each device monitors the status of itself and adjacent devices in real time. Once a fault (such as sensor failure, communication interruption, etc.) is detected in a certain device, the faulty device will immediately send a fault notification to other devices. At the same time, the sensor failure detection unit of the system will also assist in judging device faults by monitoring abnormal changes in the data and take corresponding measures in a timely manner.

[0169] Implementation of Fault Tolerance Strategy: When a certain device fails, the system adjusts the working strategy according to the type and severity of the fault. If only a certain sensor fails, the system will adjust the data fusion algorithm according to the strategy of the sensor failure detection unit, reduce the dependence on the data of the faulty sensor, and rely on other normal sensors to maintain the positioning and data collection functions. If a certain device completely fails, the system will re-allocate tasks, reasonably allocate the tasks of the faulty device to other normal devices, and ensure that the entire system can still continue to operate and complete the surveying and setting-out tasks.

[0170] The high-precision surveying and setting-out intelligent positioning method based on multi-source data fusion relies on a highly integrated and powerful system. Through the collaborative collection, deep fusion, intelligent calculation, path planning, and real-time feedback correction of multi-source data, it realizes high-precision positioning during the building surveying and setting-out process, significantly improving the efficiency and quality of engineering construction. The specific steps include:

[0171] Step S1: Multi-source data acquisition and preprocessing: Use the laser scanning unit, GNSS positioning unit, inertial measurement unit, and vision sensor unit in the data acquisition module to synchronously acquire 3D laser point cloud data, satellite positioning data, inertial navigation data, and visual image data of the construction site; these sensors are like the "perceptual antennas" of the system, obtaining information about the construction site from different dimensions; the acquired data will be time-stamped aligned to ensure the consistency of each data source in the time dimension, providing an accurate time reference for subsequent fusion processing; at the same time, perform preprocessing operations such as denoising and filtering on the original data to remove noise and outliers caused by sensor errors, environmental interference, etc., and improve data quality;

[0172] Step S2: Data fusion and spatio-temporal alignment: The data fusion processing module adopts the double-reference alignment algorithm of Beidou timestamp and ORB-SLAM3 time axis to establish a unified spatio-temporal coordinate system, and accurately match the multi-source data in time and space; use an improved hybrid Kalman filter-graph attention network model to perform feature-level fusion on the multi-source data; through this model, comprehensively consider the characteristics and correlations of the data of each sensor, mine the key information in the data, and generate a fused point cloud map; in the fusion process, the calculation of cross-modal attention weights adopts a multi-sensor feature cross-validation algorithm to ensure that the model can adaptively focus on the key information in the data of different sensors and improve the fusion effect;

[0173] Step S3: Intelligent positioning solution: The intelligent positioning solution module receives the fused point cloud map and combines the data of the inertial measurement unit, and uses a variety of algorithms to achieve high-precision positioning; use the RTK / PPK dual-mode differential positioning algorithm to solve the absolute coordinates, and flexibly switch between RTK (real-time kinematic differential positioning) and PPK (post-processing kinematic differential positioning) modes according to the satellite signal conditions at the construction site to ensure accurate absolute coordinates can be obtained in various scenarios; use the laser point cloud-assisted VIO-SLAM algorithm to calculate the relative positioning coordinates. In areas where satellite signals are blocked or lost, such as inside buildings or canyons, this algorithm fuses visual and inertial data to achieve continuous positioning and tracking; apply a robust estimation model to eliminate the influence of abnormal observations and ensure the stability and reliability of the positioning results, and finally output accurate three-dimensional space coordinates;

[0174] Step S4, Lofting Path Generation and Optimization: The BIM model parsing unit in the lofting path generation module converts the engineering BIM model into three-dimensional path constraint conditions in the construction coordinate system, extracts the coordinates of lofting points, the lofting sequence, the connection relationships between points, and the constraint conditions during the construction process; the path planning unit generates a lofting path based on these constraint conditions using a multi-objective optimization algorithm; this algorithm comprehensively considers factors such as the total path length, the maximum path curvature, and the safety score, and finds the optimal balance among multiple objectives by dynamically adjusting the weight coefficients to generate the optimal lofting path that meets the engineering requirements;

[0175] Step S5, Real-time Feedback and Dynamic Deviation Correction: The visualization terminal module receives the lofting path and real-time positioning data; the AR display unit superimposes and displays the theoretical / measured position deviation vector, enabling the user to intuitively understand the deviation between the current position and the theoretical lofting position; the dynamic deviation correction unit performs three-level feedback control based on the deviation value; when the deviation value is small, a yellow warning box is displayed on the AR interface; when the deviation value is slightly larger, a local path correction instruction is generated and vibration warning is given; when the deviation value is too large, the system self-check is started and the positioning output is frozen until manual confirmation; through this real-time feedback and dynamic deviation correction mechanism, it is ensured that the lofting process always maintains high precision;

[0176] Step S6, Multi-device Collaboration and Data Synchronization: When multiple devices work together, a distributed consensus algorithm is used to synchronize the positioning data to ensure the real-time performance and consistency of the positioning data of each device; task allocation is performed according to the functional characteristics and actual working status of each device, such as the laser scanning device is responsible for collecting spatial data, and the GNSS positioning device determines the absolute coordinates, etc.; the devices cooperate with each other, share data in real time, and jointly complete the surveying and lofting tasks; when a certain device fails, the system can detect it in time and adjust the working strategy, and ensure that the entire system can still operate normally by adjusting the data fusion algorithm or reallocating tasks.

[0177] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. High-precision survey and layout intelligent positioning system based on multi-source data fusion, characterized by: include: Data acquisition module: including laser scanning unit, GNSS positioning unit, inertial measurement unit and visual sensor unit, synchronously collects 3D laser point cloud data, satellite positioning data, inertial navigation data and visual image data of the construction site, and outputs raw data streams with time stamp alignment; Data fusion processing module: connected to the data acquisition module, and executes in sequence: Multi-source data spatiotemporal alignment unit: uses Beidou timestamp and ORB-SLAM3 time axis dual reference alignment algorithm to establish a unified spatiotemporal coordinate system; Feature-level fusion unit: executes the improved hybrid Kalman filter-graph attention network model to generate a fused point cloud map; Sensor failure detection unit: monitors the health status of each sensor in real time and dynamically adjusts the fusion strategy; Intelligent positioning solution module: receives the fused point cloud map and connects to the inertial measurement unit of the data acquisition module, and is configured as follows: The absolute coordinates are calculated through the RTK / PPK dual-mode differential positioning algorithm; Calculate relative positioning coordinates through the VIO-SLAM algorithm assisted by laser point cloud; Apply the robust estimation model to eliminate the influence of abnormal observations and output three-dimensional space coordinates; Lofting path generation module: communicates with the intelligent positioning solution module, including: BIM model analysis unit: converts the engineering BIM model into three-dimensional path constraints in the construction coordinate system; Path planning unit: Generates the layout path based on the multi-objective optimization algorithm; Wireless communication module: connected with each module, configured as: 5G communication unit: upload data to the cloud platform at a rate of ≥1Gbps; LoRa self-organizing network unit: establish a Mesh network in the signal blocking area to achieve direct communication between devices; Visual terminal module: receives the layout path and real-time positioning data, including: AR display unit: superimposes and displays theoretical / actual position deviation vectors; Dynamic deviation correction unit: performs three-level feedback control to achieve real-time path correction; The feature-level fusion unit adopts an improved hybrid Kalman filter-graph attention network model, and its state update formula is: ,in, for Fusion state estimates at all times; , is the state transfer matrix of online calibration; A Kalman gain matrix that is dynamically adjusted based on sensor confidence; is the graph network coupling coefficient; is the cross-modal attention weight; is the trainable parameter matrix; for The state value at the moment; for Control input at all times; for The measured value at the moment; is the measurement matrix; is the set of neighbor nodes of sensor node i; Neighbor node The status value of The multi-objective optimization algorithm executed by the path planning unit includes: ； Constraints: ; in, , is the total length of the path; , is the maximum curvature of the path; , is the safety scoring function; , is the dynamic weight coefficient; To plan a route; For path Positioning accuracy; For path The maximum slope of is the minimum value of the safety score; For the path The curvature of the moment; For the path Safety-related distances of points; is the number of points on the path.

2. The high-precision survey and setting-out intelligent positioning system based on multi-source data fusion according to claim 1 is characterized by: The cross-modal attention weight is calculated using a multi-sensor feature cross-validation algorithm: ,in, , is the query / key vector projection matrix; is the Euclidean distance of sensor nodes; is the environmental interference attenuation factor; , Encode vector for sensor type; For sensor nodes The eigenvector of is the dimension of the feature vector, ; is the normalized exponential function.

3. The high-precision survey and setting-out intelligent positioning system based on multi-source data fusion according to claim 1 is characterized by: When the sensor failure detection unit executes the fault-tolerant control strategy: When the first When a sensor fails, the fusion algorithm is automatically adjusted to: ,in, is the failure indicator factor, and , Indicates A sensor completely failed; It is the set of neighbor nodes that exclude failed sensors.

4. The high-precision survey and setting-out intelligent positioning system based on multi-source data fusion according to claim 1 is characterized by: The dynamic correction unit performs three-level control, including: Primary Control , The interface displays a yellow warning box; Secondary Control , generate local path correction instructions and vibrate alarms; Three-level control , start the system self-check and freeze the positioning output until manual confirmation; in, is the deviation between the theoretical position and the measured position.

5. The high-precision survey and setting-out intelligent positioning system based on multi-source data fusion according to claim 1 is characterized by: When multiple devices work together, a distributed consensus algorithm is used to synchronize positioning data: ,in, is the convergence rate parameter, For equipment exist Positioning data at the moment; For equipment exist Positioning data at the moment; For equipment Neighboring devices exist Positioning data at the moment; is the Kalman gain; For equipment The measured value of is the measurement matrix.

6. A high-precision survey and layout intelligent positioning method based on multi-source data fusion, characterized by: The high-precision survey and layout intelligent positioning method is based on the high-precision survey and layout intelligent positioning system based on multi-source data fusion described in any one of claims 1-5.

Images