Dynamic identification and correction methods for construction progress deviations in smart construction sites

Abstract

This invention relates to a method for dynamic identification and correction of construction progress deviations in smart construction sites. The method divides the target processes of the construction plan into process nodes, and pre-sets a key evidence set and a credibility level for each process node. Multi-source evidence corresponding to the key evidence set is periodically collected at the construction site and cross-verified. Based on the consistency of the evidence, it is determined whether the evidence is valid, insufficient, or conflicting. The progress status of the process node is updated within a pre-set review window according to the judgment result. Deviations are identified by combining the start time window, the completion time window, and the progress status. Corrections are triggered and executed according to a pre-set correction level sequence. After correction, evidence is re-collected and deviation identification is repeated. This invention, through multi-source evidence consistency judgment and a hierarchical correction mechanism, achieves accurate identification, reliable confirmation, and adaptive correction of construction progress, significantly improving the accuracy, stability, and efficiency of progress determination and project management.

Description

Technical Field

[0001] This invention relates to the field of construction management and information control technology, specifically a method for dynamic identification and correction of construction progress deviations in smart construction sites. Background Technology

[0002] While existing methods for dynamic identification and correction of construction progress deviations in smart construction sites incorporate BIM models, computer vision, and intelligent algorithms to some extent, they still exhibit significant shortcomings and drawbacks in practical engineering applications. Taking the BIM-based and computer vision-based dynamic optimization method and system for construction progress disclosed in patent document CN120747233A as an example, its core technical path mainly relies on the automatic registration of construction site images and BIM models, the mapping of visual recognition results to components, and progress deviation monitoring based on the critical path and deviation propagation matrix. Furthermore, it uses Bayesian networks and Monte Carlo simulations for risk prediction. This scheme emphasizes the combination of model-driven and algorithmic prediction in its overall architecture; however, it still exhibits significant limitations and coarse-grained characteristics in terms of evidence for progress deviation identification and correction. First, this type of method is highly dependent on the accuracy of computer vision recognition results and their mapping with BIM components. If there are obstructions, changes in lighting, high similarity of components, or fluctuations in image acquisition quality at the construction site, the visual recognition error will be directly amplified and transmitted to the progress judgment and deviation analysis stages. However, the existing solutions lack fine-grained modeling of the reliability differences between evidence from different sources, and do not systematically distinguish and process abnormal or supplementary data, resulting in insufficient stability of progress judgment results under complex working conditions.

[0003] Secondly, this approach focuses on overall monitoring of schedule deviations using the critical path method and deviation propagation matrix. Its deviation assessment is primarily based on time delay analysis at the process or path level, without dynamically confirming the consistency of multi-source evidence for individual process nodes within the review window. This means that once the schedule status is determined by the model, the system lacks a clear mechanism for locking and unlocking the schedule status if new, highly credible, or conflicting evidence emerges, making it difficult to effectively avoid frequent fluctuations in schedule conclusions or long-term fixation of erroneous judgments. Thirdly, while this type of method introduces risk prediction and resource optimization, its correction strategy mainly focuses on macro-level adjustments to construction sequence and resource allocation. It lacks clear verification objectives and cutoff conditions for the correction process itself. Correction actions often need to be continuously executed until the end of the review cycle or manual intervention, easily leading to over-correction or under-correction, and lacking a mechanism to terminate corrections early based on the sufficiency and consistency of evidence.

[0004] Furthermore, this scheme does not propose a coverage calculation method based on evidence category weights. The roles of various types of evidence in progress confirmation are assumed to be equal or implicitly included in the model parameters. It fails to explicitly distinguish the differences in credibility, importance, and timeliness among video evidence, equipment records, personnel records, and third-party confirmations. This can easily lead to low-credibility or outdated evidence disproportionately impacting progress conclusions in real-world smart construction sites. Finally, this type of scheme focuses more on algorithm prediction and model optimization, lacking targeted design for practical problems commonly encountered in construction management, such as evidence gaps, evidence conflicts, and the need for supplementary evidence recording. When data collection is incomplete or sources are inconsistent, the system often has to compensate by reducing model accuracy or relying on manual verification, making it difficult to form a closed-loop dynamic identification and correction mechanism. Summary of the Invention

[0005] The purpose of this invention is to provide a method for dynamic identification and correction of construction progress deviations in smart construction sites, thereby addressing some of the shortcomings and deficiencies pointed out in the background art.

[0006] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows: a method for dynamic identification and correction of construction progress deviations in smart construction sites, including: dividing the target process of the construction plan into process nodes, pre-setting key evidence sets and evidence source credibility levels for each process node; periodically collecting multi-source evidence corresponding to the key evidence sets at the construction site and performing mutual verification, and judging the evidence as valid, insufficient, or conflicting based on the consistency of the evidence.

[0007] Based on the judgment result, update the progress status of the process node within the preset review window; identify the start deviation or completion deviation by combining the start time window, the completion time window and the progress status; trigger and execute the correction according to the preset correction level order; recollect evidence after correction and repeat the deviation identification.

[0008] Furthermore, for each piece of evidence collected, an evidence digest code containing the source identifier, collection time, collection location, and object identifier is generated and recorded in chronological order. During mutual verification, the sequential continuity and repetition of the digest codes are checked. If there is a break in the sequence, repetition, or the collection time is earlier than the first recording time, the corresponding evidence is marked as supplementary evidence and the credibility level is downgraded. In the consistency determination, supplementary evidence has a lower priority than non-supplementary evidence.

[0009] Furthermore, when the consistency determination is a conflict of evidence, or when the evidence is continuously determined to be insufficient within the preset review window, a verification instruction is generated and executed. The verification instruction limits the verification period, location, and object, and requires the formation of verification evidence from at least two different sources. If the verification evidence still does not meet the consistency conditions, the status of the corresponding process node is updated to "revoked," and enhanced confirmation rules are enabled. The enhanced confirmation rules include raising the credibility level threshold required for the evidence to be valid or increasing the number of required evidence categories.

[0010] Furthermore, when the correction is triggered, a correction instruction carrying a correction verification target is generated. The correction verification target is the evidence category and consistency conditions that need to be supplemented in the review window. After correction, the correction is deemed effective only if the re-collected evidence meets the correction verification target. Otherwise, when the deviation occurs continuously to the preset number of times or the process node status repeatedly migrates between disputed and withdrawn to the preset number of times, the correction is upgraded step by step according to the preset correction level order and the correction is executed again.

[0011] Furthermore, the corrective verification target includes a corrective verification cutoff condition, which is to achieve a preset evidence coverage rate or a preset number of evidence consistency confirmations within the review window; when the cutoff condition is met in advance within the review window, the subsequent evidence collection or the execution of the corrective action is terminated, and the progress status of the corresponding process node is locked as confirmed until the next process node is entered.

[0012] Furthermore, the step-by-step escalation according to the preset correction level includes setting an escalation suppression condition. The escalation suppression condition is that the evidence obtained through re-collection indicates that the deviation is caused by an evidence gap and there is no evidence conflict. When the escalation suppression condition is met, only evidence completion correction is performed and the review window is extended.

[0013] Furthermore, the evidence coverage rate is calculated by evidence category and is defined as the ratio of the number of evidence categories that have been obtained and passed the consistency verification within the review window to the number of evidence categories required by the correction verification target; when the evidence coverage rate reaches a preset threshold and there is no evidence conflict, it is determined that the correction verification cutoff condition is met.

[0014] Furthermore, the number of evidence consistency confirmations is the number of times that the same process node continuously obtains confirmations that meet the consistency conditions within the review window, and the evidence sources corresponding to two adjacent confirmations are different at least once; when the number of evidence consistency confirmations reaches the preset number, it is determined that the correction verification cutoff condition is met and the termination of subsequent evidence collection or termination of correction action is triggered.

[0015] Furthermore, the progress status lock is confirmed to include lock holding conditions and unlocking conditions; the lock holding condition is that no conflicting evidence with a credibility level higher than the credibility level of the evidence on which the confirmed status is based appears during the lock period; the unlocking condition is that when evidence that meets the conflict determination conditions and has a credibility level higher than the evidence on which it is based appears during the lock period, the lock is released and the progress status of the process node is updated to disputed.

[0016] Furthermore, after unlocking and updating the progress status of the process node to "in dispute," a verification instruction is generated in the preset review window to limit the verification period, location, and object, and at least two types of verification evidence from different sources are obtained; when the verification evidence meets the key evidence set and consistency conditions, the progress status is restored to "confirmed," otherwise it is updated to "revoked" and the correction level is upgraded.

[0017] The beneficial effects of this invention are as follows: The method for dynamic identification and correction of construction progress deviations in smart construction sites, by refining the construction plan into process nodes and pre-setting key evidence sets and credibility levels for each process node, combined with the periodic collection and cross-verification of multi-source evidence at the construction site, achieves continuous, objective, and verifiable dynamic perception of the construction progress status. By judging the consistency of evidence within a pre-set review window and distinguishing between valid, insufficient, and conflicting evidence, it effectively avoids misjudgments of progress caused by a single data source or subjective reporting, improving the accuracy and reliability of construction progress identification results. Simultaneously, the introduction of evidence digest codes and their sequential continuity verification mechanism allows for the identification and credibility reduction of supplementary evidence, further enhancing the anti-tampering capability and data credibility foundation of the progress judgment process.

[0018] Furthermore, this invention constructs a closed-loop correction mechanism that is scalable, suppressable, and terminateable by setting correction levels, correction verification targets, and correction verification cutoff conditions. This enables the handling of construction schedule deviations to possess adaptive and refined control capabilities. When deviations are caused by evidence gaps, priority is given to supplementing the evidence and extending the review window to avoid over-correction. When evidence conflicts or repeated disputes occur, the judgment criteria are raised by strengthening confirmation rules and verification instructions, thereby effectively reducing the probability of erroneous and repeated corrections. At the same time, through a progress status locking and unlocking mechanism, it is ensured that confirmed progress remains stable in the absence of highly credible conflicting evidence, reducing the management costs caused by frequent status switching. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the logic for identifying and correcting construction progress deviations in this invention.

[0020] Figure 2 This is a functional relationship diagram of the multi-source evidence governance of construction progress in this invention.

[0021] Figure 3 This is a timing diagram of evidence judgment and progress confirmation within the correction and review window of this invention.

[0022] Figure 4 This is a schematic diagram of the process node progress determination based on the sequential continuity of evidence digest codes and the identification of supplementary evidence in Embodiment 1 of the present invention.

[0023] Figure 5 This is a schematic diagram of the verification instruction generation and process node status cancellation process in the case of evidence conflict and insufficient evidence in Embodiment 1 of the present invention.

[0024] Figure 6 This is a schematic diagram of the correction triggering, correction verification target verification, and correction level upgrade and suppression process in Embodiment 1 of the present invention.

[0025] Figure 7 This is a schematic diagram illustrating the parallel determination of evidence coverage and the number of evidence consistency confirmations under the corrective verification cutoff condition in Embodiment 2 of the present invention.

[0026] Figure 8 This is a schematic diagram of the weighted evidence coverage calculation and correction verification cutoff determination based on evidence category weights in Embodiment 2 of the present invention.

[0027] Figure 9 This is a schematic diagram of the state evolution of the progress status locking, unlocking, verification, and status recovery or cancellation process in Embodiment 2 of the present invention. Detailed Implementation

[0028] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] Combined with appendix Figure 1 This invention provides a dynamic identification and correction method for construction progress deviations in smart construction sites. It structurally breaks down the target processes in the construction plan, dividing each target process into independent and manageable process nodes. Each process node corresponds to a specific construction goal or stage result in the construction plan. For each process node, a set of key evidence related to its completion status is pre-configured. This set of key evidence characterizes whether the process node has started or been completed as planned. Key evidence may include on-site image data, video data, equipment operation data, personnel operation records, material entry and exit records, or sensor-collected data. Simultaneously, based on the acquisition method, collection stability, and historical reliability of the evidence sources, corresponding credibility levels are preset for different evidence sources to reflect their credibility in progress determination.

[0030] During construction, multi-source evidence corresponding to the key evidence set is automatically or semi-automatically collected at the construction site according to a preset collection cycle or event triggering rules. The collected multi-source evidence is then associated and stored with the corresponding work process nodes. Subsequently, a mutual verification operation is performed on the multi-source evidence corresponding to the same work process node. By comprehensively comparing the consistency of evidence content, temporal correlation, and spatial correlation, it is determined whether different pieces of evidence can corroborate each other. During the mutual verification process, when evidence from different sources and whose credibility levels meet preset conditions are consistent in key construction characteristics, the set of evidence is determined to be valid. When the number of collected evidences is insufficient or the evidence types do not cover the preset key evidence set, and an effective mutual verification relationship cannot be formed, the state is determined to be insufficient evidence. When there are irreconcilable contradictions between different pieces of evidence in key construction characteristics, time, or location, and the credibility level of the contradictory evidence meets the conflict determination requirements, the state is determined to be evidence conflict, thus providing a clear data foundation for subsequent updates and corrections to the construction progress status.

[0031] After determining the consistency of evidence corresponding to each work process node, the determination result is used as input to update the progress status of the work process node within a pre-defined review window. The review window defines the time frame for progress status assessment. Within this time frame, the system continuously receives and analyzes changes in evidence related to the work process node, thereby avoiding misjudgments of progress status due to single-time data collection anomalies or short-term construction fluctuations. Based on the determination results of whether the evidence is valid, insufficient, or conflicting, the progress status of the work process node is updated to confirmed, disputed, or pending confirmation, respectively, allowing the progress status of the work process node to be dynamically adjusted as the evidence changes.

[0032] Building upon the progress status updates, the system further compares and analyzes the progress status of each work node with its corresponding start and finish time windows. If a work node remains in a state of not having started or with insufficient evidence within the start time window, it is identified as a start-up deviation; if a work node fails to reach the confirmed status beyond the finish time window, it is identified as a finish deviation. By introducing a joint judgment mechanism of time windows and progress status, potential progress risks can be identified early in construction, while avoiding rigid judgments caused by relying solely on planned time nodes.

[0033] When a start-up or completion deviation is detected, the system automatically triggers and executes corresponding corrective measures according to a pre-set corrective level sequence. These measures can include different levels of correction methods such as evidence supplementation, on-site verification, manual confirmation, or adjustments to construction resources. After the correction is completed, the system restarts the evidence collection process for the work process nodes, performs mutual verification of the new evidence generated after the correction, and repeats the progress status update and deviation identification process within the review window. This forms a dynamic identification and correction mechanism driven by a closed-loop evidence system, enabling construction progress management to continuously self-correct and gradually stabilize.

[0034] Combined with appendix Figure 2 For each piece of evidence collected at the construction site, a unique evidence digest code is generated when it is created or added to the system. This evidence digest code includes at least the evidence source identifier, the time of evidence collection, the location of evidence collection, and the identifier of the corresponding work process object. The evidence digest code is then stored in association with the corresponding evidence. The evidence digest codes are recorded continuously in the order in which the evidence first enters the system, reflecting the chronological relationship between evidence collection and storage, thus forming a traceable evidence time sequence. This provides fundamental data support for subsequent evidence verification and progress determination.

[0035] During the cross-verification of multi-source evidence, the system verifies the temporal continuity of the evidence and the duplication of digest codes based on the evidence digest codes. When a break in the sequence of evidence digest codes, duplication of digest codes, or the evidence's collection time being earlier than its first record time in the system are detected, it is determined that the evidence was not formed during the normal collection process but was entered through a post-collection supplement. For evidence determined to be supplementary evidence, the system marks it as supplementary and correspondingly lowers its credibility level, reducing its weight in subsequent evidence consistency judgments. Simultaneously, when performing consistency judgments on the same work process node, the system prioritizes non-supplementary evidence for cross-verification analysis. Supplementary evidence is only included in the judgment scope when the number of non-supplementary evidence is insufficient or cannot form a valid cross-verification relationship, thereby effectively reducing the interference of supplementary evidence on the construction progress judgment results.

[0036] When performing multi-source evidence verification for a specific work process node, if the consistency judgment result is evidence conflict, it indicates that there are contradictory relationships between different pieces of evidence regarding key construction characteristics, time information, or spatial information that cannot be mutually corroborated. Alternatively, if the evidence remains insufficient within the preset review window and a valid consistency judgment result cannot be formed, the system determines that the progress status of the current work process node has a high degree of uncertainty. In this case, the system automatically generates and issues a corresponding verification instruction to further confirm the actual construction status of the work process node.

[0037] Verification instructions constrain the scope and method of verification. They explicitly define the time period, construction location, and corresponding work process to be verified, thus preventing irrelevant data from interfering with the verification results. Furthermore, verification instructions require the generation of at least two types of verification evidence from different sources under specified conditions to ensure that the verification results have a basis for multi-source cross-validation. After collection, the verification evidence participates again in the consistency assessment process to correct or confirm any conflicts or insufficiencies in the original evidence.

[0038] If the verification evidence still fails to meet the preset consistency conditions, indicating that the effective progress status of the process node cannot be confirmed even under enhanced verification conditions, the system will update the progress status of the corresponding process node to "cancelled," signifying that the current progress result of the process node lacks a basis for confirmation. Simultaneously, the system will activate enhanced confirmation rules to address potential subsequent recurring disputes. These enhanced confirmation rules raise the required credibility threshold for evidence establishment or increase the number of evidence categories involved in the judgment, thus requiring subsequent progress confirmations to meet stricter evidentiary conditions.

[0039] When the system triggers a corrective action based on the progress status and time window determination results of a process node, it automatically generates a corresponding corrective instruction, which carries a clear corrective verification objective. The corrective verification objective serves as the criterion for determining the effectiveness of the corrective action. Specifically, it includes the types of evidence that need to be supplemented within the review window and the consistency conditions that the evidence must meet. This ensures that the corrective action has verifiable and quantifiable criteria, avoiding situations where the effectiveness of corrective measures cannot be objectively assessed after implementation.

[0040] After corrective measures are implemented, the system restarts the evidence collection process for the work node, performs cross-verification of the re-collected evidence generated after the correction, and compares the verification results with the corrective verification target. Only when the re-collected evidence simultaneously meets the requirements for the supplementary evidence category and the corresponding consistency conditions within the review window is the system deemed the current corrective operation effective, and the system maintains or updates the progress status of the work node accordingly. If the re-collected evidence fails to meet the corrective verification target, the system will continuously monitor the number of deviations and the migration of the work node's progress status between the disputed and withdrawn states. When deviations occur consecutively to a preset number, or the work node's status repeatedly migrates between the two states to a preset number, the system will upgrade the corrective measures step by step according to the pre-set corrective level order, and generate and execute new corrective instructions to gradually increase the intensity of the correction and the confirmation standards.

[0041] To address the process of corrective actions escalating sequentially according to preset correction levels, the system further sets escalation suppression conditions to distinguish the root causes of deviations and avoid unnecessary escalation of correction levels. These escalation suppression conditions determine whether the current deviation is primarily caused by insufficient evidence collection or inadequate evidence coverage, rather than by substantial conflicts between different pieces of evidence. When the system analyzes re-collected evidence after corrective action, if the re-collected evidence shows no contradictions in content regarding the relevant evidence at each process node, and only indicates insufficient evidence quantity or incomplete coverage of the preset key evidence set, then the escalation suppression conditions are satisfied.

[0042] When the escalation suppression conditions are met, the system does not perform a step-by-step escalation of the correction level. Instead, it only takes corrective measures to fill evidence gaps, such as extending the evidence collection time, increasing the collection frequency, or introducing additional evidence sources. Simultaneously, the system extends the review window corresponding to that process node to provide sufficient time for evidence completion. Within the extended review window, the system continuously performs cross-verification and progress status updates on the completed evidence. This improves the completeness and accuracy of progress determination without increasing the intensity of correction or management intervention costs, avoiding over-correction or misjudgment due to temporary evidence insufficiency.

[0043] Combined with appendix Figure 3 The corrective verification objectives further include corresponding corrective verification cutoff conditions to clarify when the corrective action can end and when the progress status can be confirmed. The corrective verification cutoff conditions are set as achieving a preset evidence coverage rate or a preset number of evidence consistency confirmations within the review window, thus constraining the corrective results from two dimensions: evidence completeness and evidence stability. Specifically, the evidence coverage rate characterizes the degree to which key evidence categories are supplemented after correction, and the number of evidence consistency confirmations characterizes the consistency and stability of multi-source evidence in continuous judgments, thereby avoiding premature termination of the correction due to a single accidental consistency.

[0044] During the corrective action execution process, the system continuously monitors the evidence coverage and consistency confirmation within the review window. If the corrective action verification deadline is met before the review window closes, the system immediately terminates subsequent evidence collection operations or terminates any incomplete corrective actions to reduce unnecessary data collection and management intervention. After terminating the corrective action, the system locks the progress status of the corresponding process node to the confirmed state and maintains it in this stable state until the process node is completed and the next process node begins.

[0045] To quantitatively assess the extent to which key evidence is supplemented during the correction process, the system calculates evidence coverage according to evidence category and defines the evidence coverage rate as the weighted evidence coverage rate (CR). The weighted evidence coverage rate comprehensively reflects the contribution of different evidence categories in terms of credibility, importance, and timeliness. Its calculation formula is as follows:

[0046]

[0047] in, A non-empty set of evidence categories required for the correction verification objective, used to represent all evidence categories that need to be covered to complete the correction determination; The set of evidence categories that have been obtained and passed consistency verification within the review window, and that satisfy... By using the aforementioned set relationships, the calculation of evidence coverage is always limited to the scope of the correction and verification objectives, thus avoiding irrelevant evidence from interfering with the results.

[0048] In the calculation of weighted evidence coverage, each evidence category Each corresponds to a weight The weight is used to reflect the comprehensive value of the evidence category in corrective verification, and its calculation formula is as follows:

[0049]

[0050] in, Indicates the category of evidence The numerical results of the credibility level of the corresponding evidence source are used to reflect the reliability of this type of evidence in the course of historical use; Indicates the category of evidence The preset importance coefficient is used to reflect the criticality of this evidence category in the determination of the progress of the process node; This indicates the current judgment time and evidence category within the review window. The time difference between the most recent valid data collection moments that passed the consistency check is used to characterize the freshness of the evidence. When no evidence category exists within the review window... When the valid collection time passes the consistency check, classify the evidence category. The time-sensitive items are calculated based on the initial value, let This ensures that the weight calculation remains deterministic even in boundary cases.

[0051] In the above weight calculation formula, , , These are preset non-negative weighting coefficients, with at least one being positive, used to adjust the weighting of the credibility level of the evidence source, the importance of the evidence category, and the decay of evidence validity. The proportion of influence in; The preset aging decay coefficient, exponential term This is used to characterize the effectiveness of evidence as it gradually diminishes over time. By introducing the aforementioned weighting mechanism, different types of evidence are no longer simply treated equally in the evidence coverage calculation, but are instead measured differently based on their credibility, importance, and timeliness. When the calculated weighted evidence coverage rate (CR) reaches a preset threshold, and there are no evidence conflicts in the evidence judgment process at the corresponding process node, the system determines that the correction verification cutoff condition has been met, and accordingly concludes that the correction objective has been achieved.

[0052] To constrain the stability of process node progress determination, the system introduces the number of evidence consistency confirmations as a key metric for corrective verification. The number of evidence consistency confirmations characterizes the continuity and reliability of the consensus reached by multiple sources of evidence regarding the progress status of the same process node within the review window. It is defined as the number of consecutive confirmations meeting the consistency conditions within the review window corresponding to the same process node. In practice, the system sequentially records the consistency determination results of each mutual verification output. When a determination meets the preset consistency conditions, it is recorded as a valid confirmation. To avoid false stability caused by repeated reporting from a single evidence source, the system further requires that the evidence sources corresponding to two adjacent valid confirmations differ at least once, ensuring that consecutive confirmations are cross-supported by evidence from different sources, thereby enhancing the objectivity and credibility of the consistency confirmation results.

[0053] When the number of evidence consistency confirmations accumulated within the review window reaches a preset threshold, the system determines that the process node has formed a stable and reliable progress assessment result in the current correction phase, and accordingly confirms that the correction verification cutoff condition is met. After the cutoff condition is met, the system automatically triggers the termination of subsequent evidence collection operations or terminates the execution of ongoing correction actions, allowing the progress status to enter the stable confirmation phase, thereby reducing unnecessary duplicate collection and repeated corrections.

[0054] When the progress status of a process node is locked to a confirmed state after meeting the correction verification deadline, the system configures corresponding lock holding and unlocking conditions for this locked state to retain necessary error correction capabilities while ensuring the stability of the progress results. The lock holding conditions constrain the validity of the confirmed state during the locking period. If no conflicting evidence with a higher credibility level than the evidence on which the confirmed state was based appears during the locking period, the system maintains the progress status of that process node unchanged, thereby avoiding frequent interruptions to the confirmed progress status due to low-credibility or sporadic evidence.

[0055] During the lockout period, if new evidence emerges that meets the conflict determination criteria, and the credibility level of this evidence is higher than that of the evidence upon which the confirmed status was based, the system determines that this evidence has the priority to overturn the original confirmation result. In this case, the system automatically unlocks the progress status of the process node and updates the progress status of that process node to "Under Dispute," indicating that the current progress result needs to be re-verified. By introducing an unlocking mechanism based on credibility level comparison, it can be ensured that only conflicting evidence with higher credibility can trigger a status rollback, thus achieving a balance between stability and error correction.

[0056] After the progress status is unlocked and updated to "Under Dispute," the system automatically generates a corresponding verification instruction within a pre-defined review window. This instruction is used to conduct targeted verification of the actual construction status of the work process node. The verification instruction clearly defines the verification time period, construction location, and work process object, and requires the acquisition of verification evidence from at least two different sources to ensure that the verification results have a basis for multi-source cross-verification. After the verification evidence is collected, the system performs a consistency judgment on the verification evidence. When the verification evidence meets the key evidence set requirements for the work process node and satisfies the consistency conditions, the system restores the progress status of the work process node to the "Confirmed" status. When the verification evidence still fails to meet the consistency conditions, the progress status of the work process node is updated to the "Revoked" status, and an upgrade of the correction level is triggered to re-intervene in progress management through stronger correction and confirmation rules.

[0057] Example 1:

[0058] In this embodiment, during the main structure construction phase of a large residential project, the system manages the concrete pouring process of a specific floor as a process node and pre-sets a set of key evidence for this process node, including on-site video and image evidence, concrete pouring equipment operation records, construction personnel work records, and concrete material arrival records. During construction, the system continuously collects multi-source evidence related to this process node from different sources, and automatically generates a corresponding evidence digest code for each piece of evidence when it enters the system. The evidence digest code includes at least the evidence source identifier, collection sequence information, construction area location identifier, and corresponding process object identifier, and is numbered and recorded according to the order in which the evidence first enters the system, thus forming a sequence as follows: Figure 4 The continuous evidence sequence shown by the main broken line.

[0059] Within an evidence collection cycle, the system sequentially receives evidence digest codes numbered 1, 2, 3, and 4. Evidence digest code 1 originates from on-site video equipment, evidence digest code 2 from concrete pump truck equipment, evidence digest code 3 from the construction worker's work record system, and evidence digest code 4 from the materials management system. For example... Figure 4As shown in the continuous points on the left, the above evidence is continuous and consistent in terms of digest code order, collection order and process object identification. During the mutual verification process, the system confirmed that there was no duplicate digest code or broken order, and determined that the above evidence was all normally collected evidence, and participated in the subsequent consistency judgment with a confidence level of 3.

[0060] Subsequently, during the subsequent verification process, the system received another piece of work record evidence uploaded by the construction worker. Its evidence digest code number was 2, which was the same as the previously recorded evidence digest code number, and the collection order of this evidence was significantly earlier than the initial recording order of this source in the system. For example... Figure 4 As indicated by the arrow pointing to the abnormal node, when the system was verifying the sequence continuity and repeatability of the evidence digest codes, it detected that the evidence had duplicate digest codes and an abnormal collection order, thus determining that the evidence did not meet the requirements of the normal collection process. Based on the above verification results, the system automatically marked the evidence as supplementary evidence and downgraded its credibility level from the original level 3 to level 1. This credibility level change process occurred during... Figure 4 This is reflected in the confidence level ladder trajectory on the right.

[0061] When performing evidence consistency assessment on this work process node, the system prioritizes evidence not marked as supplementary for cross-verification analysis according to preset rules. At this point, the three pieces of non-supplementary evidence collected in the first round from on-site video equipment, concrete pump truck equipment, and the material management system maintain a high degree of consistency in construction content, location, and process status. The average completion rate calculated is approximately 0.80, corresponding to about 80% completion of concrete pouring. Although the supplementary evidence shows the process is completed, because it is marked as supplementary and has a lower credibility level, the system only uses it as a supplementary reference and not as the primary basis for judgment. Ultimately, based on the consistency results formed by the non-supplementary evidence, the system determines the progress status of this work process node as in progress rather than completed, thus avoiding misjudgments of progress caused by premature reporting of supplementary evidence.

[0062] During subsequent correction and review processes, the system continues to collect new on-site video evidence and equipment operation evidence, and generates new evidence digest codes 5 and 6 for them, such as... Figure 4 As shown in the continuous data points on the right, the newly added evidence digest codes maintain a continuous order without any anomalies. The new non-supplementary evidence forms a stable consistency with the aforementioned non-supplementary evidence, further confirming that the process node has not yet been completed. Figure 4 As can be seen from the changes in the evidence digest code sequence and the credibility level trajectory shown, the present invention can effectively identify supplementary evidence and reduce its interference with the progress determination results, thereby improving the reliability of progress identification.

[0063] The system is based on Figure 4The consistent results of the supplementary evidence from China and Africa determined the progress status of the concrete pouring process node on that floor to be in progress, and it continued into the review stage. In the subsequent review window, the system again collected multi-source evidence to confirm the process status, at which point new evidence was successively acquired from on-site video equipment, construction workers' mobile terminals, and the material management system. The new video evidence showed that there were still unfinished areas in the pouring area, the equipment operation record showed that the concrete pump truck had accumulated a volume of 120 cubic meters, while the work record uploaded by the construction workers showed that this process was completed and had entered the curing stage. The above-mentioned evidence from different sources showed significant inconsistencies in the description of the construction completion status. Figure 5 The radar chart shows numerical differences between different dimensions, and the system marks the consistency judgment results as evidence conflicts based on this.

[0064] Meanwhile, in another round of review, the system only collected on-site video evidence and some equipment operation data in two consecutive review cycles, failing to obtain construction personnel work records and material consumption records, resulting in insufficient coverage of key evidence categories. For example... Figure 5 As shown in the evidence coverage dimension, the evidence coverage rate is only 0.50, which is lower than the preset threshold. The system continuously determines the evidence to be insufficient within the review window. For both the evidence conflict and insufficient evidence scenarios, the system automatically generates corresponding verification instructions and executes them immediately.

[0065] The verification instruction clearly defines the scope of verification, focusing solely on the concrete pouring process at this floor and limiting the verification location to the construction area of ​​that floor. The verification objects are the completion status of the pouring and the actual amount of concrete used. Simultaneously, the instruction requires the generation of at least two different sources of verification evidence to eliminate uncertainties in the original judgment. According to the verification instruction, the system collected new high-resolution on-site image evidence and actual material supply data exported from the concrete supply system during the verification process. The verification results show that there are still clearly unpoured areas in the on-site images, while the cumulative usage in the material supply data is 135 cubic meters, lower than the planned usage of 150 cubic meters. Both are consistent in terms of construction completion status, and this consistency result is... Figure 5 The verification phase is represented by a curve.

[0066] Subsequently, the system performed a consistency check on the verification evidence. It found that while the verification evidence was consistent with each other, its conclusion conflicted with some of the previous evidence, and after considering all the evidence, it still could not meet the consistency conditions required for the original evidence to be valid. Based on this determination, the system updated the progress status of this process node to "cancelled," indicating that the previously formed progress conclusion lacked a basis for confirmation.

[0067] After updating the progress status to "Cancelled," the system automatically activates enhanced confirmation rules, setting stricter criteria for subsequent progress confirmations. Specifically, the system raises the credibility level threshold required for evidence to be valid from Level 2 to Level 3, and increases the number of required evidence categories from 3 to 4. New requirements include complete material consumption records and third-party supervision confirmation records. The enhanced confirmation rules are implemented in... Figure 5 This is reflected in the threshold changes shown. After the enhanced confirmation rule takes effect, the system continues to perform evidence collection and mutual verification for that process node. Only when the subsequently collected multi-source evidence simultaneously meets the requirements of a higher credibility level threshold and more evidence category coverage is the progress status of that process node allowed to be reconfirmed.

[0068] After the progress status of a work process node is updated to "Cancelled" and enhanced confirmation rules are enabled, the system automatically identifies a completion deviation for that node based on the difference between the planned and actual progress, and triggers a corrective action. Simultaneously with triggering the corrective action, the system generates a corrective instruction carrying the corrective verification objective, which constrains the scope and verification standards of this corrective action. The corrective verification objective explicitly requires the completion of four types of evidence within the review window: on-site video and image evidence, actual concrete supply records, construction personnel work records, and third-party supervision confirmation records. Furthermore, these pieces of evidence must meet consistency requirements regarding construction completion status, construction location, and work process object. Figure 6 The heatmap of the evidence completion process visually illustrates the changes in evidence completion status during different corrective attempts.

[0069] After the corrective action was issued, the construction site implemented targeted corrective measures as required by the instruction. These measures included supplementing the site with multi-angle video footage, simultaneously exporting complete material supply data from the concrete supply system, and organizing on-site confirmation by supervisors. After the corrective action was completed, the system restarted the evidence re-collection process for that work step and obtained new evidence data in the verification window. The newly added video evidence showed that all pouring areas were complete, the concrete supply system recorded a cumulative supply of 152 cubic meters, exceeding the planned supply of 150 cubic meters, and both the construction personnel's work records and the supervisor's confirmation records indicated that the work step was completed. After cross-verifying the re-collected evidence, the system confirmed that the four types of evidence were consistent in key construction characteristics. The calculated evidence coverage rate was 1.00, and the consistency difference was less than the preset threshold, meeting the evidence category completion and consistency conditions required by the corrective action verification objective. Therefore, the corrective action was deemed effective.

[0070] In another contrasting scenario, such as Figure 6As shown in the first two lines, when the system re-collected evidence after correction, it only supplemented two types of evidence: video evidence and material supply records. Construction worker operation records and supervisor confirmation records remained missing, resulting in an evidence coverage rate of 0.50. Although the evidence from different sources did not contradict each other regarding the completion status of the construction, and the consistency difference was small, the re-collected evidence still failed to meet the correction verification objective. Based on this, the system recorded the number of deviations. When two consecutive correction attempts failed to meet the correction verification objective, it entered the escalation judgment stage according to the preset correction level order.

[0071] During the escalation of the correction level, the system simultaneously determines whether the escalation suppression condition is met. After analyzing the re-collected evidence, the system found that although there were gaps in the evidence categories in the first two correction attempts, the obtained evidence did not conflict with the completed construction status, and the consistency difference was less than the preset threshold, indicating that the deviation was mainly caused by evidence gaps rather than evidence conflicts. Based on this, the system determined that the escalation suppression condition was met. When the escalation suppression condition is met, the system no longer escalates the correction level, but only implements evidence completion-related correction measures, such as extending the review window and increasing the frequency of evidence collection, allowing more time for subsequent re-collected evidence to be completed.

[0072] Within the extended review window, the system finally supplemented the missing construction worker work records and supervisor confirmation records, and performed a consistency check again to confirm that all kinds of evidence were consistent in the construction completion status, thereby determining that the correction was effective and ending the correction process. Figure 6 The evidence-complete heatmap and corresponding calculation results show that the present invention can effectively correct construction progress deviations while avoiding over-correction, fully demonstrating the engineering applicability and stability of the smart construction site construction progress deviation dynamic identification and correction method in complex construction scenarios.

[0073] Example 2:

[0074] In this embodiment, after completing the correction process in Embodiment 1, the system enters the correction verification and progress confirmation stage. For this process node, the system simultaneously sets the correction verification target when generating the correction instruction, and further clarifies the correction verification deadline conditions within the correction verification target to control the termination timing of the correction process and the final confirmation of the progress status. Combined with... Figure 7 The diagram illustrates the triggering of the correction verification deadline. Within the review window, the system monitors two paths in parallel: evidence coverage and the number of evidence consistency confirmations. When either condition is met, it triggers an early termination, terminating subsequent evidence collection or the execution of correction actions, and enters the progress status locking process. Figure 7The document also provides calculation prompts consistent with the data in the embodiment, including trigger round markers for the evidence coverage threshold and the consistency confirmation number threshold, as well as prompts for the ratio of the supply quantity to the planned usage quantity, to corroborate the feasibility and verifiability of the correction verification cutoff determination in this embodiment.

[0075] In this embodiment, the system sets two types of judgment paths for the correction verification cutoff conditions for this process node: one is to achieve a preset evidence coverage rate within the review window, and the other is to achieve a preset number of evidence consistency confirmations within the review window. Specifically, the system sets the number of evidence categories required for the correction verification target to four: on-site video image evidence, actual concrete supply records, construction personnel operation records, and third-party supervision confirmation records. The evidence coverage rate threshold is set to 0.75, corresponding to at least three types of valid evidence being obtained to trigger the cutoff judgment. At the same time, the system sets the evidence consistency confirmation number threshold to three times, requiring three consecutive consistency judgment results to be obtained within the review window, and adjacent confirmations to come from at least different evidence sources. Figure 7 Using the relative rounds within the review window as the horizontal axis, the triggering logic of the two paths mentioned above is displayed synchronously. The evidence coverage rate is represented by a stepped curve, and the number of evidence consistency confirmations is represented by a broken line curve. The coverage rate threshold of 0.75 and the confirmation number threshold of 3 are marked with dashed lines in the figure.

[0076] In the review window following the completion of the correction, the system first collected three types of evidence: on-site video equipment, the concrete supply system, and the mobile terminals of construction workers. The video evidence showed that all pouring areas were complete, the concrete supply system recorded a cumulative supply of 151 cubic meters, exceeding the planned supply of 150 cubic meters, and the construction workers' work records indicated that the process was completed and had entered the curing stage. After performing a consistency check on the above evidence, the system confirmed that the three types of evidence were consistent in terms of construction completion status and process objects. At this point, the evidence coverage rate was calculated to be 3 / 4 = 0.75, reaching the preset evidence coverage rate threshold. Figure 7 The calculation results and explanations below provide a numerical verification relationship consistent with this paragraph, namely, the total number of evidence categories is 4, the coverage threshold is 0.75, and the coverage rate of path A is 3 / 4=0.75 when it reaches 3 types of evidence. It also gives a hint that the supply ratio is 151 / 150=1.007, which is used to characterize the matching trend between supply records and completion status.

[0077] Based on the above calculations, the system determines that the deadline for corrective verification has been met and immediately terminates the subsequent evidence collection process, as well as the ongoing corrective actions. The system then updates and locks the progress status of this work process node from the "in dispute" state to the "confirmed" state, indicating that the current progress conclusion is supported by sufficient evidence. After the progress status is locked as confirmed, the system will not trigger the corrective or verification process for this work process node again until the construction plan moves to the next work process node. Figure 7 By marking the trigger rounds for path A, the system can terminate early when the coverage reaches a threshold, thereby shortening the invalid collection and repeated correction within the review window and improving the efficiency of progress confirmation.

[0078] In another review path, the system did not quickly reach the evidence coverage threshold at the beginning of the review window. However, by continuously collecting evidence from multiple sources, it obtained three consecutive consistency judgment results within the review window. Specifically, within the review window, the system first formed a consistency confirmation based on video evidence and material supply records, then formed a second consistency confirmation based on material supply records and supervision confirmation records, and finally formed a third consistency confirmation based on video evidence and supervision confirmation records. All three consistency confirmations indicated that the concrete pouring process had been completed, and the evidence sources corresponding to two adjacent confirmations differed at least once, thus meeting the requirement for the number of evidence consistency confirmations. Figure 7 The arrow markings indicating the trigger round when the consistency confirmation count curve for path B reaches 3 are consistent with the mechanism of 3 consecutive confirmations in this segment. The explanation below the figure describes how path B triggers the cutoff condition when it reaches 3 consecutive confirmations in the 3rd round, illustrating that even if the coverage target is not quickly met, the cutoff can still be achieved early through the number of consistency confirmations.

[0079] Based on this, the system determines that the deadline for correction verification is also met, and terminates the evidence collection and correction action before the end of the review window, locking the progress status of the corresponding process node as confirmed.

[0080] To further enhance the granularity and engineering applicability of the correction verification results during the system's error correction process, a weighted evidence coverage calculation method based on evidence category weights is introduced to quantitatively determine the correction verification cutoff conditions. This weighted evidence coverage considers not only whether evidence categories have been obtained, but also the credibility level of the evidence source, the importance of the evidence category, and the timeliness of the evidence, thereby avoiding the simplistic equal weighting of different pieces of evidence in progress assessment. Combined with... Figure 8 The diagram shows the decomposition of evidence weights and the calculation of weighted evidence coverage. Figure 8The data is presented in a grouped bar chart, showing the values ​​of each type of evidence in terms of credibility contribution, importance contribution, timeliness contribution, and total weight. The real-time calculation results of the numerator, denominator, and weighted coverage are given on the right side of the chart, forming a verifiable quantitative basis.

[0081] For the concrete pouring process nodes of each floor, the system sets the evidence categories required for the corrective verification target into four types: on-site video and image evidence, actual concrete supply records, construction personnel operation records, and third-party supervision confirmation records. The resulting evidence category set for the corrective verification target is denoted as follows: Each element corresponds to one of the four types of evidence mentioned above.

[0082] Within the review window, the system actually acquired and passed the consistency verification for three categories of evidence: on-site video and image evidence, actual concrete supply records, and construction worker operation records. Third-party supervision confirmation records have not yet been uploaded. Therefore, the set of evidence categories that have been acquired and passed the consistency verification is as follows: And it is obviously satisfied. The system defines the evidence coverage rate as the weighted evidence coverage rate (CR), and its calculation formula is as follows:

[0083]

[0084] in Indicates the category of evidence The weight is used to reflect the overall contribution of this type of evidence to progress confirmation.

[0085] The weights of evidence categories are calculated using the following formula:

[0086]

[0087] in, Indicates the category of evidence The numerical results corresponding to the credibility level of the evidence source. Indicates the category of evidence The preset importance coefficient, This represents the time difference between the current determination time and the most recent valid collection time that passed the consistency check for this type of evidence. , , The weighting coefficients are non-negative and at least one of them is positive. This is the time-related decay coefficient. Figure 8 The above three contributions have been visually broken down to facilitate the auditing and tracing of the source of weights during project implementation.

[0088] The system settings are as follows: , , , For the four categories of evidence, the system set and collected the following data:

[0089] The credibility level of the on-site video image evidence is 3, the importance coefficient is 3, and the time difference between the most recent consistency check and the current judgment time is 1.

[0090]

[0091]

[0092]

[0093] The reliability level of the actual concrete supply record is 3, the importance coefficient is 4, and the time difference is 0.

[0094]

[0095]

[0096]

[0097] The reliability level of the construction workers' work records is 2, the importance coefficient is 2, and the time difference is 2.

[0098]

[0099]

[0100]

[0101] The third-party supervision confirmation record has a credibility level of 4 and an importance coefficient of 4. However, no valid collection record that has passed the consistency check has been obtained within the review window. The system calculates its timeliness item according to the initial value, making... :

[0102]

[0103]

[0104]

[0105] Based on the above calculation results, the system calculates the numerator and denominator respectively:

[0106]

[0107]

[0108] This yields the weighted evidence coverage rate:

[0109]

[0110] The system sets the weighted evidence coverage threshold to 0.65, and there are no evidence conflicts within this review window. Since the calculated weighted evidence coverage CR has reached the preset threshold, the system determines that the correction verification cutoff condition is met. Without fully supplementing all evidence categories, the system prematurely terminates the subsequent evidence collection and correction actions, and simultaneously locks the progress status of this process node to the confirmed state. Figure 8 The calculation results and explanations below provide numerator and denominator values ​​consistent with this section, specifying that the numerator is 7.245 and the denominator is 10.645. The calculated CR is 0.68, which is higher than the threshold of 0.65, thus consistent with the early cutoff conclusion of this embodiment.

[0111] After the system determines the cutoff condition for corrective verification is met based on weighted evidence coverage and locks the progress status of the process node as confirmed, the system simultaneously introduces the evidence consistency confirmation count as another independent and parallel-effective corrective verification cutoff determination path to further enhance the stability and anti-interference capability of the progress confirmation results. The evidence consistency confirmation count is defined as the number of consecutive confirmations that meet the consistency condition for the same process node within the review window, requiring that the evidence sources relied upon by two adjacent confirmations differ at least once to avoid systematic bias caused by repeated confirmations from a single evidence source. Figure 9 The diagram shows the stacked area illustrating the evolution of the progress status percentage of each process node. Figure 9 Using the relative stages of the review process as the horizontal axis and the percentage of status as the vertical axis, the graph shows the evolution of the percentage of statuses such as confirmed lock, disputed, under verification, restored confirmed, and revoked in the process. The graph also marks the key nodes of stage 1 triggering unlocking and entering dispute, stage 2 generating verification instructions, and stage 3 verification diversion, making the logical relationship between lock holding, unlocking, verification, and restoration or revocation escalation visible.

[0112] For the concrete pouring process, the system sets a preset threshold of three confirmations for evidence consistency. Within the review window, the system continuously performs consistency checks on newly collected multi-source evidence, generating multiple confirmation results sequentially. Specifically, the system first confirms consistency based on on-site video evidence and actual concrete supply records, confirming the completion of the pouring area and a supply volume of 151 cubic meters. Subsequently, the system confirms consistency based on the actual concrete supply records and third-party supervision records, both indicating the completion of the process. Finally, the system confirms consistency based on on-site video evidence and construction worker work records, ensuring consistency between the video evidence and work records in the completion status. In these three confirmations, at least one of the adjacent confirmations has a different source of evidence, and all meet the consistency criteria. Therefore, the system determines that the number of evidence consistency confirmations has reached the preset number and can proceed accordingly. Figure 7 The path B mechanism shown triggers an early termination.

[0113] Once the number of evidence consistency confirmations reaches a preset threshold, the system immediately determines that the correction verification deadline has been met, even if the review window has not yet ended. It then prematurely terminates subsequent evidence collection and correction actions, and locks the progress status of that process node to the confirmed state. This progress status lock indicates that the progress conclusion of the current process node has received sufficient and stable evidence support within the review window, and the system will no longer trigger correction or verification operations for that process node. Figure 9 Phase 4 shows an increase in the percentage of recovered confirmed states returning to confirmed locked states, which indicates that the system completes state closure and returns to locked state under the condition of consistent verification.

[0114] After the progress status is locked as confirmed, the system simultaneously sets lock maintenance and unlocking conditions for this locked status to address potential abnormal evidence situations. The lock maintenance condition is that no conflicting evidence with a credibility level higher than the credibility level of the evidence on which the confirmed status was based appears during the lock period. In this embodiment, the credibility level of the main evidence on which the confirmed status was based is 3. Therefore, during the lock period, if the system only collects scattered conflicting evidence with a credibility level no higher than 3, the unlocking operation is not triggered, and the progress status remains confirmed. Figure 9 The initial state is illustrated by a confirmed lock percentage of 1.00 before phase 0 to phase 1, and the percentage before phase 1 without any disputes is used to represent the operational scenario where the lock holding conditions are not violated.

[0115] In another operational scenario, during the lockout period, the system acquired new evidence from a third-party specialized testing device. This evidence was rated with a credibility level of 4 and indicated that some areas had incomplete pouring. After performing a conflict determination on this evidence, the system confirmed that it conflicted with the already confirmed status, and that its credibility level was higher than the credibility level of the evidence upon which the confirmed status was based. Based on this, the system met the unlocking conditions, automatically released the progress status lock of the process node, and updated the progress status of that process node to "Under Dispute." Figure 9 In Phase 1, the percentage of unlocking into the disputed area is represented by 0.20. The text annotation on the right side of the figure clearly states that in Phase 1, the detection of highly credible conflict evidence triggers the unlocking into the disputed area.

[0116] After updating the progress status of the work process node to "Under Dispute," the system automatically generates and executes a verification instruction within a new preset verification window to re-verify the actual completion status of the work process. The verification instruction explicitly limits the verification period to the current verification window, the verification location to the corresponding floor's concrete pouring construction area, and the verification objects to the pouring completion status and actual concrete usage. It also requires obtaining verification evidence from at least two different sources. Based on the verification instruction, the system collects new on-site video image evidence and material supply data exported from the concrete supply system. The video evidence shows that all pouring areas have been completed, and the cumulative usage in the material supply data is 153 cubic meters, exceeding the planned usage of 150 cubic meters. Figure 9 In Phase 2, the sample in dispute enters the verification process with a percentage of 0.20 in the verification process. The verification supply ratio is marked as 153 / 150=1.02 on the right side of the figure. The supply quantity of 153 and the planned usage of 150 are also given to support the quantitative review in the verification phase.

[0117] After performing consistency checks on the above verification evidence, the system confirms that the two types of verification evidence are consistent in terms of construction completion status and process objects, and meet the key evidence set and consistency conditions. Based on this, the system restores the progress status of the process node from disputed to confirmed. Figure 9 In Phase 3, the verification process is further divided into two categories: 0.15% for confirmed restoration and 0.05% for withdrawn verification. This represents the two possible branches of consistency and inconsistency that may occur during the verification phase. The calculation results for the restoration percentage (15 / 20 = 0.75) and the withdrawal percentage (5 / 20 = 0.25) are provided in the text annotation on the right side of the diagram to illustrate the proportional relationship of the verification result distribution. If, in another comparison scenario, the verification evidence still fails to meet the consistency condition, or if high-confidence-level conflicting evidence reappears, the system updates the progress status of that process node to withdrawn and simultaneously triggers a correction level upgrade process to re-address the progress deviation problem of that process node through higher-level correction measures. Figure 9The 0.05% repatriation in Phase 4 represents the possibility of triggering an escalation and re-entering the dispute review path after withdrawal, thus maintaining consistency with the withdrawal and escalation process in this section.

Claims

1. A method for dynamic identification and correction of construction progress deviations in smart construction sites, characterized in that... include: The target process of the construction plan is divided into process nodes, and key evidence sets and credibility levels of evidence sources are preset for each process node; At the construction site, multiple sources of evidence corresponding to the key evidence set are collected periodically and cross-verified. The evidence is determined to be valid, insufficient, or conflicting based on the consistency of the evidence. Based on the judgment result, update the progress status of the process node within the preset review window; identify the start deviation or completion deviation by combining the start time window, the completion time window and the progress status; trigger and execute the correction according to the preset correction level order; recollect evidence after correction and repeat the deviation identification.

2. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 1, characterized in that... For each piece of evidence collected, an evidence digest code containing the source identifier, collection time, collection location, and object identifier is generated and recorded in chronological order. During mutual verification, the continuity and repetition of the digest code sequence are checked. If there is a break in the sequence, repetition, or the collection time is earlier than the first recording time, the corresponding evidence is marked as supplementary evidence and the credibility level is downgraded. In the consistency judgment, supplementary evidence has a lower priority than non-supplementary evidence.

3. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 1, characterized in that... When the consistency determination is a conflict of evidence, or when the evidence is continuously determined to be insufficient within the preset review window, a verification instruction is generated and executed. The verification instruction limits the verification period, location and object, and requires the formation of verification evidence from at least two different sources. If the verification evidence still does not meet the consistency conditions, the status of the corresponding process node is updated to "revoked", and enhanced confirmation rules are enabled. The enhanced confirmation rules include raising the credibility level threshold required for the evidence to be established or increasing the number of required evidence categories.

4. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 1, characterized in that... When the correction is triggered, a correction instruction carrying the correction verification target is generated. The correction verification target is the evidence category and consistency conditions that need to be supplemented in the review window. After correction, the correction is deemed effective only if the re-collected evidence meets the correction verification target. Otherwise, when the deviation occurs continuously to the preset number of times or the process node status repeatedly migrates between dispute and cancellation to the preset number of times, the correction is upgraded step by step according to the preset correction level order and the correction is executed again.

5. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 4, characterized in that... The corrective verification target includes a corrective verification cutoff condition, which is to achieve a preset evidence coverage rate or a preset number of evidence consistency confirmations within the review window. When the cutoff condition is met in advance within the review window, subsequent evidence collection or corrective action execution is terminated, and the progress status of the corresponding process node is locked as confirmed until the next process node is entered.

6. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 4, characterized in that... The step-by-step escalation according to the preset correction level includes setting escalation suppression conditions. The escalation suppression conditions are that the evidence obtained through re-collection is caused by evidence gaps and there are no evidence conflicts. When the escalation suppression conditions are met, only evidence completion corrections are performed and the review window is extended.

7. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 5, characterized in that... The evidence coverage rate is calculated by evidence category and is defined as the ratio of the number of evidence categories that have been obtained and passed the consistency verification within the review window to the number of evidence categories required by the correction verification target. When the evidence coverage rate reaches a preset threshold and there is no evidence conflict, it is determined that the correction verification cutoff condition is met.

8. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 5, characterized in that... The number of evidence consistency confirmations is the number of times that the same process node continuously obtains confirmations that meet the consistency conditions within the review window, and the evidence sources corresponding to two adjacent confirmations are different at least once; when the number of evidence consistency confirmations reaches the preset number, it is determined that the correction verification cutoff condition is met and the termination of subsequent evidence collection or termination of correction action is triggered.

9. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 5, characterized in that... The progress status lock is confirmed to include lock holding conditions and unlocking conditions; the lock holding condition is that no conflicting evidence with a credibility level higher than the credibility level of the evidence on which the confirmed status is based appears during the lock period; The unlocking condition is that when evidence that meets the conflict determination conditions and has a higher credibility level than the evidence on which the evidence is based appears during the locking period, the lock is released and the progress status of the process node is updated to "in dispute".

10. The method for dynamic identification and correction of construction progress deviations in smart construction sites according to claim 9, characterized in that... After unlocking and updating the progress status of the process node to "in dispute", a verification instruction is generated in the preset review window to limit the verification time period, location and object, and at least two types of verification evidence from different sources are obtained. When the verification evidence meets the key evidence set and consistency conditions, the progress status will be restored to confirmed; otherwise, it will be updated to revoked and the correction level will be upgraded.

Images