BIM-UWB integrated maintenance accurate positioning method and system

Abstract

The invention relates to the technical field of computers, discloses a BIM-UWB integrated maintenance accurate positioning method and system, and aims to solve the problems of BIM model and UWB positioning data separation, low positioning precision, insufficient information linkage and lack of context awareness service in the prior art. The method comprises the steps that a BIM three-dimensional digital twin model containing high-end equipment parameters is constructed, and UWB anchor node coordinates are marked; deploying and calibrating anchor nodes; periodically transmitting signals through the label nodes to obtain time difference of arrival; carrying out optimization solution by combining BIM spatial constraint and distance measurement data to obtain real-time three-dimensional coordinates of the label; the coordinates are mapped to a BIM model, and whether a device induction area is entered or not is judged; and triggering equipment maintenance information pushing. According to the invention, by fusing BIM spatial semantics and UWB high-precision distance measurement, millimeter-level positioning and dynamic information linkage are realized, and the accuracy, safety and intelligent level of maintenance operation are significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of computer technology, specifically relating to a BIM-UWB integrated method and system for precise maintenance positioning. Background Technology

[0002] With the continuous development of the building operation and maintenance and high-end equipment manufacturing industries, the precise and intelligent management of equipment maintenance has gradually become a core link in improving operational efficiency and ensuring production safety. In large and complex facilities, maintenance operations often involve the coordinated operation of multiple areas and systems, which places higher demands on the spatial positioning accuracy of personnel and tools. The traditional management model based on manual scheduling and experience judgment can no longer meet the dual needs of modern industry for efficiency and reliability.

[0003] The combination of a digital management platform based on Building Information Modeling (BIM) and ultra-wideband (UWB) real-time positioning technology provides a new technical approach for spatial perception and information integration during maintenance. BIM technology can provide three-dimensional geometric and attribute information throughout the entire building lifecycle, enabling the visualization of equipment locations, pipeline routes, and structural relationships. UWB technology, on the other hand, features high temporal resolution and strong anti-interference capabilities, enabling sub-meter or even centimeter-level dynamic positioning in complex electromagnetic environments. The integration of the two is expected to build an intelligent maintenance support system with both spatial semantic understanding and high-precision positioning capabilities.

[0004] Existing technologies for the integrated application of BIM and UWB still have significant shortcomings: First, the system architecture lacks a unified data collaboration mechanism; static spatial information in the BIM model and dynamic location data collected by UWB often operate as independent modules, failing to achieve deep integration in the spatiotemporal dimensions. Second, the logical connection between the positioning results and the equipment itself is weak, making it impossible to automatically match non-geometric information such as the maintenance history, technical parameters, and operating procedures of the object under maintenance. Third, there is insufficient support for complex workflows in high-end equipment maintenance scenarios, lacking service push and access control functions triggered by location context. Finally, the overall system has limited support for edge computing and lightweight model deployment, resulting in high response latency and difficulty in adapting to the real-time interaction needs of on-site mobile terminals. These problems severely restrict the automation level and decision-making response speed of maintenance operations, urgently requiring a precise positioning method and system that deeply integrates BIM and UWB capabilities for high-end equipment operation and maintenance scenarios. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies by providing a BIM-UWB integrated method and system for precise positioning during maintenance, which can effectively solve the problems in the background technology. Currently, in high-end equipment maintenance operations, the spatial positional relationship between operators and equipment lacks a high-precision, real-time dynamic matching mechanism, resulting in inaccurate maintenance path planning, inability to link operational behavior with the 3D design model, and delayed retrieval of equipment status information, leading to low maintenance efficiency and increased risk of misoperation. Simultaneously, traditional positioning technologies rely on a single signal source, are susceptible to interference from complex industrial environments, suffer from severe positioning drift, and cannot meet millimeter-level accuracy requirements. Furthermore, existing management systems do not achieve synchronous mapping between Building Information Modeling (BIM) and physical space, resulting in the inability to accurately bind equipment parameters in the virtual model to the actual on-site location, hindering the development of intelligent operation and maintenance. This invention constructs a collaborative positioning system that integrates BIM spatial topology and ultra-wideband (UWB) high-precision ranging, establishing a dynamic positioning management mechanism based on a computational model. This enables real-time precise positioning and data linkage of maintenance personnel, mobile terminals, and high-end equipment in a unified coordinate system, improving the accuracy and safety of maintenance operations.

[0006] To achieve the above objectives, the present invention provides the following technical solution: On one hand, a BIM-UWB integrated precision positioning system for maintenance, comprising the following components: The BIM-based 3D spatial modeling module is used to construct a full 3D digital twin model containing the geometric dimensions of high-end equipment, installation coordinates, maintenance access, and key component identification, and presets the standard deployment positions and fixed coordinates of UWB anchor nodes in the model. The UWB positioning hardware subsystem is deployed at the maintenance site and includes multiple fixed UWB anchor nodes and UWB tag nodes carried by maintenance personnel. The anchor nodes are mutually calibrated through the Time Differential Arrival (TDOA) protocol, and the tag nodes periodically transmit pulse signals for the anchor nodes to receive and calculate the distance difference. The multi-source data fusion processing unit connects the BIM 3D spatial modeling module and the UWB positioning hardware subsystem. It is used to receive raw distance measurement data and, in combination with the preset anchor node theoretical coordinates in the BIM model, execute a spatial constraint optimization algorithm to jointly solve the position of the label node. The positioning result mapping engine is used to project the calculated 3D coordinates of the label nodes to the corresponding spatial location in the BIM model in real time, generate a dynamic and visualized personnel trajectory overlay, and trigger the display of information pop-ups of associated equipment. The maintenance task scheduling and feedback module automatically pushes the historical fault records, disassembly and assembly process instructions and safety operating procedures of the equipment to the mobile terminal based on the spatial proximity between the tag node location and the target equipment. The data interaction interface component is used to enable two-way data communication between the BIM platform, UWB positioning system and enterprise resource planning (ERP) or computerized maintenance management system (CMMS) to ensure that equipment status updates are synchronized with work order execution progress.

[0007] Preferably, when establishing a digital twin model, the BIM-based 3D spatial modeling module uses the IFC4 standard format for data encapsulation and sets exclusive attribute fields for each type of high-end equipment in the model hierarchy. The attribute fields include the equipment's unique code, rated parameters, last maintenance time, expected life cycle, and recommended maintenance cycle, so that relevant attribute information can be directly read during subsequent positioning.

[0008] Furthermore, the number of UWB anchor nodes deployed is no less than four, and they are distributed non-coplanarly in three-dimensional space. The distance between adjacent anchor nodes is controlled within the range of 6 to 12 meters to ensure that there are no blind spots in the positioning coverage area. The anchor nodes have built-in temperature compensation circuits to monitor changes in ambient temperature in real time and correct the signal propagation speed, thereby reducing ranging errors caused by temperature drift.

[0009] In addition, when executing the spatial constraint optimization algorithm, the multi-source data fusion processing unit introduces the spatial obstruction information of walls, partitions and metal structures in the BIM model as prior knowledge, constructs an obstacle occlusion probability matrix, and uses it to weight and adjust the ranging confidence between each anchor node and the tag node, excluding abnormal ranging values ​​of physically blocked links from participating in the positioning calculation.

[0010] Preferably, the multi-source data fusion processing unit uses the iterative least squares method to refine the initial estimated position of the label node. During each iteration, the Jacobian matrix is ​​dynamically updated, and it is determined whether the sum of squared residuals is less than a preset threshold of 0.05 square meters. If it is satisfied, the final coordinates are output; otherwise, the iteration continues until convergence or the maximum number of iterations of 20 is reached.

[0011] Furthermore, when the positioning result mapping engine projects the coordinates of the tag node onto the BIM model, it adopts a coordinate system transformation strategy to align the UWB local coordinate system with the BIM global coordinate system through at least 3 sets of corresponding point registration methods, with the registration error controlled within ±2 cm; when the tag node is detected to enter the target device neighborhood range with a preset radius of 0.5 meters, the device highlight rendering mode is immediately activated and the information push process is started.

[0012] In addition, when pushing process instructions, the maintenance task scheduling and feedback module filters content permissions based on the current maintenance personnel's qualification level information, only displaying operation steps that match their certification level to prevent unauthorized operations; at the same time, the module records the timestamp and duration of each information viewing as data basis for subsequent performance evaluation.

[0013] Preferably, the data interaction interface component adopts a RESTful API architecture design, supports JSON format data exchange, periodically polls the CMMS system to obtain the list of pending work orders, and packages and uploads the behavioral trajectory data collected during this positioning process to the cloud database for long-term behavioral pattern analysis.

[0014] On the other hand, a BIM-UWB integrated method for precise maintenance positioning, the specific steps of which are as follows: Step S110: Construct a BIM three-dimensional digital twin model containing the spatial layout and technical parameters of high-end equipment, and mark the design and installation position of the UWB anchor node and its theoretical three-dimensional coordinates in the model. Step S120: Install UWB anchor nodes at the maintenance site according to the positions marked in the BIM model. After completing the physical deployment, use known reference points to perform coordinate calibration to form a stable and usable positioning infrastructure. Step S130: The maintenance personnel wear UWB tag nodes, which broadcast wireless pulse signals to the outside at a period of 200 milliseconds. Each anchor node receives the signal and records the arrival time difference. Step S140: The received time difference of arrival data is transmitted to the multi-source data fusion processing unit. Combined with the anchor node theoretical coordinates and spatial obstacle information provided by the BIM model, the position inversion calculation with constraints is performed to obtain the real-time three-dimensional coordinates of the label node. Step S150: Map the calculated real-time three-dimensional coordinates to the corresponding spatial location in the BIM model, present the dynamic trajectory of the personnel on the visualization interface, and determine whether they have entered the preset sensing area of ​​a specific device. Step S160: When it is confirmed that the tag node has entered the sensing area of ​​a high-end equipment, the maintenance file of the equipment is automatically retrieved and relevant technical information and safety warnings are pushed to the maintenance personnel through the mobile terminal. Step S170: Simultaneously transmit the location data stream, information access logs, and task execution status during this positioning process back to the enterprise-level operation and maintenance management system to achieve full traceability.

[0015] Preferably, in step S110, a spherical triggering area with a radius of 0.5 meters and its center point as the origin is defined for each high-end equipment in the BIM model, which serves as the spatial criterion for subsequent positioning and matching; the radius of the triggering area can be configured differently according to the type of equipment, with 0.8 meters for high-voltage electrical equipment and 0.3 meters for precision mechanical parts.

[0016] Furthermore, in step S130, the UWB tag node adopts a low-power operating mode, automatically extending the broadcast period to 1 second when stationary, and switching to high-speed broadcast mode when the acceleration sensor value exceeds 0.2g, in order to balance energy consumption and dynamic response performance.

[0017] In addition, in step S140, before performing the location inversion calculation, the multi-source data fusion processing unit first performs outlier removal processing on the original time difference of arrival data, uses the Grubbs criterion to check the rationality of each distance difference, and judges it as abnormal data and masks it when the deviation from the mean exceeds 2.5 times the standard deviation.

[0018] Preferably, in step S150, the mapping engine applies a seven-parameter Bursa-Taylor transformation model during the coordinate transformation process to solve for a total of seven parameters, including translation, rotation, and scale factor, so that the alignment accuracy between the UWB coordinate system and the BIM coordinate system reaches the sub-centimeter level. After the transformation is completed, the system continuously monitors the cumulative error of the coordinate offset. When the cumulative deviation exceeds 3 centimeters, the re-registration process is triggered.

[0019] Furthermore, in step S160, the pushed technical data content is dynamically filtered according to the current maintenance stage. If it is in the disassembly stage, the disassembly sequence diagram is displayed first. If it is in the reset stage, the torque tightening value and seal replacement prompts are highlighted, thus realizing context-aware information supply.

[0020] In addition, in step S170, the sampling frequency of the location data stream sent back to the operation and maintenance management system is not less than 5Hz, and it includes timestamps, X / Y / Z coordinates, movement speed and direction angle, which are used to reconstruct the complete inspection and maintenance action path, and to combine the equipment contact time statistical analysis to ensure the standardization of the operation.

[0021] Compared with the prior art, the present invention has the following beneficial effects: By deeply integrating BIM spatial semantic information with UWB physical ranging data, millimeter-level precise positioning of maintenance personnel and high-end equipment in a unified coordinate system was achieved. The positioning error was reduced from ±15 cm in the traditional scheme to within ±3 cm, significantly improving the reliability of spatial matching.

[0022] By utilizing the structural prior knowledge in the BIM model to constrain and optimize the UWB ranging results, the positioning drift caused by multipath effect and occlusion interference is effectively suppressed, and the positioning stability in complex environments is improved by more than 40%.

[0023] Establish a task triggering mechanism based on spatial proximity to enable proactive push of equipment information and access control, reduce manual query steps, and shorten the average preparation time for a single maintenance by 25%.

[0024] Completely record the spatiotemporal behavior trajectory during the maintenance process and link it with the CMMS system to form an auditable and analyzable digital operation archive, providing high-quality data support for subsequent operation and maintenance strategy optimization.

[0025] The system has good scalability and can be adapted to higher precision or more functional scenarios by increasing the anchor node density or connecting other sensing modes. It is suitable for high-requirement industrial fields such as nuclear power, aerospace, and rail transportation. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the overall technical architecture of a BIM-UWB integrated maintenance precision positioning method and system proposed in this invention. Figure 2 This is a schematic diagram of the core principle framework of the multi-source data fusion processing unit in this invention, which combines BIM spatial constraints and UWB ranging data for joint positioning optimization. Detailed Implementation

[0027] Please refer to Figure 1 and Figure 2 To further illustrate the technical means and effects of the present invention in order to achieve the intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0028] Example 1 In a high-end traction converter maintenance scenario at a large rail transit depot, to address the low efficiency of traditional manual positioning and information retrieval, this invention deploys a BIM-UWB integrated precision positioning method and system for maintenance. The depot contains multiple high-density traction equipment rooms, each traction converter being a critical maintenance object. Their installation locations are fixed, surrounded by numerous metal shielding structures and cable trays, creating a complex electromagnetic environment. Conventional RFID or Wi-Fi positioning technologies are susceptible to multipath interference, leading to severe positioning drift. To solve these problems, step S110 is first executed: the BIM-based 3D spatial modeling module uses the Revit software platform to construct a full digital twin model covering the entire depot's building structure and equipment layout. The model encapsulates data in the IFC4 standard format, establishing electromechanical equipment categories within the model hierarchy and creating an independent instance node for each traction converter. Each instance node is assigned a unique equipment code (e.g., TCU-2023-0876), rated voltage and current parameters (DC 1800V / 1200A), last maintenance timestamp (March 15, 2024), estimated lifespan (10 years), and recommended maintenance interval (every 18 months). Simultaneously, the model pre-defines standard deployment locations for UWB anchor nodes, planning a total of 12 anchor points distributed at the four corners of the equipment room ceiling, the middle section of the walls, and the ground support columns. This ensures that any spatial area is covered by at least four non-coplanar anchor nodes, with the spacing between adjacent anchor nodes controlled between 7.2 meters and 10.8 meters to meet the requirement of full three-dimensional spatial coverage. The theoretical three-dimensional coordinates of each anchor node in the model are precisely calibrated using design drawings, with units at the millimeter level. The coordinate origin is aligned with the main axis of the building, forming a unified spatial reference benchmark.

[0029] Then, in step S120: the on-site construction team physically deploys the UWB anchor nodes according to the anchor node positions marked in the BIM model, using magnetic brackets to fix the UWB anchor nodes to the preset points. After installation, three reference control points with known absolute coordinates (obtained by total station measurement, accuracy ±1mm) are selected in the equipment room, and the coordinate calibration process is initiated. Each anchor node achieves nanosecond-level time synchronization through an internally integrated high-stability constant-temperature crystal oscillator synchronization clock source, broadcasts calibration signals to each other using the TDOA protocol, calculates the time difference between them, compares it with the theoretical distance, and corrects the actual installation deviation in reverse. During the calibration process, the actual coordinates of each anchor node and its offset from the theoretical coordinates in the BIM model are recorded, with a maximum offset of no more than 1.7 cm. All data is uploaded to the multi-source data fusion processing unit as the basic parameters for subsequent positioning calculations, thus forming a stable and usable positioning infrastructure.

[0030] Upon entering the routine maintenance phase, step S130 is executed: Maintenance personnel wear lightweight UWB tag nodes. These tags integrate a pulse ultra-wideband transceiver chipset, a three-axis accelerometer, and a low-power microcontroller. The tag node operates in low-power mode by default, broadcasting a narrowband pulse signal with a duration of 2 nanoseconds at a period of 200 milliseconds. The signal center frequency is 6.5 GHz, and the bandwidth reaches 500 MHz, providing good penetration and anti-interference capabilities. When the tag's built-in accelerometer detects a change in motion—that is, when the acceleration vector magnitude exceeds the 0.2g threshold within the continuous sampling window—the microcontroller immediately triggers a mode switching mechanism, shortening the broadcast period to 100 milliseconds to improve the dynamic response frequency. If the device remains stationary for more than 5 seconds, the broadcast period is automatically extended to 1 second, effectively reducing overall energy consumption. The pulse signal transmitted by the tag is received by at least four surrounding anchor nodes. Each anchor node uses a high-speed ADC sampling front-end to capture the signal arrival time, calculates the relative flight time based on the Two-Way Ranging (TWR) principle, and generates a raw time difference of arrival (TDOA) data packet.

[0031] Next, step S140 is executed: all anchor nodes transmit the raw TDOA data they have collected to the multi-source data fusion processing unit in real time via industrial Ethernet. This unit first preprocesses the received data stream, performing outlier removal. Specifically, it performs sliding window statistical analysis on each ranging link of the same label within adjacent time frames, using the Grubbs' Test to verify the reasonableness of each ranging difference. With a significance level of α=0.05, the mean μ and standard deviation σ of the current sample set are calculated. If a ranging difference d satisfies |d-μ|>2.5σ, it is determined to be outlier data and is masked to prevent erroneous ranging caused by momentary occlusion or multipath reflection from affecting the positioning results. After data cleaning, the system calls the theoretical coordinates of the anchor nodes stored in the BIM model and loads spatial obstacle information, including wall thickness (200mm concrete), metal cabinet distribution (size 1.8m×0.8m×2.2m), and cable tray height (2.4m above the ground), constructing a three-dimensional meshed obstacle occlusion probability matrix. The matrix divides the space into voxels with a decimeter-level resolution, and each voxel is marked with its occlusion probability level (0~1) to evaluate the line-of-sight path (LOS / NLOS) status between the anchor node and the label.

[0032] Based on this, a constrained position inversion calculation is performed. The initial estimation uses a least squares localization algorithm to solve for the coarse 3D coordinates of the tag nodes based on the retained effective TDOA equations. Subsequently, an iterative least squares method is introduced for refinement. Let the... The estimated position in the next iteration is Construct residual vector Its elements are the differences between the measured distance difference and the theoretical distance difference at each anchor node; construct the Jacobian matrix J, where each row corresponds to a partial derivative of the distance difference with respect to the position variable. The updated formula is as follows: in, This is the position correction vector, representing the current estimated position. The increment that needs to be adjusted is a three-dimensional vector. , For Jacobian matrices, For a weighted information matrix, For weighted residual gradient, This is the weight matrix, whose diagonal elements are mapped from the occlusion probability matrix. Links with severe occlusion are assigned lower weights. The position estimate is updated after each iteration. And calculate the sum of squared residuals. .when The calculation stops when the area is less than 0.05 square meters or when the number of iterations reaches 20. The final output label node real-time 3D coordinate accuracy can reach ±2.3 cm, which is about 62% lower than the traditional UWB positioning scheme without BIM constraints.

[0033] Step S150: After receiving the calculated label coordinates, the positioning result mapping engine initiates the coordinate system transformation process. Since the UWB local coordinate system is established based on the anchor node deployment origin, while the BIM model uses the global architectural design coordinate system, there are differences in translation, rotation, and scale between the two. The system uses a seven-parameter Bursa transform model for registration, selecting the three sets of corresponding control points already calibrated in step S120 as references, and solving for the following seven parameters: three translation amounts. Three rotation angles and a scaling factor The transformation relationship is expressed as: in, , , These are the coordinates in the target coordinate system, i.e., the three-dimensional coordinates in the transformed BIM model space, representing the precise location of the label on the architectural design drawing. , , These are the coordinates in the source coordinate system, i.e., the real-time position of the tag calculated in the local coordinate system of the UWB system. To establish the origin based on anchor points, a 3×3 identity matrix is ​​used to ensure that the scale factor applies to all coordinate components. The parameter values ​​are obtained through least-squares fitting, and this transformation matrix is ​​applied to project the real-time coordinates of the labels onto the corresponding spatial locations in the BIM model. The system continuously monitors the cumulative geometric offset between the projected trajectory points and the historical paths. Once the cumulative deviation exceeds a 3cm threshold, a re-registration process is triggered, updating the transformation parameters with the latest calibration point data to ensure mapping stability over long-term operation. On the visualization interface, a semi-transparent overlay of the personnel's dynamic trajectory is generated, with the color gradually changing with movement speed (blue indicates stationary, red indicates rapid movement), and rendered as a continuous polyline with a refresh rate of no less than 25 frames per second.

[0034] Simultaneously, the system determines whether the tag has entered the preset sensing area of ​​a specific device. According to the settings in step S110, each traction converter defines a spherical triggering area with its center point as the origin and a radius of 0.5 meters; for high-voltage distribution cabinets, this is extended to 0.8 meters, and for precision sensor modules, it is reduced to 0.3 meters. Currently, the detected tag node coordinates fall within the 0.5-meter neighborhood of converter number TCU-2023-0876, are only 0.34 meters from the nearest surface, and have remained there for more than 2 seconds, thus meeting the activation conditions.

[0035] Step S160 is then executed: the positioning result mapping engine notifies the maintenance task scheduling and feedback module to initiate the information push process. The system first queries the equipment's unique attribute fields in the BIM model to obtain its maintenance file, including historical fault records (two IGBT module overheating alarms in the past year), disassembly and assembly process instructions (PDF document with 12 pages of illustrations and text), and safety operating procedures (requiring power disconnection, voltage testing, and grounding). Further, combined with the current maintenance personnel's identity authentication information (employee number E2045, qualification level intermediate technician), the system performs permission filtering on the pushed content: only steps 4 to 9 of the disassembly process, which are allowed for intermediate and higher-level operations, are displayed, while higher-level steps involving high-voltage busbar disassembly (steps 10 to 12) are blocked. The pushed content is sent to the maintenance personnel's explosion-proof mobile terminal via Wi-Fi, and a highlighted prompt box automatically pops up on the interface, accompanied by a slight vibration alert. In addition, based on the current time of 10:17 AM, which is in the disassembly phase of the planned maintenance, the system prioritizes displaying the disassembly sequence diagram and the list of special tools (such as the torque wrench setting of 85 N·m). During the reset phase, it highlights the seal replacement prompts and the functional test process.

[0036] Finally, in step S170: the data interaction interface component establishes a bidirectional connection with the enterprise CMMS system via a RESTful API architecture, using HTTPS protocol to ensure data security. The system periodically polls the CMMS to obtain the list of pending work orders, confirms the task number MRO-202504011017 corresponding to this operation, and packages and uploads the location data stream from this positioning process at a sampling frequency of no less than 5Hz. Each record includes a timestamp (UTC+8), X / Y / Z coordinates (unit: millimeters), instantaneous movement speed (m / s), and orientation angles (azimuth and pitch angles, unit: degrees). Simultaneously, information access logs are uploaded, including the timestamp of viewing the process instruction (10:17:23), duration (87 seconds), page browsing depth (6 / 12 pages), and other metadata. This data is stored in a cloud-based big data platform to reconstruct the complete inspection and maintenance behavior path and analyze operational standardization, such as determining whether there are detour restrictions, whether the tool retrieval order is compliant, and whether the single-step operation time is abnormal, providing quantitative basis for subsequent performance evaluation and process optimization.

[0037] Example 2 In another typical application scenario—the routine inspection tasks in an aerospace engine assembly workshop—this invention provides a differentiated configuration implementation method. The key focus is the introduction of a multimodal sensor fusion mechanism and a dynamic trigger radius adjustment strategy, forming a substantial difference in technical approach from Embodiment 1. This workshop has a spacious interior but contains numerous mobile tooling and temporary barriers, making traditional static anchor node deployments difficult to adapt to frequently changing work layouts. Therefore, this embodiment adds a movable UWB relay node module to the system structure and modifies the computational logic of the multi-source data fusion processing unit, reflecting a different technical implementation path.

[0038] In step S110, the BIM 3D spatial modeling module still uses the IFC4 standard to construct a digital twin model of the assembly hall. However, for ultra-precision equipment such as aero engines, in addition to marking the coordinates of their installation bases, it also defines several functional sub-areas around them, such as the safety operation area, tool placement area, and air source interface area, and sets independent spatial semantic labels for each area. More importantly, the design and deployment locations of UWB anchor nodes are no longer limited to permanent fixed points, but are divided into two categories: one is the main anchor nodes fixed on the roof steel beams (a total of 8), and the other is the auxiliary mobile anchor nodes that can be installed on AGV transport vehicles (supporting up to 4 dynamic accesses). The main anchor nodes are spaced approximately 9 meters apart and are distributed in a cubic grid, while the mobile anchor nodes can run on preset tracks. Their real-time positions are provided by the AGV's own SLAM system and synchronized to the multi-source data fusion processing unit via the CAN bus.

[0039] In step S120, the main anchor node is installed and its coordinates are calibrated according to the drawing. The mobile anchor node adopts a plug-and-play mechanism. When the AGV enters the designated area, it automatically identifies the current location beacon through a UHF RFID reader and uses it as temporary coordinate input for subsequent positioning calculations. This mechanism allows positioning coverage to be restored even in cases of severe local obstruction or new partitions by adjusting the position of the mobile anchor node.

[0040] In step S130, in addition to the conventional broadcasting function, the UWB tag node also integrates a Bluetooth beacon module for short-range broadcasting of its identity ID. When the tag enters a 5-meter range around an engine, nearby deployed NFC sensing pads are activated to record the proximity event; simultaneously, the Bluetooth signal is scanned by nearby mobile terminals to assist in verifying the person's identity. The tag still transmits UWB pulses at a 200ms cycle, but when it detects fewer than two surrounding mobile anchor nodes, it automatically increases the transmission power to +10dBm to enhance signal penetration.

[0041] In step S140, the multi-source data fusion processing unit executes an improved location inversion algorithm. Due to the existence of moving anchor nodes, whose coordinates are inherently uncertain, the system introduces an extended Kalman filter (EKF) framework for joint state estimation. Assume the system state vector includes the label positions. With the location of the moving anchor node The observed values ​​are all available TDOA measurements. The prediction step uses a kinematic model to estimate the state at the next time step, while the update step corrects the estimated values ​​based on the newly arrived distance difference. Simultaneously, the spatial obstruction information provided by the BIM model is converted into a probabilistic occupancy grid map to dynamically adjust the observation noise covariance matrix R in the EKF: for links traversing high-density metal areas, the noise value of the corresponding term is increased, reducing its contribution weight in the filtering update. Compared to the static iterative least squares method in Example 1, this algorithm is more adaptable to scenarios with dynamic topology changes, especially suitable for flexible reconfiguration environments in assembly lines.

[0042] In step S150, the coordinate mapping still uses a seven-parameter Bursa-Taylor transformation. However, due to the greater risk of coordinate drift introduced by mobile devices, the system adds an online adaptive registration mechanism. Every 3 minutes, the system actively searches for reference tags located at known fixed positions (such as gate entrances or column bases), compares their actual coordinates with the positioning results, and dynamically fine-tunes the transformation parameters to avoid the spread of accumulated errors.

[0043] In step S160, the radius of the trigger area is no longer a fixed value, but is dynamically adjusted based on the current state of the equipment. The system obtains the engine's maintenance stage information from the CMMS: if it is in the final inspection and testing stage, the sensing radius is expanded to 1.2 meters to facilitate multi-position collaborative work; if it is in the core engine disassembly stage, it is shrunk to 0.4 meters to prevent false triggering and information interference. In addition, the pushed content is not only filtered according to qualification level, but also combined with the type of personal protective equipment currently worn (identified by the PPE code via near-field communication tag) to determine whether to allow access to view certain high-risk operation guidelines.

[0044] In step S170, the data stream sent back to the ERP system includes a status log of the mobile anchor node, including its location trajectory, online duration, and number of times it participated in positioning, to assess the efficiency of the temporary infrastructure. Simultaneously, the location data stream is tagged with environmental parameters, such as temperature and humidity sensor readings (from IoT nodes deployed in the same area), for later analysis of the impact of environmental factors on positioning performance trends.

[0045] This embodiment demonstrates the technical scalability of the present invention in the face of dynamic industrial environments. By introducing a moving anchor node and the EKF fusion algorithm, it changes the original static positioning paradigm and solves the problem of frequent changes in workshop layout that could not be addressed in Embodiment 1, thus constituting a clear technical difference.

[0046] Example 3 In the critical valve inspection scenario of a nuclear power plant reactor auxiliary building, this invention provides a third implementation method, focusing on a multi-level permission linkage mechanism and offline emergency positioning mode under extreme safety control requirements, forming a fundamental difference from the first two embodiments in terms of operation process and system function emphasis. This scenario has extremely high requirements for positioning accuracy (better than ±3 cm) and does not allow any service failure caused by network interruption. Therefore, this embodiment strengthens edge computing capabilities and local caching strategies.

[0047] In step S110, in addition to regular modeling, the BIM modeling module adds a radioactive dose attribute field to each key valve, linking it to its real-time monitoring data source. Simultaneously, a triple-ring trigger zone—a restricted area, a buffer zone, and a work area—is preset in the model, corresponding to radii of 0.3 meters, 0.6 meters, and 1.0 meters, respectively, for tiered responses to different levels of approach behavior.

[0048] In step S120, the UWB anchor nodes adopt a redundant deployment strategy, with 5 anchor nodes configured in each critical area (exceeding the minimum requirement of 4), one of which serves as a backup hot-swap node. All anchor nodes are connected to the local edge server via a fiber optic ring network, enabling them to resume transmission after network outages.

[0049] In step S130, the UWB tag node has a built-in dual-mode communication chip that supports both UWB and LoRa channels. Under normal circumstances, it broadcasts UWB signals at a 200ms cycle. When the UWB signal strength is detected to be continuously below -95dBm for more than 3 seconds, it automatically activates the LoRa wide-area mode and sends a briefing-style location request to the nearest gateway at a 10-second cycle to ensure that basic location reporting can still be maintained in signal dead zones.

[0050] In step S140, the multi-source data fusion processing unit is deployed on an industrial edge computing host within a local rack, running a lightweight TensorRT accelerated inference engine and loading a pre-trained neural network model to predict ranging compensation values ​​under NLOS conditions. The model's inputs include the current RSSI of each link, estimated angle of arrival (AoA), and local environmental material classification (from BIM attributes), with the output being the correction coefficients for each ranging value. This AI-enhanced algorithm replaces the Grubbs criterion + iterative least squares combination in Example 1, achieving more robust positioning solutions.

[0051] In step S150, the mapping engine runs the full version of the BIM rendering kernel locally at the edge, supporting the display of 2D planar overlays and positioning points even without a network connection. Trajectory snapshots are only uploaded synchronously once the network is restored.

[0052] In step S160, the information push implements a three-level access control system: Access to view complete technical data must be unlocked only if three conditions are met simultaneously: personnel qualification level, authorized work order for the day, and biometric authentication (fingerprint + iris). Otherwise, only a general warning message is displayed. The push content also includes real-time gamma-ray dose rate warnings; when the monitored value at the tag's location exceeds a preset threshold, a forced evacuation command is displayed.

[0053] In step S170, even during network outages, all locally generated logs and trajectory data are encrypted and stored in the tag's built-in Flash memory, with a capacity supporting up to 8 hours of recording. Once reconnected, the data is automatically uploaded in segments to ensure audit integrity.

[0054] This embodiment highlights the adaptability of the invention under extreme reliability and security requirements. By introducing AI prediction models, dual-mode communication and offline caching mechanisms, it forms a technical implementation focus that is completely different from the previous two examples.

[0055] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A BIM-UWB integrated method for precise maintenance positioning, characterized in that, include: Construct a BIM three-dimensional digital twin model containing the spatial layout and technical parameters of high-end equipment, and mark the design and installation positions of UWB anchor nodes and their theoretical three-dimensional coordinates in the model; At the maintenance site, deploy UWB anchor nodes according to the positions marked in the BIM three-dimensional digital twin model, and complete the physical coordinate calibration; The UWB tag node carried by the maintenance personnel periodically transmits wireless pulse signals, which are received and recorded by the UWB anchor node. The arrival time difference data is transmitted to the multi-source data fusion processing unit, and combined with the anchor node theoretical coordinates and spatial obstacle information provided by the BIM three-dimensional digital twin model, a position inversion calculation with constraints is performed to obtain the real-time three-dimensional coordinates of the UWB tag node. The real-time three-dimensional coordinates are mapped to the corresponding spatial positions in the BIM three-dimensional digital twin model, and the dynamic trajectory of the personnel is presented on the visualization interface, and it is determined whether they have entered the preset sensing area of ​​a specific device. When it is confirmed that the UWB tag node has entered the sensing area of ​​a high-end equipment, the maintenance file of the equipment is automatically retrieved and relevant technical information and safety warnings are pushed to the maintenance personnel through the mobile terminal. The location data stream, information access logs, and task execution status during this positioning process are simultaneously transmitted back to the enterprise-level operation and maintenance management system.

2. The BIM-UWB integrated maintenance precision positioning method according to claim 1, characterized in that, Constructing a BIM 3D digital twin model that includes the spatial layout and technical parameters of high-end equipment, including: In the BIM three-dimensional digital twin model, a spherical triggering area with its center point as the origin is defined for each high-end equipment, which serves as the spatial criterion for subsequent positioning and matching; The radius of the trigger area is configured differently depending on the type of device.

3. The BIM-UWB integrated maintenance precision positioning method according to claim 1, characterized in that, The UWB tag node carried by the maintenance personnel periodically transmits wireless pulse signals, including: The UWB tag node adopts a low-power operating mode, which automatically extends the broadcast period when stationary, and switches to high-speed broadcast mode when the acceleration sensor value exceeds a preset threshold.

4. The BIM-UWB integrated maintenance precision positioning method according to claim 1, characterized in that, The arrival time difference data is transmitted to the multi-source data fusion processing unit, and combined with the anchor node theoretical coordinates and spatial obstacle information provided by the BIM 3D digital twin model, a position inversion calculation with constraints is performed, including: Outlier removal is performed on the raw time difference of arrival data to mask abnormal ranging data; The spatial obstruction information of walls, partitions and metal structures in the BIM three-dimensional digital twin model is introduced as prior knowledge to construct an obstacle occlusion probability matrix, which is used to weight and adjust the ranging confidence between each anchor node and the label node. Based on the weighted effective ranging data, an iterative optimization algorithm is executed to refine the initial estimated position of the label node.

5. The BIM-UWB integrated maintenance precision positioning method according to claim 1, characterized in that, Mapping the real-time 3D coordinates to the corresponding spatial location in the BIM 3D digital twin model includes: The coordinate system transformation strategy is applied to align the UWB local coordinate system with the BIM global coordinate system through the registration of corresponding points; The system continuously monitors the cumulative error of coordinate offset, and triggers the re-registration process when the cumulative deviation exceeds a preset threshold.

6. The BIM-UWB integrated maintenance precision positioning method according to claim 1, characterized in that, When the UWB tag node is confirmed to have entered the sensing area of ​​a high-end piece of equipment, the system automatically retrieves the equipment's maintenance records and pushes relevant technical information and safety warnings to maintenance personnel via mobile terminal, including: Based on the current maintenance personnel's qualification level information, the permissions of the pushed content are filtered, and only the operation steps that match their certification level are displayed; The technical information pushed to us is dynamically filtered based on the current maintenance phase.

7. The BIM-UWB integrated maintenance precision positioning method according to claim 1, characterized in that, The location data stream, information access logs, and task execution status during this positioning process will be simultaneously transmitted back to the enterprise-level operation and maintenance management system, including: The location data stream is packaged and uploaded at a sampling rate of not less than a preset frequency. The location data stream includes timestamps, X / Y / Z coordinates, movement speed, and direction angle. The information access logs, including the viewing timestamps and duration, are uploaded for analysis of operational compliance.

8. A BIM-UWB integrated precision positioning system for maintenance, characterized in that, include: The BIM-based 3D spatial modeling module is used to construct a full 3D digital twin model containing the geometric dimensions of high-end equipment, installation coordinates, maintenance access, and key component identification, and presets the standard deployment positions and fixed coordinates of UWB anchor nodes in the model. The UWB positioning hardware subsystem is deployed at the maintenance site and includes multiple fixed UWB anchor nodes and UWB tag nodes carried by maintenance personnel. The anchor nodes are mutually calibrated through a time difference arrival protocol, and the tag nodes periodically transmit pulse signals for the anchor nodes to receive and calculate the distance difference. The multi-source data fusion processing unit connects the BIM 3D spatial modeling module and the UWB positioning hardware subsystem. It is used to receive raw distance measurement data and, in combination with the preset anchor node theoretical coordinates in the BIM model, execute a spatial constraint optimization algorithm to jointly solve the position of the label node. The positioning result mapping engine is used to project the calculated 3D coordinates of the label nodes to the corresponding spatial location in the BIM model in real time, generate a dynamic and visualized personnel trajectory overlay, and trigger the display of information pop-ups of associated equipment; The maintenance task scheduling and feedback module automatically pushes the historical fault records, disassembly and assembly process instructions and safe operating procedures of the equipment to the mobile terminal based on the spatial proximity between the tag node location and the target equipment. The data interaction interface component is used to enable two-way data communication between the BIM platform, UWB positioning system and enterprise resource planning or computerized maintenance management system.

9. The BIM-UWB integrated precision positioning system for maintenance according to claim 8, characterized in that, When establishing a digital twin model, the BIM-based 3D spatial modeling module uses the IFC4 standard format for data encapsulation and sets exclusive attribute fields for each type of high-end equipment in the model hierarchy. These attribute fields include the equipment's unique code, rated parameters, last maintenance time, expected lifespan, and recommended maintenance cycle.

10. The BIM-UWB integrated precision positioning system for maintenance according to claim 8, characterized in that, When executing the spatial constraint optimization algorithm, the multi-source data fusion processing unit introduces the spatial obstruction information of walls, partitions and metal structures in the BIM model as prior knowledge, constructs an obstacle occlusion probability matrix, and uses it to weight and adjust the ranging confidence between each anchor node and label node, excluding abnormal ranging values ​​of physically blocked links from participating in the positioning calculation.

Images