Construction progress determination method and device based on artificial intelligence, and electronic equipment

Abstract

The application provides a construction progress determination method and device based on artificial intelligence and an electronic device, and belongs to the technical field of computers. The method comprises collecting the type and installation position information of a current hanging building component; and providing the information to a corresponding building object agent. After receiving the information, the building object agent verifies whether the position is within its own space range and whether the current process conforms to the pre-logic. If the verification is reasonable, the progress state of the building object agent is updated and synchronized to a construction organization agent. If the verification is not reasonable, the data is rejected and re-matching is triggered. The construction organization agent transfers the data to other building object agents for verification or dynamically creates a new agent until the matching is successful. Finally, the construction organization agent collects the progress states of the building object agents to generate the overall construction progress. Thus, without relying on a pre-set BIM model, distributed rationality verification and automatic error correction of construction behavior are achieved.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and more specifically, to a method, apparatus, and electronic device for determining construction progress based on artificial intelligence. Background Technology

[0002] During construction, the seamless transition between different work processes directly impacts the overall project schedule and cost control. Accurately understanding the progress of each process allows for the timely scheduling of subsequent processes, ensuring a smooth and orderly workflow.

[0003] Currently, construction sites often rely on manual statistics and experience-based judgments. For example, construction progress is estimated by observing visible indicators such as the frequency of tower crane hook movements, material transport volume, and the integrity of the scaffolding, combined with experience. Specifically, the intensity of hoisting operations is judged by the number of tower boom swings, or the building's topping-out stage is inferred from the progress of scaffolding dismantling. Furthermore, progress reporting is delayed, and the lack of real-time, automated monitoring methods can easily lead to problems such as material shortages or delays in procedures. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a construction progress determination method, device and electronic device based on artificial intelligence, so as to realize automatic and real-time monitoring of construction progress and improve the accuracy of monitoring of construction progress.

[0005] Firstly, an artificial intelligence-based method for determining construction progress is provided, applied to an intelligent agent set. The intelligent agent set includes multiple building object intelligent agents and at least one construction organization intelligent agent. Each building object intelligent agent corresponds to a building component or a construction location, used to verify the rationality of construction actions performed on that component or location. The construction organization intelligent agent coordinates the logical relationships between the various building object intelligent agents. The method includes: Data is collected from the site using sensors on the tower crane cluster. The data includes the type, unique identifier, and installation location information of the currently suspended building components. The on-site data is provided to a candidate building object intelligent agent; after receiving the on-site data, the candidate building object intelligent agent verifies whether the current construction behavior is reasonable. If the verification result is reasonable, update the progress status maintained by the candidate building object agent and synchronize the update result to the construction organization agent; If the verification result is unreasonable, the data will be rejected and a rematching process will be triggered. The on-site data will be transferred to other candidate building object agents for verification. The construction organization agent summarizes the progress status of each building object agent and generates the construction progress of the entire construction site.

[0006] Optionally, when all existing candidate building object agents are deemed unreasonable based on the field data, the method further includes: Generate an anomaly alarm and suspend automatic scheduling of the building component, pending manual intervention; Alternatively, the construction organization agent can create a new building object agent, initialize the properties of the new agent to take over the verification and progress management of the building component.

[0007] Optionally, providing on-site data to a candidate building object agent includes: Analyze the spatial coordinates of building component types and installation locations in the on-site data; Based on the type of building component and its corresponding spatial coordinates, the on-site data is input into the candidate building object intelligent agent that matches the type and spatial coordinates.

[0008] Optionally, verifying the reasonableness of the construction activities includes: Verify whether the installation location indicated by the installation location information is within the spatial range corresponding to the candidate building object intelligent agent; And, verify whether the current construction behavior conforms to the preset logical sequence of procedures; When both the spatial scope verification and the process logic sequence verification pass, the current construction behavior is deemed reasonable.

[0009] Optionally, verifying whether the current construction activity conforms to the preset sequence of procedures includes: The candidate building object intelligent agent sends a query request for the preceding process to the construction organization intelligent agent; The construction organization AI agent returns the completion status of the previous processes for the current component; The candidate building object intelligent agent determines whether the current construction behavior meets the preset process logic sequence based on the returned status.

[0010] Optionally, creating a new building object agent includes: Load basic attributes from a pre-set agent template; Initialize the verification rules and corresponding spatial range based on the characteristics of the current building components; Register the newly created agent to the agent set and immediately assign it to the rematching process for further verification.

[0011] Optionally, the rematching process may also include: Record a mismatch event for the candidate building object agent whose original data was provided. When the cumulative number of mismatch events for the same candidate building object exceeds a preset threshold, the construction organization agent determines that there is a deviation between the agent and the current routing strategy or its own configuration, and performs one of the following operations: Reduce the priority of this agent in the routing strategy so that it no longer prioritizes receiving data from this type of component; Alternatively, it can trigger configuration diagnostics for the agent, check for deviations in the definition of its verification rules or spatial range, and generate repair suggestions.

[0012] Optionally, the construction organization intelligent agent supports natural language interaction to receive natural language instructions from on-site management personnel; the method also includes: Perform semantic parsing and security permission verification on the received natural language instructions; If the verification passes, the verification rules or logical relationships of each building object's intelligent agent will be dynamically adjusted according to the instructions, and the adjustment results will be reported back to the management personnel.

[0013] Secondly, an artificial intelligence-based construction progress determination device is provided, applied to an intelligent agent set; the intelligent agent set includes multiple building object intelligent agents and at least one construction organization intelligent agent; each building object intelligent agent corresponds to a building component or a construction part, used to verify the rationality of construction behaviors acting on the building component or construction part; the construction organization intelligent agent is used to coordinate the logical relationships between the various building object intelligent agents; the device includes: The data acquisition module is used to collect on-site data through sensors on the tower crane cluster. The on-site data includes the type, unique identifier, and installation position information of the currently suspended building components. A module is provided to provide on-site data to a candidate building object intelligent agent; after receiving the on-site data, the candidate building object intelligent agent verifies whether the current construction behavior is reasonable. The update module is used to update the progress status maintained by the candidate building object agent if the verification result is reasonable, and synchronize the update result to the construction organization agent. The transfer module is used to reject the current data and trigger a rematch process if the verification result is unreasonable, and transfer the on-site data to other candidate building object agents for verification. The summary module is used to summarize the progress status of each building object intelligent agent through the construction organization intelligent agent, and generate the construction progress of the entire construction site.

[0014] Thirdly, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When the processor executes a program stored in memory, it implements any of the first aspects of the artificial intelligence-based construction progress determination method.

[0015] This invention provides a construction progress determination method, apparatus, and electronic device based on artificial intelligence. It constructs a construction site intelligent agent set comprising multiple building object intelligent agents and at least one construction organization intelligent agent. Tower crane sensors collect information on the type and installation location of suspended components, and a data routing unit provides this information to the corresponding building object intelligent agent. Upon receiving the information, the building object intelligent agent verifies whether the installation location falls within its own spatial range and whether the current process conforms to the preceding logic. If the verification is valid, it updates its own progress status and synchronizes it to the construction organization intelligent agent. If the verification is invalid, it rejects the data and triggers a re-matching process. The construction organization intelligent agent then transfers the data to other building object intelligent agents for verification or dynamically creates a new intelligent agent until a match is successful. Finally, the construction organization intelligent agent aggregates the progress statuses of all building object intelligent agents to generate the overall construction progress. This achieves distributed rationality verification and automatic error correction of construction activities without relying on a pre-set BIM model.

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a construction progress determination method based on artificial intelligence provided by an embodiment of the present invention is shown; Figure 2 A schematic diagram of a construction progress determination device based on artificial intelligence provided in an embodiment of the present invention is shown. Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention is shown. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0020] Considering that construction sites often rely on manual statistics and experience-based judgments—such as observing visible indicators like the frequency of tower crane hook movements, material transport volume, and the integrity of the scaffolding—and combining this with experience to estimate construction progress, specific methods might include judging the intensity of hoisting operations by the number of tower boom swings or inferring the building's topping-out stage based on the progress of scaffolding dismantling. Furthermore, progress reporting is often delayed, and the lack of real-time, automated monitoring methods can easily lead to problems such as material shortages or delays in procedures.

[0021] Based on this, embodiments of the present invention provide a method and apparatus for determining construction progress based on artificial intelligence, which are described below through embodiments.

[0022] Example 1: like Figure 1 As shown, this embodiment provides an artificial intelligence-based method for determining construction progress, which is applied to an intelligent agent set. The intelligent agent set includes multiple building object intelligent agents and at least one construction organization intelligent agent. Each building object intelligent agent corresponds to a building component or a construction location and is used to verify the rationality of construction actions performed on that component or location. The construction organization intelligent agent is used to coordinate the logical relationships between the various building object intelligent agents.

[0023] Multiple building object agents are the smallest management units corresponding to a single specific building object. Each agent is bound to only one independent management object, which can be a single prefabricated building component or a pre-defined construction area. The sole core task of each building object agent is to check whether the construction behavior currently applied to its corresponding object conforms to the rules. For example, verifying whether the hoisting component matches the corresponding position and whether the current construction conditions of the object are met. All rationality verification of individual objects is completed by this agent.

[0024] At least one construction organization agent: This is the overall coordination and management unit of the entire system. It is not responsible for verifying the rationality of individual objects, but only for coordinating and managing the logical relationships between all construction object agents, including verifying the sequence of construction procedures, adjusting and allocating data after mismatches, and summarizing and outputting the overall project progress. If the project is large, multi-level construction organization agents can be set up, such as dividing the agents into zones according to construction areas, without affecting the overall core functions.

[0025] In traditional centralized processing architectures, the central node needs to handle fine-grained physical verification of the placement and attitude of a massive number of individual components, as well as coordinate the logical dependencies of the overall process. This can easily lead to processing bottlenecks due to excessive computational coupling, resulting in delays in progress response.

[0026] This embodiment decentralizes the high-frequency, local, physical-level task of rationality verification to individual building object agents, allowing each agent to focus solely on the compliance verification of its own bound objects, thereby distributing the computational burden of the central node to the edge. At the same time, low-frequency, global, logical-level tasks such as process dependency coordination and anomaly coordination are centrally managed by the construction organization agent.

[0027] This decoupling mechanism fundamentally eliminates the computational bottleneck and logical coupling when a single central node processes massive amounts of fine-grained verification, enabling high-concurrency real-time processing and accurate tracking.

[0028] It should be understood that the specific form of building components or construction parts is not limited to precast shear walls or precast beams. They can also be cast-in-place concrete columns, steel structure nodes, floor slab areas, or specific construction flow sections that are pre-defined, as long as they can serve as independent management and verification objects. The number of construction organization intelligence agents is not limited to a single individual. When the project is large, multiple levels and multiple construction organization intelligence agents can be set up according to construction areas or professional divisions, as long as they can perform global logical coordination functions.

[0029] The specific execution flow of this method is as follows: Step S110: Collect on-site data through sensors on the tower crane cluster.

[0030] The on-site data includes the type, unique identifier, and installation location information of the currently suspended building components.

[0031] Specifically, the hoisting of prefabricated components at the prefabricated construction site is mainly carried out by tower cranes, so it is necessary to install matching sensor sets on all tower cranes involved in the hoisting operation.

[0032] When a tower crane is hoisting building components, the sensors will automatically collect two types of core on-site data: the first type is the identity and type information of the currently hoisted component, such as distinguishing whether the component is a precast shear wall or a precast beam, and specifying the specific component number; the second type is the installation and placement information of the component, that is, the coordinate information of the position where the component is located after hoisting.

[0033] The specific forms of sensors may include UWB positioning modules, LiDAR, industrial cameras, and weight sensors. Structured field data is formed by fusing multi-source heterogeneous data. The two types of data together constitute the raw field data collected in this step.

[0034] Step S120: Provide the on-site data to a candidate building object agent.

[0035] After completing the on-site data collection, the data routing unit in the system will send the on-site data to the corresponding pre-built building object intelligent agent based on the collected component information. The intelligent agent bound to the component will then conduct subsequent construction rationality verification, thereby achieving targeted data distribution and avoiding invalid calculations.

[0036] Step S130: The building object intelligent agent verifies the rationality of the construction behavior and performs corresponding processing based on the verification results.

[0037] After receiving the on-site data, the candidate building object intelligent agent will check whether the construction behavior is compliant item by item according to the pre-set reasonableness verification rules. Only when all verification items pass can it be judged as reasonable.

[0038] If the verification result is reasonable, the progress status maintained by the candidate building object agent is updated, and the update result is synchronized to the construction organization agent.

[0039] Once the verification is successful, the corresponding building object's intelligent agent will update the progress status of the components or construction parts it maintains, such as changing the status from "pending installation" to "installing" or "completed". After the modification is completed, the update result will be sent to the construction organization intelligent agent, and the overall coordinator will update the overall project progress to complete the entire process of this hoisting operation.

[0040] If the verification result is unreasonable, the data will be rejected and a rematching process will be triggered. The on-site data will then be transferred to other candidate building objects for verification.

[0041] If any verification item fails, the agent will directly reject the data and refuse the verification. At the same time, it will send a mismatch notification to the system, triggering the system's rematch mechanism. The system will then transfer the on-site data to other candidate building agents for re-verification.

[0042] This rejection and transfer mechanism ensures that mismatched data does not incorrectly update the progress status, thereby guaranteeing the accuracy of progress tracking.

[0043] Step S140: Through the construction organization intelligent agent, the progress status of each building object intelligent agent is summarized, and the construction progress of the entire construction site is generated.

[0044] Each building object agent maintains its own real-time progress status. The construction organization agent automatically collects the progress status of all building object agents according to a preset cycle. After classifying and summarizing the status of all objects, the complete actual construction progress of the entire construction site can be obtained. The whole process does not require manual on-site statistics and is fully automated.

[0045] This embodiment constructs a construction site intelligence agent set comprising multiple building object intelligence agents and at least one construction organization intelligence agent. Tower crane sensors collect information such as the type and installation location of suspended components, and this information is provided to the corresponding building object intelligence agent. Upon receiving the information, the building object intelligence agent verifies whether the installation location falls within its own spatial range and whether the current process conforms to the preceding logic. If the verification is valid, it updates its own progress status and synchronizes it to the construction organization intelligence agent. If the verification is invalid, it rejects the data and triggers a re-matching process. The construction organization intelligence agent then transfers the data to other building object intelligence agents for verification or dynamically creates a new intelligence agent until a match is successful. Finally, the construction organization intelligence agent aggregates the progress statuses of all building object intelligence agents to generate the overall construction progress. This achieves distributed rationality verification and automatic error correction of construction activities without relying on a pre-built BIM model.

[0046] Example 2: Based on Example 1, when all existing candidate building object agents verify the on-site data as unreasonable, the method further includes: generating an abnormal alarm and suspending the automatic scheduling of the building component to await manual intervention; or, the construction organization agent creates a new building object agent and initializes the attributes of the new agent to take over the verification and progress management of the building component.

[0047] Specifically, in the rematching process of Example 1, the on-site data is sequentially transferred to other candidate building object agents for verification. However, in actual construction scenarios, especially in prefabricated buildings or projects with frequent design changes, extreme abnormal scenarios inevitably occur: that is, after traversing all existing candidate building object agents of the same type, all verification results are unreasonable.

[0048] This situation typically means that the currently suspended building component is an unknown new item not predefined in the initial BIM model or schedule, or that its placement information has no overlap with the spatial range of any known components. If the system simply remains in a state of continuous looping and rematching at this point, it will cause the progress tracking of this component to deadlock, thereby hindering the calculation and output of the overall construction progress. In order to establish defense depth under extreme anomalies and prevent the system from getting stuck in a progress deadlock due to unknown components, this embodiment splits two parallel handling paths at this node.

[0049] Path A: The system generates an abnormal alarm and suspends the automatic scheduling of the building component, awaiting manual intervention.

[0050] It should be understood that the specific form of anomaly alarms is not limited to text prompts pushed to the handheld terminals of management personnel. It can also be triggered by audible and visual alarms, or highlighted and flashing markers on the digital twin interface, as long as it can attract the attention of on-site management personnel. Suspending automatic scheduling means that the system temporarily suspends the routing, distribution, and verification cycle of data for that component, returning control to human experts.

[0051] When faced with data that completely exceeds the system's cognitive boundaries, forcibly interrupting the automatic processing flow via path A can effectively prevent the system from blindly self-evolving based on erroneous assumptions, thus avoiding more serious progress statistics chaos and ensuring the system's safety and controllability in the face of unknown risks. After management intervention, the true identity and design intent of the components can be confirmed through on-site surveys before deciding on subsequent handling measures.

[0052] Path B: The construction organization agent creates a new building object agent and initializes the properties of the new agent to take over the verification and progress management of the building component.

[0053] When the system detects that the component is a design change or an unforeseen addition, and the site data itself has sufficient structured features, the construction organization agent will call the built-in agent factory module to dynamically instantiate a new building object agent.

[0054] The core mechanism of this strategy lies in enabling the system to respond instantly to unknown environments by dynamically expanding the membership of the intelligent agent set. Once a new intelligent agent is created and takes over the component, the verification process, which was previously deadlocked, will be restored. The progress status of the component can be maintained normally and synchronized globally, thereby fundamentally ensuring the continuity and closed-loop nature of construction progress tracking and preventing gaps in the overall progress summary due to the absence of individual components.

[0055] In actual deployment, the system can flexibly select trigger path A or path B based on the confidence level of abnormal data, the project's requirement for real-time progress, or the preset preferences of management personnel. For example, during the construction phase of the core load-bearing structure with extremely high safety requirements, path A can be used first; while during the conventional hoisting phase, where seamless progress tracking is pursued and design changes are frequent, path B can be used first.

[0056] This defensive strategy design ensures that the system will neither go out of control due to excessive automation nor stagnate due to excessive conservatism, thus continuously guaranteeing the robustness of the progress monitoring system.

[0057] Example 3: In traditional centralized scheduling systems or systems lacking directional routing mechanisms, once a piece of field data is collected, a full traversal approach is typically used. This data is sequentially sent to every building object agent in the agent set for tentative verification until a valid agent is found. However, this full traversal approach is problematic in large-scale prefabricated building projects. Since the number of building object agents can reach hundreds or even thousands, and dozens of hoisting data points can be generated per minute, this approach easily leads to massive amounts of invalid verification calculations. This not only severely wastes the system's computing resources but also causes delays in data distribution and verification response, ultimately hindering the real-time output of the overall construction progress.

[0058] Therefore, based on Example 1, providing on-site data to a candidate building object agent includes: parsing the building component type and spatial coordinates in the installation location information of the on-site data; and inputting the on-site data into a candidate building object agent that matches the type and spatial coordinates of the building component.

[0059] Specifically, to avoid the invalid calculations caused by full traversal, this embodiment introduces a dual data routing mechanism based on type and spatial coordinates. After the tower crane cluster sensors collect on-site data, the data routing unit first performs structured parsing on the data, extracting the building component type and the spatial coordinates from the installation location information.

[0060] After parsing, the data routing unit does not blindly traverse, but performs targeted matching based on the features of these two dimensions.

[0061] In terms of type, the system only narrows the range of data candidates to the set of building object agents that are the same type as the component. For example, if the current component type is a precast beam, the data routing unit will only look for candidates among all agents corresponding to precast beams, and directly exclude all agents corresponding to precast shear walls or cast-in-place columns, thereby significantly reducing the range of invalid verifications in the first step.

[0062] In the spatial dimension, each candidate building object intelligent agent that has passed type screening corresponds to a preset spatial region and region center. The data routing unit can calculate the spatial matching degree between the spatial coordinates of the current component and the region center of the candidate intelligent agent through the Mspace function.

[0063] The Mspace function uses an exponential decay mechanism to ensure that the closer the location coordinates are to the spatial center of the candidate agent, the higher their matching score. Simultaneously, it uses an indicator function to hard-truncate low-scoring matches that fall outside the spatial range. Only when the spatial matching exceeds a preset matching threshold is the candidate building agent finally selected, and only then will the on-site data be directed to that agent.

[0064] This embodiment achieves targeted data distribution through a dual matching mechanism of initial screening and fine screening using spatial matching degree function. It compresses the invalid computation that might have required traversing hundreds of agents into verification of only a few high-matching candidate agents, thereby significantly saving system computing resources, greatly improving the efficiency of data routing and verification, and ensuring high-concurrency real-time response capability for determining construction progress.

[0065] It should be understood that the above spatial matching degree function is only an interpretive algorithm example and not a limitation. In other embodiments, spatial nearest neighbor search algorithm based on KD tree or Boolean decision algorithm based on geofencing can also be used to achieve spatial coordinate orientation matching, as long as it can achieve the function of accurately distributing data to the candidate agent that best fits the spatial range.

[0066] Example 4: Based on Example 1, verifying whether the current construction behavior is reasonable includes: verifying whether the placement location indicated by the installation placement information is within the spatial range corresponding to the candidate building object intelligent agent; and verifying whether the current construction behavior conforms to the preset process logic sequence; when both the spatial range verification and the process logic sequence verification pass, the current construction behavior is determined to be reasonable.

[0067] Spatial verification example: The location coordinate range of the intelligent agent A is (100,100,12m)-(102,102,13m). The location coordinates collected on site are (105,105,12.5m). The coordinates exceed the spatial range of intelligent agent A, so the verification is directly judged to fail.

[0068] Process verification example: The project's preset construction sequence is "first complete the shear wall hoisting, then carry out the floor slab hoisting". When the building object intelligent agent corresponding to the 3rd floor east unit floor slab receives the hoisting data, the shear wall in that area has not yet been completed, and the process verification is directly judged to fail.

[0069] This embodiment employs a dual verification dimension of spatial location and procedural logic. This eliminates both physical placement errors caused by component misallocation and logical violations that deviate from the construction sequence, ensuring absolute compliance of construction activities from both physical and logical perspectives. Only when both spatial scope verification and procedural logic sequence verification pass does the system determine that the current construction activity is reasonable. This dual-defense mechanism eliminates the risk of missed detections that may arise from single-dimensional verification. For example, verifying only the spatial scope might allow for a correctly positioned but incorrectly executed hoisting operation, or verifying only the procedural logic might allow for a correctly executed but incorrectly positioned installation. This ensures absolute reliability of construction progress tracking across both physical and logical dimensions.

[0070] Example 5: Based on Example 2, creating a new building object intelligent agent includes: Step S1201: Load basic attributes from the preset agent template.

[0071] The construction organization intelligent agent maintains a pre-set knowledge base that stores general rule templates for various common building components. For example, when an unknown component is identified as a precast beam, the system will load the basic template for precast beams from the knowledge base. This template includes the general state machine definition, basic verification logic framework, and default physical parameter tolerance range for precast beam components.

[0072] It should be understood that the specific form of the preset template is not limited to a single type of static file. It can also be a probability distribution model generated based on historical project data statistics or a feature prototype extracted based on clustering of similar components, as long as it can provide a general initialization skeleton for the new intelligent agent.

[0073] The core mechanism of this template-based loading mechanism is that by reusing proven and mature rule frameworks, it avoids the huge computational overhead and time cost of building agent logic from scratch, thus laying the foundation for the rapid instantiation of new agents.

[0074] Step S1202: Initialize the verification rules and corresponding spatial range of the current building component based on its characteristics.

[0075] Since the general template only provides coarse-grained default parameters, it cannot accurately adapt to the personalized characteristics of the current specific unknown component. Therefore, the system needs to fine-tune the template parameters in this step.

[0076] For example: Define the feature vector of the current component and the target specification. The system calculates the loss function to measure the difference between the current template configuration and the target specification, and adjusts the verification parameter θ accordingly.

[0077] Through a fine-tuning mechanism, the system can quickly shift general basic parameters to optimal parameters that fit the current component characteristics with minimal computational cost.

[0078] For example, if the current component is a special precast beam whose dimensions have been lengthened due to design changes, the fine-tuning mechanism will automatically broaden the tolerance range for its spatial verification based on its length and weight characteristics, and adjust the weighting coefficients in the process compliance score.

[0079] It should be understood that the specific algorithm for parameter fine-tuning is not limited to meta-learning gradient updates. It can also use posterior probability update algorithms based on Bayesian inference or lookup table methods based on rule mapping, as long as it can make targeted corrections to the basic parameters according to the current features.

[0080] This step enables the instantaneous transformation of the new intelligent agent from general to specific through precise offset driven by local features. This ensures the safety of the initialization logic and endows it with the ability to accurately verify unknown components.

[0081] Step S1203: Register the newly created agent to the agent set and immediately assign it to the re-matching process for further verification.

[0082] After completing attribute loading and parameter fine-tuning, the new agent possesses full verification capabilities. The construction organization agent registers this instance of the new agent in the global agent set, updates the candidate list of routing strategies, and immediately distributes the previously deadlocked field data to the new agent, triggering dual verification of its execution space scope and process logic.

[0083] This seamless registration and assignment mechanism ensures zero latency from the creation of new intelligent agents to their online response, allowing the progress tracking process, which was originally interrupted due to a full mismatch, to be restored instantly.

[0084] This embodiment achieves rapid instantiation and immediate response of new building object intelligent agents through the organic combination of template loading and feature fine-tuning. Template loading solves the time-consuming problem of building from scratch, while fine-tuning solves the accuracy problem of mismatch between general templates and specific components. Together, they constitute the underlying support for the dynamic scalability of the system.

[0085] This design enables the construction progress monitoring system to automatically generate new management units with complete verification capabilities within seconds when faced with design changes or unforeseen additions, without waiting for manual offline modeling and configuration. This ensures the continuity and closed-loop nature of progress tracking and enables the system to adapt to complex and ever-changing construction environments.

[0086] The template loading and fine-tuning mechanisms described above are for illustrative purposes only and are not restrictive. In actual deployment, developers can flexibly choose different initialization strategies and fine-tuning algorithms based on the specific data volume and computing resources of the project.

[0087] Example 6: Specifically, in the verification and rematching closed loop described in Example 1, when the candidate building object agent is deemed unreasonable and rejects the data, although the system ensures the final destination of the data for a single instance through a transfer mechanism, if this frequent rejection behavior is ignored, the system will face a hidden risk of continuous mismatch.

[0088] For example, if a building object agent is initially configured with too small a spatial range or has overly strict verification rules, it may continuously reject reasonable data that should belong to it. This will not only waste computing power by causing data to flow ineffectively among multiple agents, but may also hinder the real-time performance of the overall progress updates.

[0089] To establish a defense depth with parameter endpoints and dynamic adjustments, and to prevent continuous mismatches caused by configuration deviations, this embodiment introduces a mismatch penalty and a route adaptive optimization mechanism.

[0090] Based on the rematching process in Example 1, the rematching process further includes: recording a mismatch event for the candidate building object intelligence agent that originally provided data; when the cumulative mismatch events of the same candidate building object intelligence agent exceed a preset threshold, the construction organization intelligence agent determines that the intelligence agent deviates from the current routing strategy or its own configuration, and performs one of the following operations: reducing the priority of the intelligence agent in the routing strategy so that it no longer prioritizes receiving data of this type of component; or, triggering a configuration diagnosis of the intelligence agent to check whether there is a deviation in the definition of its verification rules or spatial range, and generating a repair suggestion.

[0091] This embodiment, by introducing an intervention path for mismatch event recording, in principle constructs a dual defense depth from traffic avoidance to root cause repair.

[0092] Example 7: In traditional construction management systems, if managers need to adjust the verification logic of a component or modify process dependencies based on unforeseen circumstances on site, they typically need to log into a complex management backend, click through configuration menus layer by layer, and may even need some programming or database operation skills. This intervention method based on traditional configuration interfaces has an extremely high operational threshold, slow response speed, and is difficult to adapt to the rapidly changing environmental demands of construction sites.

[0093] To break down this operational barrier, this embodiment provides the construction organization's intelligent agent with a zero-code intervention mechanism for natural language interaction and security permission verification, enabling managers to directly intervene in the underlying intelligent agent rules and logic using voice or text commands that are most in line with human intuition.

[0094] Based on Example 1, the construction organization intelligent agent supports natural language interaction to receive natural language instructions from on-site management personnel; the method also includes: performing semantic parsing and security permission verification on the received natural language instructions; if the verification is successful, the verification rules or logical relationships of each building object intelligent agent are dynamically adjusted according to the instruction content, and the adjustment results are fed back to the management personnel.

[0095] Specifically, when on-site managers input natural language commands through text boxes on handheld terminals, smart speakers, or digital twin interfaces, the construction organization intelligent agent first performs semantic parsing on the received natural language commands.

[0096] This embodiment introduces the BERT model as an example of the underlying algorithm to explain the principle of semantic parsing: the system inputs natural language instructions into the pre-trained BERT model and extracts the intent and entities from the instructions.

[0097] For example, when a manager enters the instruction "Pause the installation verification of all precast beams in Area A and change to only check the welding quality", the BERT model will parse the Intent as "Modify verification rules", and the Entities will include the target area "Area A", the target component type "precast beam", and the new verification content "welding quality".

[0098] It should be understood that the specific form of semantic parsing models is not limited to the BERT model. They can also adopt prompt word engineering parsing frameworks based on large language models or intent classifiers based on rule matching, as long as they can accurately extract structured operational intents and target entities from unstructured natural language.

[0099] After parsing, the system immediately enters the security permission verification phase. Because natural language commands can directly modify underlying verification rules or global process logic, without strict security measures, the entire progress monitoring system can easily collapse due to accidental operations or malicious commands.

[0100] This embodiment also sets up a permission matrix, which defines the authorization mapping relationship between different user roles and different operation intentions.

[0101] For high-risk intents involving modifications to core logic, such as "modifying verification rules" or "skipping process checks," the permission matrix requires that the user role initiating the instruction must meet the critical permission level.

[0102] The system checks the permission matrix to determine whether the intent of the current command is within the scope of the current user's role authorization.

[0103] Based on the verification results, the system will follow two distinct processing paths. If the verification fails, meaning the current user's role and permissions are insufficient to support the intent of the instruction, the construction organization agent will refuse to execute the instruction and send an "Insufficient permissions" alarm to the management personnel terminal. Simultaneously, the unauthorized access event will be recorded for security auditing.

[0104] If the verification passes, the construction organization agent will generate a corresponding temporary rule set based on the parsed intent and entity. For example, based on the above parsing results, the system will generate a temporary verification rule replacement set for the precast beam building object agent in area A, temporarily replacing its original dual verification logic of space and process with a single verification logic that only checks the welding quality.

[0105] Subsequently, the construction organization agent sends the temporary rule set to the target building object agent. Upon receiving the new rules, the target agent immediately updates its state machine and verification engine, and performs verification according to the new rules the next time it receives on-site data.

[0106] After execution, the construction organization AI will also provide the management personnel with a natural language description of the adjustment results, such as "The verification logic of the precast beams in area A has been successfully adjusted to check only the welding quality", forming a complete interactive closed loop.

[0107] This embodiment, through the organic combination of natural language interaction and permission matrix verification, not only significantly improves the response speed of on-site management personnel to emergencies, but also maintains the security and robustness required of an industrial-grade system.

[0108] The BERT model and permission matrix mechanism described above are for illustrative purposes only and are not restrictive. In actual deployment, developers can flexibly choose semantic parsing algorithms and permission verification strategies of different strengths according to the security level and interaction requirements of the project.

[0109] This invention provides an artificial intelligence-based method for determining construction progress, applied to an intelligent agent set, which includes multiple building object intelligent agents and at least one construction organization intelligent agent; each building object intelligent agent corresponds to a building component or a construction part, used to verify the rationality of construction behaviors acting on the building component or construction part; the construction organization intelligent agent is used to coordinate the logical relationships between the various building object intelligent agents.

[0110] Based on the same inventive concept, an artificial intelligence-based construction progress determination device is provided, applied to an intelligent agent set. The intelligent agent set includes multiple building object intelligent agents and at least one construction organization intelligent agent. Each building object intelligent agent corresponds to a building component or a construction location, used to verify the rationality of construction actions performed on that building component or construction location. The construction organization intelligent agent is used to coordinate the logical relationships between the various building object intelligent agents; such as... Figure 2 As shown, it includes: The data acquisition module 201 is used to collect field data through sensors on the tower crane cluster. The field data includes the type, unique identifier, and installation position information of the currently suspended building component. Module 202 is provided to provide on-site data to a candidate building object intelligent agent; after receiving the on-site data, the candidate building object intelligent agent verifies whether the current construction behavior is reasonable. The update module 203 is used to update the progress status maintained by the candidate building object agent if the verification result is reasonable, and to synchronize the update result to the construction organization agent. The transfer module 204 is used to reject the current data and trigger the rematch process if the verification result is unreasonable, and transfer the on-site data to other candidate building object intelligent agents for verification. The summary module 205 is used to summarize the progress status of each building object intelligent agent through the construction organization intelligent agent and generate the construction progress of the entire construction site.

[0111] The artificial intelligence-based construction progress determination device provided in this embodiment of the invention can be specific hardware on the device or software or firmware installed on the device. The implementation principle and technical effects of the device provided in this embodiment of the invention are the same as those in the foregoing method embodiments. For the sake of brevity, any parts not mentioned in the device embodiments can be referred to the corresponding content in the foregoing method embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.

[0112] Based on the same technical concept, embodiments of the present invention also provide an electronic device, such as... Figure 3 As shown, it includes a processor 301, a communication interface 302, a memory 303, and a communication bus 304, wherein the processor 301, the communication interface 302, and the memory 303 communicate with each other through the communication bus 304.

[0113] Memory 303 is used to store computer programs; The processor 301, when executing the program stored in the memory 303, implements the steps of the construction progress determination method based on artificial intelligence.

[0114] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not indicate that there is only one bus or one type of bus.

[0115] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0116] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0117] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0118] The computer program product for determining construction progress based on artificial intelligence provided in this embodiment of the invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0119] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0120] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0121] In addition, the functional units in the embodiments provided by the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0122] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0123] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0124] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. All should be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for determining construction schedule based on artificial intelligence, characterized in that, It is applied to a set of intelligent agents; the set of intelligent agents includes multiple building object intelligent agents and at least one construction organization intelligent agent; each building object intelligent agent corresponds to a building component or a construction part, and is used to verify the rationality of the construction behavior acting on the building component or construction part; The construction organization agent is used to coordinate the logical relationships between various building object agents; the method includes: Field data is collected by sensors on the tower crane cluster. The field data includes the type, unique identifier, and installation location information of the currently suspended building components. The on-site data is provided to a candidate building object intelligent agent; after receiving the on-site data, the candidate building object intelligent agent verifies whether the current construction behavior is reasonable. If the verification result is reasonable, the progress status maintained by the candidate building object agent is updated, and the update result is synchronized to the construction organization agent. If the verification result is unreasonable, the data is rejected and a rematching process is triggered, and the on-site data is transferred to other candidate building object agents for verification. The construction organization agent summarizes the progress status of each building object agent and generates the construction progress of the entire construction site.

2. The method according to claim 1, characterized in that, When all existing candidate building object agents verify the site data as unreasonable, the method further includes: Generate an anomaly alarm and suspend automatic scheduling of the building component, pending manual intervention; Alternatively, the construction organization agent can create a new building object agent, initialize the attributes of the new building object agent to take over the verification and progress management of the building component.

3. The method according to claim 1, characterized in that, Providing the on-site data to a candidate building object agent includes: Analyze the spatial coordinates of the building component types and installation locations in the field data; Based on the type of the building component and its corresponding spatial coordinates, the on-site data is input into the candidate building object intelligent agent that matches the type and spatial coordinates.

4. The method according to claim 1, characterized in that, Verifying whether the construction activities are reasonable includes: Verify whether the installation location indicated by the installation location information is within the spatial range corresponding to the candidate building object intelligent agent; And, verify whether the current construction behavior conforms to the preset logical sequence of procedures; When both the spatial scope verification and the process logic sequence verification pass, the current construction behavior is deemed reasonable.

5. The method according to claim 4, characterized in that, The verification of whether the current construction behavior conforms to the preset sequence of procedures includes: The candidate building object intelligent agent sends a preliminary process query request to the construction organization intelligent agent; The construction organization intelligence agent returns the completion status of the previous process of the current component; The candidate building object intelligent agent determines whether the current construction behavior meets the preset process logic sequence based on the returned status.

6. The method according to claim 2, characterized in that, The process of creating a new building object intelligent agent includes: Load basic attributes from a pre-set agent template; Initialize the verification rules and corresponding spatial range based on the characteristics of the current building components; The newly created building object agent is registered to the agent set and immediately assigned to the rematching process for further verification.

7. The method according to claim 2, characterized in that, The rematching process also includes: Record a mismatch event for the candidate building object agent whose original data was provided. When the cumulative number of mismatch events for the same candidate building object exceeds a preset threshold, the construction organization agent determines that there is a deviation between the agent and the current routing strategy or its own configuration, and performs one of the following operations: Reduce the priority of this agent in the routing strategy so that it no longer prioritizes receiving data from this type of component; Alternatively, it can trigger configuration diagnostics for the agent, check for deviations in the definition of its verification rules or spatial range, and generate repair suggestions.

8. The method according to claim 1, characterized in that, The construction organization intelligent agent supports natural language interaction and is used to receive natural language instructions from on-site management personnel; the method further includes: Perform semantic parsing and security permission verification on the received natural language instructions; If the verification passes, the verification rules or logical relationships of each building object's intelligent agent will be dynamically adjusted according to the instructions, and the adjustment results will be reported back to the management personnel.

9. A construction progress determination device based on artificial intelligence, characterized in that, It is applied to a set of intelligent agents; the set of intelligent agents includes multiple building object intelligent agents and at least one construction organization intelligent agent; each building object intelligent agent corresponds to a building component or a construction part, and is used to verify the rationality of the construction behavior acting on the building component or construction part; The construction organization intelligent agent is used to coordinate the logical relationships between various building object intelligent agents; the device includes: The data acquisition module is used to collect field data through sensors on the tower crane cluster. The field data includes the type, unique identifier, and installation position information of the currently suspended building component. A module is provided to provide the on-site data to a candidate building object intelligent agent; after receiving the on-site data, the candidate building object intelligent agent verifies whether the current construction behavior is reasonable. The update module is used to update the progress status maintained by the candidate building object agent if the verification result is reasonable, and to synchronize the update result to the construction organization agent. The transfer module is used to reject the current data and trigger a rematch process if the verification result is unreasonable, and transfer the on-site data to other candidate building object intelligent agents for verification. The summary module is used to summarize the progress status of each building object intelligent agent through the construction organization intelligent agent, and generate the construction progress of the entire construction site.

10. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the artificial intelligence-based construction progress determination method according to any one of claims 1-7.

Images