RPA-based bank-enterprise pledge data intelligent back transmission system

Abstract

The application discloses a bank-enterprise pledge data intelligent back transmission system based on RPA and relates to the technical field of financial data processing.The system comprises the following steps: performing page positioning, field writing and back transmission submission on RPA back transmission arrangement information, performing field matching on page controls with position changes, in the network timeout and task interruption state, performing unfinished field writing and back transmission submission according to idempotent identification and recovery checkpoints, obtaining RPA back transmission execution tracks and current bank receipt information, corresponding the RPA back transmission execution tracks and the current bank receipt information according to idempotent identification, extracting the processing state of the pledge back transmission task, updating the received successful content to the bank confirmation state and releasing the corresponding back transmission debt, hanging the return reason to the original pledge back transmission task, forming the to-be-corrected content, marking the processing task as a receipt to be confirmed, and generating the bank-enterprise pledge data back transmission state.The application improves the back transmission stability and traceability.

Description

Technical Field

[0001] This invention relates to the field of financial data processing technology, and in particular to an intelligent backhaul system for bank-enterprise pledge data based on RPA. Background Technology

[0002] With the digital development of bank pledge financing business, pledge supervision is gradually shifting from manual inventory to edge IoT sensing, enterprise business system collaboration, and online processing at the bank end. The monitoring equipment continuously forms the on-site status of the pledged object, the enterprise system records the pledge contract and business changes, the bank system receives and processes the progress through receipts, and RPA technology is used for cross-system page positioning, field filling and task submission. This forms a bank-enterprise data processing model that focuses on the collection of pledged object status, business aggregation, automatic feedback and receipt tracking.

[0003] The existing technology for transmitting pledged data between banks and enterprises has two main problems: First, the on-site supervision status, enterprise registration status and bank confirmation status generated by edge IoT lack a unified temporal correlation. When status changes are transmitted between different systems, the positions are easily out of sync, making it difficult to accurately determine the objects to be transmitted and the associated evidence. Second, RPA usually executes according to a fixed page path. When control positions change, network timeouts occur and tasks are interrupted, it is easy to cause incorrect field writing, duplicate submissions and distortion of the transmitted status. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an intelligent backhaul system for bank-enterprise pledge data based on RPA to solve the problems of insufficient temporal correlation of multi-source pledge status and limited recovery capability of RPA backhaul anomalies.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides an RPA-based intelligent data feedback system for bank-enterprise pledged assets, comprising: a data acquisition module for collecting on-site supervision information, enterprise registration information, and historical bank receipt information of the pledged assets; calibrating the status according to the asset attribution and business sequence to form a physical evidence status, an enterprise supervision status, and a bank confirmation status; verifying the validity of the status and marking the time continuity to generate an initial pledged asset status image; a debt identification module for aligning the initial pledged asset status image with the time sequence and comparing status differences to form a status propagation gap; and performing back-feed debt identification and related transfer along the connection relationships between the pledged assets and the pledge contract, as well as between the supervision equipment and the ownership certificate, to generate pledged asset back-feed debt information; and a back-feed orchestration module for, based on the pledged asset back-feed debt information, locating the asset attribution and filtering the back-feed scope to obtain the pledged assets to be back-feeded, and performing status content conversion and related certificate verification. According to the connection, it is encapsulated as a pledge return task, and idempotency flags and recovery checkpoints are configured to generate RPA return orchestration information; the field return module is used to perform page positioning, field writing and return submission on the RPA return orchestration information, perform field matching for page controls with changed positions, and perform incomplete field writing and return submission according to the idempotency flags and recovery checkpoints in the case of network timeout and task interruption, and obtain the RPA return execution trajectory and current bank receipt information; the status update module is used to match the RPA return execution trajectory and current bank receipt information according to the idempotency flag, extract the processing status of the pledge return task, update the received successful content to the bank confirmation status and release the corresponding return debt, attach the return reason to the original pledge return task to form the content to be corrected, mark the task in processing as pending confirmation, and generate the bank-enterprise pledge data return status.

[0007] As a preferred embodiment of the intelligent backhaul system for bank-enterprise pledged data based on RPA described in this invention, the steps for generating the initial pledged state image are as follows: The system performs object identification parsing on-site supervision information, enterprise registration information, and historical bank receipt information of the pledged object. Based on the pledged object identifier in the enterprise registration information and the ownership of the pledge contract, the system links the on-site supervision information and the supervision equipment identifier and ownership certificate identifier in the enterprise registration information to the pledged object index, collects information fragments pointing to the same pledged object, and generates pledged object status collection information. Based on the information collected on the status of pledged objects, the evidence source of on-site supervision information is confirmed to form a physical evidence status, the registration information of enterprises is assigned to registration items to form an enterprise supervision status, the historical bank receipt information is returned to the corresponding task to form a bank confirmation status, and the chronological relationship of each status is corrected according to the time of business occurrence and the time of status update to generate pledge status hierarchical information. The system performs object identifier consistency comparison on the physical evidence state, enterprise supervision state, and bank confirmation state in the pledge status layer information, verifies the business nodes and formation time corresponding to each state, merges state fragments with duplicate content, excludes state fragments with time inversion and object mismatch, marks the validity of the retained state and marks the preceding and following positions, and generates an initial pledge status image.

[0008] As a preferred embodiment of the RPA-based intelligent backhaul system for bank-enterprise pledge data described in this invention, the steps for forming the state propagation gap are as follows: Based on the initial pledged state mirror, and using the pledged object index and the time of business occurrence as the alignment benchmark, the state fragments in the physical evidence state, corporate supervision state, and bank confirmation state are sequentially organized, and the state fragments reflecting the same business change are attached to the same time position, and the difference in state update time is corrected to generate pledged state time-series aligned data. By comparing the differences in the state content, update time, and order of changes in the pledged state time-series alignment data, the location where the existing state has changed but the associated state has not yet been synchronized is located. At the same time, the location where the state update is stopped and the misalignment is marked, forming a state propagation gap.

[0009] As a preferred embodiment of the intelligent backhaul system for bank-enterprise pledged data based on RPA described in this invention, the steps for generating pledged debt information are as follows: Based on the state propagation gap, the business attribution of the gap is determined along the connection between the pledged object and the pledge contract, the scope of evidence involved in the gap is determined along the connection between the regulatory equipment and the ownership certificate, the back-transferred debt value is calculated according to the duration of the state propagation gap, the degree of state difference and the scope of related transfer, and the back-transferred debt value is marked to the location of the state propagation gap to generate the initial back-transferred debt distribution; Based on the initial distribution of returned debt, the returned debt value is transferred along the associated direction between the pledged object, pledge contract, monitoring equipment and ownership certificate, and the transfer path pointing to the same pending synchronization matter is merged. The source location, propagation direction, scope of influence and returned debt value of the returned debt are retained to generate pledged returned debt information.

[0010] As a preferred embodiment of the intelligent data feedback system for bank-enterprise pledge based on RPA as described in this invention, the steps for obtaining the pledge object to be fed back are as follows: Extract the source location, debt transfer path and pledge object index from the status propagation gap of pledged debt information, trace the pledge business to which the status change belongs along the connection between the pledge object and the pledge contract, classify the status propagation gap into the pledge object, and generate gap attribution location information. Based on the location information of the gap, and combined with the transfer status of the returned debt on the regulatory equipment and ownership certificate nodes, the pledged objects that have been confirmed by the bank and have not formed a valid status change are excluded, and the pledged objects that need to be supplemented, changed and confirmed in execution status are retained, and the information of the pledged objects to be returned is generated.

[0011] As a preferred embodiment of the intelligent backhaul system for bank-enterprise pledged data based on RPA described in this invention, the steps for generating RPA backhaul orchestration information are as follows: Based on the information of the pledged object to be returned, extract the state fragments and supporting evidence on both sides of the state propagation gap, convert the state content according to the meaning of the fields on the return page of the target financial institution, establish the linkage between the converted field content and the supporting evidence, and encapsulate it into a pledge return task. The pledge return tasks are arranged in order according to the formation sequence of state propagation gaps and the return debt transition path. Idempotent flags are configured in combination with pledge object index, state update time and task content. Recovery checkpoints are set at field write and return submission nodes to generate RPA return orchestration information.

[0012] As a preferred embodiment of the RPA-based intelligent data feedback system for bank-enterprise pledges described in this invention, the steps of performing page positioning, field writing, and feedback submission are as follows: Based on the RPA back-to-archive information, the RPA process is started according to the execution order of the pledge back-to-archive task. The page path and field content to be written of the target financial institution are extracted, and the pledge business page is entered. The field filling position is locked according to the page level, field name and control attribute, and page positioning data is generated. Based on the page location data, the field content in the staking return task is written into the locked page control. The page display content is compared with the field content to be written. For page controls whose positions have changed, the field name, input format and adjacent tags are matched to determine the available controls and complete the return submission, generating RPA task submission information.

[0013] As a preferred embodiment of the intelligent backhaul system for bank-enterprise pledged data based on RPA described in this invention, the steps for obtaining the RPA backhaul execution trajectory and current bank receipt information are as follows: Based on RPA task submission information, the execution nodes of page positioning, field writing and return submission are recorded. When network timeout and task interruption occur, the idempotency flag is used to determine the task submission status, and the unfinished field writing and return submission are continued to be executed according to the recovery checkpoint, generating the RPA return execution trajectory. Based on the RPA execution trajectory, extract the task acceptance information and processing feedback returned by the bank's page, bind the processing feedback to the pledge return task according to the idempotency identifier, organize the task submission nodes and bank processing content, and generate the current bank receipt information.

[0014] As a preferred embodiment of the RPA-based intelligent data feedback system for bank-enterprise pledge data described in this invention, the steps for extracting the processing status of the pledge feedback task are as follows: Extract the idempotency identifier from the RPA back-transmission execution trajectory and the current bank receipt information, attach the current bank receipt information to the pledge back-transmission task with the same idempotency identifier, and organize the execution sequence according to the task submission time and receipt formation time to generate back-transmission task receipt association data. The bank acceptance feedback and processing progress in the data associated with the return task receipts are analyzed, and the pledge return tasks are divided into successfully received tasks, returned tasks, and tasks in process. The receipt content and original execution position of each task are retained, and the processing status information of the return task is generated.

[0015] As a preferred embodiment of the intelligent backhaul system for bank-enterprise pledged data based on RPA described in this invention, the steps for generating the backhaul status of bank-enterprise pledged data are as follows: Based on the status information of the back-transmission task, the pledge status in the successfully received task is written into the bank confirmation status of the initial pledge status mirror, the back-transmission debt mark on the status propagation gap is removed, and the return reason is extracted from the receipt content of the returned task. The return reason is attached to the original pledge back-transmission task to generate back-transmission task processing data. The system summarizes the fields to be modified and the evidence to be supplemented in the data of the returned task processing to form the content to be corrected. For tasks in progress, the original returned debt is retained and a receipt confirmation mark is configured. The system collects the task processing status, debt release status and content to be corrected, and generates the data return status of bank-enterprise pledge data.

[0016] The beneficial effects of this invention are as follows: By constructing a pledge status mirror, aligning the time sequence, and transmitting debt back, the difference between the edge IoT status and the bank / enterprise status is located, which is used to screen the objects to be transmitted back and the scope of evidence, reducing the risk of missed transmission, incorrect transmission, and duplicate transmission; by arranging pledge tasks, idempotent control, restoring checkpoints, and associating with bank receipts, RPA transmission anomaly continuation and processing status update are realized, which is used to ensure the accurate transmission of pledge data and improve transmission stability and traceability. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of an intelligent data feedback system for bank-enterprise pledge data based on RPA.

[0019] Figure 2 To transmit a scatter plot showing the relationship between debt value and the degree of difference in status.

[0020] Figure 3 This diagram illustrates the impact of changes in the position of page controls on the accuracy of field postback.

[0021] Figure 4 A radar chart comparing key indicators for intelligent data feedback of bank-enterprise pledge data. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides an intelligent data feedback system for bank-enterprise pledge based on RPA, including the following steps: The data acquisition module is used to collect on-site supervision information, enterprise registration information and historical bank receipt information of the pledged objects. It performs status calibration according to the object's ownership and business time sequence to form physical evidence status, enterprise supervision status and bank confirmation status, and performs status validity verification and time continuity marking to generate an initial pledge status image.

[0026] The system performs object identification parsing on-site supervision information, enterprise registration information, and historical bank receipt information of the pledged objects. Based on the pledged object identifier in the enterprise registration information and the ownership of the pledge contract, the system links the on-site supervision information and the supervision equipment identifier and ownership certificate identifier in the enterprise registration information to the pledged object index, collects information fragments pointing to the same pledged object, and generates pledged object status collection information.

[0027] Furthermore, the format and validity of object identifiers in on-site supervision information, enterprise registration information, and historical bank receipt information are standardized. Identifiers that can represent the same pledged object are converted into a pledged object index. Using the pledged object identifiers and pledge contract attribution recorded in the enterprise registration information as the collection benchmark, the object names, contract attribution, and business occurrence locations in each information segment are verified to eliminate object dispersion caused by differences in identifier formats, duplicate registrations, and inconsistent source naming. For information segments with missing identifiers or conflicting attributions, their source marks and abnormal positions are retained, and the attachment process is suspended to avoid incorrectly classifying information of different pledged objects into the same object index.

[0028] After completing object identifier resolution, the regulatory equipment identifier and ownership certificate identifier are extracted from on-site supervision information and enterprise registration information. A linkage relationship is established according to the actual regulatory object of the equipment, the object recorded in the certificate, and the ownership of the pledge contract. When the regulatory equipment identifier matches the pledge object index, the regulatory record generated by the equipment is included in the pledge object. When the ownership certificate identifier matches the pledge contract and object identifier, the certificate status is linked to the same object index. Information fragments from different sources but pointing to the same pledge object are merged, and the source, formation time, and linkage basis of each information fragment are retained to form a pledge object status collection information that can reflect the association status of the pledge object, pledge contract, regulatory equipment, and ownership certificate.

[0029] Based on the information collected on the status of pledged objects, the evidence source of on-site supervision information is confirmed to form a physical evidence status, the registration information of enterprises is classified according to the registration items to form an enterprise supervision status, the historical bank receipt information is returned to the corresponding task to form a bank confirmation status, and the chronological relationship of each status is corrected according to the time of business occurrence and the time of status update to generate pledge status hierarchical information.

[0030] Furthermore, regarding the pledge object index in the pledge object status collection information, a consistency check is performed on the collection source, formation time, and supervised object carried by the on-site supervision information to confirm whether it was generated by the connected supervision equipment or on-site supervision process; supervision segments with valid sources and consistent object affiliation are written into the physical evidence state according to the actual location reflected, changes in inventory, and equipment response; segments with missing source markers, abnormal collection times, or inconsistent supervised objects retain abnormal markers and do not enter the valid state; the registration items, business change records, and pledge contract affiliation in the enterprise registration information are correlated and organized, and status segments that can reflect the enterprise-side pledge management process are written into the enterprise supervision state, while retaining the affiliation relationship between the registration items and the pledge object index.

[0031] For historical bank receipt information, the original pledge return task is traced according to the task identifier, pledge object index, and submission time carried on the receipt. The content of receipts that have been accepted, returned, or are still being processed by the bank is written into the bank confirmation status to avoid directly regarding the enterprise-side status without bank feedback as the bank confirmation status. After each status is formed, the actual status change order is determined by the time of the business occurrence, and the delayed upload and supplementary entry are identified by the status update time. The positions of status fragments with conflicting order are adjusted, and the original formation time and the corrected continuation position are retained. The physical evidence status, enterprise supervision status, and bank confirmation status are hierarchically collected according to the pledge object and business time sequence to generate pledge status hierarchical information.

[0032] It should be noted that physical evidence status is used to record the status information of the pledged object formed by actual supervision sources such as on-site supervision equipment and collection terminals, reflecting the objective existence status of the pledged object in the actual supervision environment.

[0033] Enterprise Supervision Status: This is used to record the status of pledged objects, pledge contracts and related management matters registered in the enterprise's business system, reflecting the enterprise's registration and management of pledged business.

[0034] Bank confirmation status: This is used to record the acceptance, processing, and confirmation status information returned by the bank system for pledge return tasks, reflecting the bank's actual reception of changes in pledge status.

[0035] The pledge status layering information is organized according to the pledge object index and business time sequence, and the physical evidence status, corporate supervision status and bank confirmation status are layered and organized, and the structured status data of each status source, formation time and receiving position are retained.

[0036] The system performs object identifier consistency comparison on the physical evidence state, enterprise supervision state, and bank confirmation state in the pledge status layer information, verifies the business nodes and formation time corresponding to each state, merges state fragments with duplicate content, excludes state fragments with time inversion and object mismatch, marks the validity of the retained state and marks the preceding and following positions, and generates an initial pledge status image.

[0037] Furthermore, based on the pledge object index in the pledge status hierarchy information, the object identifiers in the physical evidence status, corporate supervision status, and bank confirmation status are compared item by item to check whether each status belongs to the same pledge object and the same pledge contract. When the object identifiers are consistent, the business node where each status is located is found along the business occurrence sequence to verify whether the status content can inherit the changes of the previous node, and to determine whether the status falls into a reasonable time sequence position in combination with the formation time. Status fragments with the same content and duplicate sources are merged according to the object index, business node, and formation time, retaining information that can reflect the original source and status formation process, and avoiding the duplicate recording of the same status.

[0038] After merging the states, the temporal sequence and object affiliation relationships between adjacent states are checked. States whose formation time is earlier than the preceding business node and lack supplementary data are marked as time-reversed states. States whose object identifiers, contract affiliations, and business nodes cannot support each other are marked as object mismatch states and excluded from the valid state sequence. The remaining states are marked for validity based on source validity, business continuity, and temporal consistency. The preceding and subsequent acceptance positions are recorded according to the actual formation order of the states, so that the change process of each pledged object in different state layers can be continuously presented, generating an initial pledged state image.

[0039] It should be noted that the initial pledge state mirror is a structured set of states organized according to the pledge object index and business time sequence, consisting of physical evidence state, corporate regulatory state, and bank confirmation state, used to continuously represent the formation, change, and succession relationship of the pledge object state.

[0040] The debt identification module is used to perform time-series alignment and state difference comparison on the initial pledge state mirror, form state propagation gaps, and transmit debt identification and related transfers along the connection relationship between the pledge object and the pledge contract, as well as between the supervision equipment and the ownership certificate, to generate pledged return debt information.

[0041] Based on the initial pledged state mirror, and using the pledged object index and the time of business occurrence as the alignment benchmark, the state fragments in the physical evidence state, corporate supervision state, and bank confirmation state are sequentially organized. State fragments reflecting the same business change are attached to the same time position, and the difference in state update time is corrected to generate pledged state time-series aligned data.

[0042] Furthermore, using the pledged object index in the initial pledged state mirror as the attribution benchmark, state fragments belonging to the same pledged object in the physical evidence state, corporate supervision state, and bank confirmation state are included in the same comparison range; state fragments are organized from earliest to latest according to the time of business occurrence; combining business nodes, state content, and preceding and following positions, it is determined whether state fragments from different sources reflect the same business change; mutually supportive state fragments are attached to a unified time position, and the source mark and original formation time are retained to avoid splitting the same state change into multiple time-series nodes due to differences in collection frequency.

[0043] When a state segment is successfully attached, the actual change position of the pledged state is marked by the time the business occurs, and the position where each information source completes its recording is marked by the time the state is updated. The offset between the time the business occurs and the time the state is updated is compared. For update time differences caused by transmission delays, system supplementation, and bank acknowledgment delays, only the timing of the state segment is adjusted, without changing the original business occurrence order. For state segments with advanced update times and lacking preceding state support, a timing anomaly mark is retained. The corrected state segments are reorganized according to a unified time position, and the state sequence relationship and update time offset are recorded to generate pledged state timing alignment data.

[0044] By comparing the differences in the state content, update time, and order of changes in the pledged state time-series alignment data, the location where the existing state has changed but the associated state has not yet been synchronized is located. At the same time, the location where the state update is stopped and the misalignment is marked, forming a state propagation gap.

[0045] Furthermore, using the unified time position in the pledge status time-series aligned data as the comparison unit, the status fragments in the physical evidence status, corporate supervision status, and bank confirmation status are aggregated according to the pledge object index; the content changes, update time, and direction of each status fragment are compared along the business occurrence sequence; when a certain status has recorded new business changes, while the other statuses are still stuck on the previous content, the time position that has not been synchronized is marked as the difference position; for status fragments with the same content but significantly delayed update time, it is determined whether it is a normal delay or an abnormal status transmission by combining its source mark and original formation time, so as to avoid directly identifying data upload delay as a propagation failure.

[0046] After identifying the discrepancy locations, check the continuity of state changes along the preceding and subsequent acceptance locations of each state segment; if a state has not generated any subsequent updates for a long period after forming new content, mark the state update termination location; if the same business change recorded in different states is attached to inconsistent time nodes and cannot be corrected through the original formation time, mark the acceptance misalignment location; merge the discrepancy locations, state update termination locations, and acceptance misalignment locations according to the pledged object and business time sequence, retain the level of the missing state, the preceding state, and the direction to be synchronized, forming a state propagation gap.

[0047] Based on the state propagation gap, the business attribution of the gap is determined along the connection between the pledged object and the pledge contract, and the scope of evidence involved in the gap is determined along the connection between the regulatory equipment and the ownership certificate. The back-transferred debt value is calculated based on the duration of the state propagation gap, the degree of state difference, and the scope of related transitions, and the back-transferred debt value is marked at the location of the state propagation gap to generate the initial back-transferred debt distribution.

[0048] Furthermore, based on the index of the pledged object where the status propagation gap is located, the source of the gap is traced along the linkage path between the pledged object and the pledge contract; the pledge business node and contract acceptance status at the location where the gap occurs are verified, and the status propagation gap is classified into the corresponding pledge business scope; then, evidence nodes supporting the status change are found along the linkage path between the pledged object and the supervision equipment and ownership certificate, and it is checked whether the evidence nodes have formed a valid status, whether there is an update lag, and whether they can connect the status fragments before and after the gap; evidence nodes that can affect the establishment of the gap and the subsequent feedback content are included in the evidence scope, and the business attribution, source path, and evidence acceptance location are preserved.

[0049] The duration is determined by the interval from the first occurrence of the state propagation gap to the current processing time. The state content before and after the gap and the synchronization of each state layer are compared to determine the degree of state difference. The associated nodes and path levels that the gap can be transmitted to are counted along the linkage path between the pledged object, pledge contract, supervision equipment and ownership certificate to determine the scope of associated migration. The duration, degree of state difference and scope of associated migration are normalized and weighted to calculate the returned debt value. The returned debt value, together with the gap source location and migration direction, is marked to the state propagation gap. Each debt mark is organized according to the pledged object and business time sequence to generate the initial returned debt distribution.

[0050] It should be noted that the connection between the pledged object and the pledge contract is established by linking the pledged object identifier, contract number and business attribution information in the enterprise registration information, binding the specific pledged object with the corresponding pledge contract, and forming an object contract association relationship that reflects the business attribution.

[0051] Linkage between monitoring equipment and ownership certificates: Based on the monitoring equipment identifier, ownership certificate identifier and pledge object index in the on-site monitoring information, an association is established to bind the monitoring equipment collection status with the corresponding ownership certificate information, forming an equipment certificate association relationship that reflects the source of physical monitoring evidence.

[0052] The formula for calculating the value of repatriated debt is: ; in, Indicates the first The backhaul debt value of each state propagation gap. Indicates the first A state propagation gap, Indicates the first The normalized value of the duration of each state propagation gap. Indicates the first Normalized value of the degree of state difference in each state propagation gap. Indicates the first Normalized value of the associated transition range of each state propagation gap. Indicates the first Evidence supporting the closure of the propagation gap in a given state; Indicates the first The transition efficiency coefficient of each state propagation gap.

[0053] The formula adopts a combined structure of "persistent debt base, evidence closure inhibition, and differential transition coupling". It mainly considers that after the state propagation gap is formed, the backpropagation debt will accumulate with the increase of duration and will also be affected by the combined influence of state differences and related diffusion. At the same time, sufficient evidence closure can reduce the need for repeated backpropagation. Using the normalized value of duration as the base term can ensure that the gap that has been out of sync for a long time always retains a certain level of debt. The remaining time space and the degree of evidence incompatibility jointly limit the subsequent increment, which can prevent the backpropagation debt value from being infinitely amplified. The overlapping deduction term between the degree of state difference and the scope of related diffusion can eliminate the repeated inclusion of the same abnormal influence, so that the calculation results can more accurately reflect the persistence, degree of incompatibility and actual diffusion range of the state propagation gap.

[0054] The duration normalized value is obtained by normalizing the interval from the moment the gap first forms to the current processing time, serving as the time base for the backhauled debt; the state difference degree normalized value is determined by the state content before and after the gap and the degree of asynchrony between each state layer, used to characterize the debt increment caused by state differences; the associated migration range normalized value is obtained by normalizing the number of associated nodes and path levels through which the debt migration passes, used to characterize the scope of the gap's influence; the evidence closure support value is determined by the completeness of the evidence nodes connecting the state fragments before and after the gap, used to weaken the debt increment that has already obtained sufficient evidence support; the migration effectiveness coefficient is determined by the proportion of effective migration paths to all associated paths, used to exclude the influence of paths that only have a connection relationship but do not form an actual asynchronous state. All parameters participate in the calculation to obtain the backhauled debt value of the nth state propagation gap.

[0055] Figure 2Using the degree of state difference as the x-axis and the returned debt value as the y-axis, this graph illustrates the distribution of returned debt requirements under different state difference conditions due to state propagation gaps. Hollow circles represent pledged objects that do not require returned debt after checks on object ownership, bank confirmation status, and effective state changes. Intersections represent pledged objects requiring status supplementation, status change, or status confirmation. The horizontal dashed line represents the adaptive returned debt threshold formed based on the sample distribution. Overall, samples with low state difference and returned debt values ​​below the threshold are mainly concentrated in the area where returned debt is not required. As the degree of state difference increases, samples requiring returned debt gradually cluster towards areas with higher returned debt values. The local cross-distribution indicates that the returned debt value is not solely determined by the degree of state difference but is also influenced by the duration of the state propagation gap, the scope of associated transitions, the evidence closure support value, and the transition validity coefficient. This result demonstrates that returned debt calibration can comprehensively distinguish between genuine pending synchronization matters and equipment fluctuations, duplicate data collection, and completed bank confirmation matters, improving the accuracy of screening pledged objects awaiting returned debt.

[0056] Based on the initial distribution of returned debt, the returned debt value is transferred along the associated direction between the pledged object, pledge contract, monitoring equipment and ownership certificate, and the transfer path pointing to the same pending synchronization matter is merged. The source location, propagation direction, scope of influence and returned debt value of the returned debt are retained to generate pledged returned debt information.

[0057] Furthermore, the debt marker position in the initial returned debt distribution is used as the starting point for transfer; the business chain to which the debt belongs is traced along the connection direction between the pledged object and the pledge contract, and the debt impact is transmitted to the contract node that accepts the same pledge status; then the evidence chain is extended along the connection direction between the monitoring equipment and the ownership certificate to determine whether there are any unsynchronized states or incomplete returned matters among the related nodes; debt transfer markers are established for related nodes that can continue the propagation gap of the original state, and the transfer starting point, the nodes passed through, and the direction of transmission are recorded; for related nodes that have not formed a state difference or have completed bank confirmation, the debt transfer is terminated to avoid the boundless diffusion of returned debt.

[0058] Based on the object affiliation, business nodes, and status changes of the items to be synchronized, the various migration paths are compared. Migration paths with different origins but ultimately pointing to the same item to be synchronized are merged, retaining the main path that reflects the cause of debt formation and the propagation process. Debt values ​​that repeatedly pass through the same associated nodes are merged to prevent the same state propagation gap from being counted repeatedly. The scope of debt impact is determined based on the associated levels and unsynchronized node ranges of each path, and the source location, propagation direction, scope of impact, and merged return debt value are linked to the corresponding item to be synchronized. The information is then organized according to the pledge object and business sequence to generate pledged return debt information.

[0059] The backhaul orchestration module is used to locate the object and filter the backhaul scope based on the pledged debt information, obtain the pledged object to be backhauled, and perform state content conversion and related evidence attachment, encapsulate it into a pledged backhaul task, configure idempotency flags and recovery checkpoints, and generate RPA backhaul orchestration information.

[0060] Extract the source location, debt transfer path, and pledge object index from the status propagation gap of pledged debt information. Trace the pledge business to which the status change belongs along the connection between the pledge object and the pledge contract, classify the status propagation gap into the pledge object, and generate gap attribution location information.

[0061] Furthermore, the source location, debt migration path, and pledge object index of each state propagation gap are extracted from the pledged debt information. Starting from the source location, the state fragments before and after the gap are formed are searched to identify the changed state content and the receiving position that has not yet been synchronized. The related nodes passed by the gap are checked along the debt migration path to distinguish between the original gap and the related gap formed by migration. Gaps with the same pledge object index and originating from the same state change are grouped into a unified positioning range, and the gap starting point, migration direction, and position to be synchronized are retained.

[0062] Tracing the contract nodes and business transition links to which status changes belong along the linkage between the pledged object and the pledge contract, verifying the consistency between the pledged object index, contract ownership, and status change content, and eliminating gaps such as contract termination, object release from pledge, and linkage conflicts; linking the status propagation gaps verified through ownership to the corresponding pledged object, and marking its business node and debt transfer path; and generating gap ownership location information by aggregating the gap source location, direction to be synchronized, and contract transition relationship according to the pledged object ownership.

[0063] Based on the location information of the gap, and combined with the transfer status of the returned debt on the regulatory equipment and ownership certificate nodes, the pledged objects that have been confirmed by the bank and have not formed a valid status change are excluded, and the pledged objects that need to be supplemented, changed and confirmed in execution status are retained, and the information of the pledged objects to be returned is generated.

[0064] Furthermore, the index of the pledged object and the status propagation gap in the gap attribution location information are used as the starting point for screening; along the migration path of the back-transmitted debt on the regulatory equipment node and the ownership certificate node, it is checked whether the related nodes still retain the unsynchronized state and valid debt mark; the status fragments formed by the related nodes are compared with the bank's confirmed state to determine whether the relevant state has been received by the bank; the back-transmission scope is terminated for pledged objects whose bank confirmed state has been updated and whose back-transmitted debt has been released; the gap attribution mark is revoked for pledged objects formed by equipment fluctuations, repeated collection and time offset but which have not caused actual status changes.

[0065] The system checks for gaps in the status propagation of pledged objects, the endpoint of debt transfer, and the bank confirmation location. If a missing status has not yet been written into the bank confirmation status, the pledged object is marked as a status supplement object. If there is a difference between the bank confirmation content and the current valid status, the pledged object is marked as a status change object. If the return task has been submitted but there is still a lack of a clear bank processing conclusion, the pledged object is marked as a status confirmation object. The system compiles the object index, gap location, pending status, and scope of related evidence for the pledged objects to be retained, and generates information on pledged objects to be returned.

[0066] Based on the information of the pledged object to be returned, extract the state fragments and supporting evidence on both sides of the state propagation gap, convert the state content according to the meaning of the fields on the return page of the target financial institution, establish the linkage between the converted field content and the supporting evidence, and encapsulate it into a pledge return task.

[0067] Furthermore, using the object index and status propagation gap in the information of the pledged objects to be returned as the processing entry point, status fragments are extracted along the preceding and subsequent receiving positions of the gap, distinguishing between the formed valid status and the target status that still needs to be synchronized; from the status collection information of the pledged objects, regulatory records and voucher records that can support the status changes are searched, and the source of evidence, the time of formation and the object's ownership are verified. Evidence fragments that are unrelated to the status propagation gap, have mismatched formation time and conflicting object identifiers are excluded. The retained evidence is attached to the pledged objects to be returned according to the status change position, forming a correlation between the status content and the supporting evidence.

[0068] Establish a status content conversion relationship based on the field names, input formats, and business meanings in the target financial institution's feedback page. Convert the enterprise-side status description into field content that the bank's page can recognize, while retaining the status meaning and source location before and after conversion. Check whether the converted field content can fully reflect the status propagation gap, attach supporting evidence to the field positions that require evidence explanation, and set supplementary markers for content with missing fields, insufficient evidence, and inconsistent status meanings. Encapsulate the checked field content, supporting evidence, pledge object index, and status propagation gap into a task to generate a pledge feedback task.

[0069] It should be noted that the pledge return task is a structured execution unit that encapsulates the pledge status to be synchronized, the field content of the bank page, supporting evidence, the pledge object index, and the status propagation gap, and is used to drive RPA to perform field writing and return submission.

[0070] The pledge return tasks are arranged in order according to the formation sequence of state propagation gaps and the return debt transition path. Idempotent flags are configured in combination with pledge object index, state update time and task content. Recovery checkpoints are set at field write and return submission nodes to generate RPA return orchestration information.

[0071] Furthermore, tasks are arranged according to the formation time of the state propagation gap associated with the pledge return task. First, the succession relationship of each task under the same pledge object is determined, and then the influence direction between tasks is checked along the debt transfer path. Tasks originating from the same state change and whose return content has a succession relationship are maintained in their original order. Tasks pointing to the same pending synchronization matters are merged. Tasks with conflicting state content are temporarily suspended from arrangement, while retaining the conflicting position and source task, thus forming a task execution queue that can reflect the order of state changes and debt propagation.

[0072] The idempotency flag is generated by combining the staking object index, status update time, and task content characteristics, so that the same staking status can be identified as the same return task when triggered repeatedly. Recovery checkpoints are set at the field write completion position and the return submission completion position, respectively, to record the range of completed fields, the page dwell position, and the submission status. When the task is interrupted, the idempotency flag is used to determine whether the submission has been completed, and the incomplete content is processed from the most recent recovery checkpoint. The task execution queue, idempotency flag, and recovery checkpoint are aggregated to generate RPA return orchestration information.

[0073] The field return module is used to perform page positioning, field writing, and return submission for RPA return orchestration information. It performs field matching for page controls whose positions have changed. In the event of network timeout or task interruption, it performs incomplete field writing and return submission based on the idempotency flag and recovery checkpoint, and obtains the RPA return execution trajectory and current bank receipt information.

[0074] Based on the RPA backhaul orchestration information, the RPA process is initiated according to the execution order of the pledge backhaul tasks. The page path and field content to be written of the target financial institution are extracted, and the pledge business page is entered. The field filling position is locked according to the page level, field name and control attribute, and page positioning data is generated.

[0075] Furthermore, using the task execution queue in the RPA backhaul orchestration information as the starting basis, the pledge backhaul tasks are retrieved one by one according to the orchestration order, and the access portal of the target financial institution is called to complete identity authentication and session establishment; the page path and the content of the fields to be written are extracted from the pledge backhaul tasks, and the pledge business function area is entered sequentially along the page path. It is checked whether the current page identifier is consistent with the business area specified by the task. The execution of tasks with page jump interruption, login failure, and business area mismatch is suspended and the current access position is retained to form a page access state that can be used for field positioning.

[0076] When the page access status is confirmed to be valid, the scope of control retrieval is narrowed down layer by layer according to the page hierarchy. The business area where the field to be written is located is determined by the field name, and the input box, selection item and attachment entry are distinguished by the control attributes and the text of the adjacent page. For fields with the same name or controls whose positions have changed, the hierarchy of the control, the input content and the meaning of the task field are compared to eliminate candidate controls that do not meet the filling requirements. The content of the field to be written is bound to the control that has passed the filtering, and the page position, control identifier and field affiliation are recorded to generate page positioning data.

[0077] Based on the page location data, the field content in the staking return task is written into the locked page control. The page display content is compared with the field content to be written. For page controls whose positions have changed, the field name, input format and adjacent tags are matched to determine the available controls and complete the return submission, generating RPA task submission information.

[0078] Furthermore, the field filling positions in the page positioning data are bound to the fields to be written in the pledge return task. The field content is written to the locked page controls according to the task execution order. After each field writing is completed, the page echo content is collected and compared with the original field content to be written for character consistency and format consistency. The page is checked for truncation, automatic replacement, and input failure. For fields whose echo content is consistent with the field content to be written, the writing completion status is recorded. For fields with differences, the current input is cleared and rewritten. If differences still exist after continuous writing, the current task is paused and the abnormal field position is retained to form the field writing status.

[0079] If the position, level, or attribute of the original locked control changes during the field writing process, the control is re-searched within the current business area. Candidate controls are filtered using field name, allowed input format, and adjacent label text. The business meaning of the candidate controls is matched with the content of the field to be written, excluding controls with similar names but different uses. The matched controls are updated as available controls and the field writing continues. After all fields have been displayed and compared, the writing status of required content and attached evidence is checked, triggering a submission feedback and collecting the acceptance prompt, submission time, and idempotency indicator returned by the page, generating RPA task submission information.

[0080] Figure 3The performance of field feedback under continuously changing bank collateral business page conditions is compared between the baseline group and the complete solution group, using the page control position change rate as the x-axis and the page control positioning accuracy rate and field correct writing rate as the y-axis. Solid dotted lines represent the page control positioning accuracy rate of the baseline group, dashed square lines represent the field correct writing rate of the baseline group, dashed triangular lines represent the page control positioning accuracy rate of the complete solution group, and dotted diamond lines represent the field correct writing rate of the complete solution group. As the page control position change rate increases from zero to 50%, the positioning accuracy rate and field correct writing rate of the baseline group continuously decrease, while the two indicators of the complete solution group remain at a high level. This result indicates that the present invention, by locking the field filling position based on page hierarchy, field name, and control attributes, and combining input format, adjacent labels, and page echo content to re-match page controls with changing positions, can avoid field errors caused by fixed position positioning failures, ensuring that the field content in the collateral feedback task is accurately written to the target financial institution's page.

[0081] It should be noted that the baseline group used a fixed page path and control position to complete field positioning and writing, and did not introduce multi-feature matching and page echo comparison after position changes. Therefore, the positioning accuracy and field correct writing rate decreased significantly when the page control changed.

[0082] The complete solution group uses the page hierarchy, field name, control attribute, input format and adjacent tag joint matching of the present invention, and performs verification in combination with the page echo content, so it has a stronger adaptability to changes in the position of page controls.

[0083] Based on RPA task submission information, the execution nodes of page positioning, field writing, and return submission are recorded. When network timeout or task interruption occurs, the idempotency flag is used to determine the task submission status, and the incomplete field writing and return submission are continued to be executed according to the recovery checkpoint, generating the RPA return execution trajectory.

[0084] Furthermore, using the task identifier and submission status in the RPA task submission information as the entry point for execution records, the operation nodes formed by page location, field writing, and return submission are continuously recorded according to the page access order. The execution time, page position, and completion status are written for each operation node. When a field is successfully written, the range of completed fields is updated. When a return submission is triggered, the submission entry and page feedback are recorded, so that the operation process of the same staking return task is continuously connected in chronological order, forming a node sequence that can reflect the task execution progress.

[0085] When network timeout or task interruption occurs, extract the idempotency flag of the current staking back transmission task and query existing submission records to determine whether the bank page has accepted the current task. If it has been accepted, stop repeated submissions and retain the acceptance node. If it has not been accepted, locate the most recent recovery checkpoint, check the field writing status completed before the checkpoint, and continue to execute the incomplete field writing and back transmission submission after the checkpoint. Gather the normal execution nodes, the location of the exception, the recovery processing path, and the final submission status in chronological order to generate the RPA back transmission execution trajectory.

[0086] Based on the RPA execution trajectory, extract the task acceptance information and processing feedback returned by the bank's page, bind the processing feedback to the pledge return task according to the idempotency identifier, organize the task submission nodes and bank processing content, and generate the current bank receipt information.

[0087] Furthermore, the final submission node in the RPA backhaul execution trajectory is used as the entry point for receipt retrieval. This allows entry into the task acceptance area and processing feedback area of ​​the bank's page, collecting the acceptance identifier, processing status, and feedback time returned by the page. The idempotency identifier carried in the page feedback is compared with the idempotency identifier in the pledge backhaul task. If the identifiers match, the processing feedback is attached to the pledge backhaul task. If the identifiers are missing or conflict, the feedback source and page position are preserved to prevent the bank feedback from being incorrectly assigned to other backhaul tasks.

[0088] The time sequence between page location, field writing, submission, and bank feedback is organized along the RPA backhaul execution trajectory, connecting task submission nodes and bank processing content to the same task chain; duplicate bank feedback is merged according to the time of formation, retaining the latest processing content and previous change records, and tasks that have not yet reached a clear processing conclusion are kept in a pending confirmation status; the pledge backhaul task, idempotency identifier, submission node, and bank processing content are collected to generate the current bank receipt information.

[0089] The status update module is used to match the RPA back-transmission execution trajectory and the current bank receipt information according to the idempotency identifier, extract the processing status of the pledge back-transmission task, update the successful reception content to the bank confirmation status and release the corresponding back-transmission debt, attach the return reason to the original pledge back-transmission task to form the content to be corrected, mark the tasks in processing as pending confirmation, and generate the bank-enterprise pledge data back-transmission status.

[0090] Extract the idempotency identifier from the RPA back-transmission execution trajectory and the current bank receipt information, attach the current bank receipt information to the pledge back-transmission task with the same idempotency identifier, and organize the execution sequence according to the task submission time and receipt formation time to generate back-transmission task receipt association data.

[0091] Furthermore, the idempotency identifiers carried by the pledge return task are extracted from the RPA return execution trajectory, and the idempotency identifiers carried by the bank processing records are extracted from the current bank receipt information. The format uniformity and integrity of the identifiers on both sides are checked. Taking the pledge return task as the search starting point, the bank processing records with the same identifiers are searched in the current bank receipt information, and the successfully matched bank processing records are attached to the original pledge return task. The source location and abnormal markers are retained for bank processing records with missing identifiers, conflicting identifiers, and duplicate identifiers to avoid the mixing of bank processing content from different tasks.

[0092] Organize the time sequence according to the submission time of the pledge return task and the formation time of the bank processing record, check whether the bank processing record is later than the task submission node, and connect the bank processing records returned consecutively for the same task according to the order of formation; merge the content of duplicate feedback, retain the latest processing content and previous status changes, and suspend the inclusion of bank processing records whose formation time is earlier than the task submission time or lack submission node support; collect the pledge return task, bank processing record and time sequence relationship, and generate return task receipt association data.

[0093] The bank acceptance feedback and processing progress in the data associated with the return task receipts are analyzed, and the pledge return tasks are divided into successfully received tasks, returned tasks, and tasks in process. The receipt content and original execution position of each task are retained, and the processing status information of the return task is generated.

[0094] Furthermore, taking the pledge return task in the data associated with the return task receipt as the processing unit, the status marker, feedback content and formation time in the bank's acceptance feedback are extracted, and the processing progress is organized along the time sequence of the same task; the bank's feedback content is associated with the task submission node to check whether the bank has completed data reception, whether it has returned a return prompt, and whether the processing status is still ongoing. Duplicate feedback is merged according to the formation order, retaining the latest content that can reflect the current processing stage, while recording the changes in the previous processing to avoid the same task forming a duplicate status due to multiple feedbacks.

[0095] Staking return tasks that have returned a receipt completion mark and whose submitted content has not been returned are classified as successful receipt tasks; staking return tasks whose feedback content includes a return prompt and indicates the failure location are classified as return tasks; and staking return tasks that have been accepted but have not yet reached a final processing conclusion are classified as processing tasks. The task type is associated with the bank receipt content, the task submission node, and the original RPA execution location, and the status classification basis and processing time are retained to generate return task processing status information.

[0096] Based on the status information of the backhaul task, the pledge status in the successfully received task is written into the bank confirmation status of the initial pledge status mirror, the backhaul debt mark on the status propagation gap is removed, and the return reason is extracted from the receipt content of the returned task. The return reason is attached to the original pledge backhaul task to generate backhaul task processing data.

[0097] Furthermore, taking the successful reception task in the backhaul task processing status information as the update object, the pledge object and status propagation gap in the initial pledge status mirror are located according to the idempotency identifier. The pledge status that the bank has confirmed to receive is attached to the bank confirmation status of the same pledge object, and the bank receipt formation time and task submission position are retained. The updated bank confirmation status is compared with the content to be synchronized in the status propagation gap. When the received content can cover the status to be synchronized, the backhaul debt mark on the gap position is removed. When the received content only covers part of the status, the backhaul debt of the unsynchronized part is retained.

[0098] For returned tasks in the status information of returned task processing, the field location and failure reason pointed to by the return prompt are identified from the bank receipt content. The original pledge return task is traced according to the idempotency identifier, and the return reason is attached to the field content or supporting evidence location in the original task. The submission content and execution trajectory of the original task are retained, the content that needs to be modified and the scope of re-return are marked, and the status update records of successfully received tasks, the debt release records of returned tasks and the reason attachment records of returned tasks are collected to generate return task processing data.

[0099] The system summarizes the fields to be modified and the evidence to be supplemented in the data of the returned task processing to form the content to be corrected. For tasks in progress, the original returned debt is retained and a receipt confirmation mark is configured. The system collects the task processing status, debt release status and content to be corrected, and generates the data return status of bank-enterprise pledge data.

[0100] Furthermore, the returned tasks in the data processing of returned tasks are decomposed into content. The original submission content is found along the field position and supporting evidence position associated with the reason for return. The pending items are distinguished by field format mismatch, missing status expression and insufficient evidence support. Content that can be corrected by adjusting the field is marked as fields to be modified, and content that needs to be re-collected or supplemented is marked as evidence to be supplemented. The data is then merged according to the pledge object index and the original pledge return task, retaining the return position, correction scope and re-return entry point to form the content to be corrected.

[0101] For tasks in the processing status information, check the bank's acceptance progress for tasks in progress, retain the original returned debts at the status propagation gap, and avoid prematurely removing the debt mark before the bank has formed a final feedback; configure a receipt confirmation mark on pledged return tasks, and record the most recent feedback time and subsequent query position; collect the debt release status of successfully received tasks, the content to be corrected for returned tasks, and the confirmation status of tasks in progress according to the pledge object and task identifier, and generate a bank-enterprise pledge data return status that can reflect the current processing progress of each pledged return task.

[0102] Figure 4 The overall performance of the benchmark group and the complete solution group was compared in six aspects: recall rate of state propagation gaps, accuracy rate of screening collateral objects to be returned, correct field writing rate, success rate of anomaly recovery, accuracy rate of bank receipt association, and completeness rate of processing trajectory. The dotted lines represent the benchmark group, and the solid lines represent the complete solution group using the complete technical process of this invention. The complete solution group maintained a high and relatively balanced level of indicators across all evaluation dimensions, while the benchmark group showed significant contraction in screening collateral objects to be returned, field writing, bank receipt association, and processing trajectory recording. These results indicate that initial collateral status mirroring, collateral status time-series aligned data, and debt transition can improve the ability to locate state propagation gaps and screen collateral objects to be returned; page control field matching, idempotency identification, and recovery checkpoints can improve the stability of RPA anomaly continuation and field return; and bank receipt association and processing status updates can maintain the continuous correspondence between task submission, bank feedback, and debt release processes, thereby reducing the overall risk of missed, incorrect, and duplicate returns.

[0103] It should be noted that the need for back transmission refers to the situation where the pledge status corresponding to the status propagation gap has not yet been written into the bank's confirmed status, and there are still valid pending synchronization matters after the debt is back-transmitted. It is necessary to generate the pledge object information to be back-transmitted and encapsulate it as a pledge back transmission task.

[0104] No need for backhaul means that the relevant pledge status has been confirmed by the bank, or the status has only changed due to equipment fluctuations, repeated data collection, and time offsets without any actual change, so no pledge backhaul task will be generated.

[0105] In summary, this invention achieves the location of differences between the edge IoT status and the bank / enterprise status through: constructing a pledge status mirror, aligning the time sequence, and transmitting debt back. This is used to screen the objects to be transmitted and the scope of evidence, reducing the risks of missed, incorrect, and duplicate transmissions. Through pledge task orchestration, idempotent control, recovery checkpoints, and bank receipt association, it achieves RPA transmission anomaly continuation and processing status updates, ensuring the accurate transmission of pledge data and improving transmission stability and traceability.

[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A smart data feedback system for bank-enterprise pledged loans based on RPA, characterized in that, include: The data acquisition module is used to collect on-site supervision information, enterprise registration information and historical bank receipt information of the pledged object, perform status calibration according to object ownership and business time sequence, form physical evidence status, enterprise supervision status and bank confirmation status, and perform status validity verification and time inheritance mark to generate an initial pledge status image. The debt identification module is used to perform time-series alignment and state difference comparison on the initial pledge state mirror, form state propagation gaps, and transmit debt identification and related transfers along the connection relationship between the pledge object and the pledge contract, as well as between the supervision equipment and the ownership certificate, to generate pledge return debt information. The backhaul orchestration module is used to locate the object and filter the backhaul range based on the pledged debt information, obtain the pledged object to be backhauled, and perform state content conversion and related evidence attachment, encapsulate it into a pledged backhaul task, configure idempotency flags and recovery checkpoints, and generate RPA backhaul orchestration information. The field return module is used to perform page positioning, field writing and return submission on RPA return orchestration information, perform field matching on page controls with changed positions, and perform incomplete field writing and return submission based on idempotency flags and recovery checkpoints in the event of network timeout or task interruption, and obtain RPA return execution trajectory and current bank receipt information. The status update module is used to match the RPA back-transmission execution trajectory and the current bank receipt information according to the idempotency identifier, extract the processing status of the pledge back-transmission task, update the successful reception content to the bank confirmation status and release the corresponding back-transmission debt, attach the return reason to the original pledge back-transmission task to form the content to be corrected, mark the tasks in processing as pending confirmation, and generate the bank-enterprise pledge data back-transmission status.

2. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 1, characterized in that, The steps for generating the initial staking state image are as follows: The system performs object identification parsing on-site supervision information, enterprise registration information, and historical bank receipt information of the pledged object. Based on the pledged object identifier in the enterprise registration information and the ownership of the pledge contract, the system links the on-site supervision information and the supervision equipment identifier and ownership certificate identifier in the enterprise registration information to the pledged object index, collects information fragments pointing to the same pledged object, and generates pledged object status collection information. Based on the information collected on the status of pledged objects, the evidence source of on-site supervision information is confirmed to form a physical evidence status, the registration information of enterprises is assigned to registration items to form an enterprise supervision status, the historical bank receipt information is returned to the corresponding task to form a bank confirmation status, and the chronological relationship of each status is corrected according to the time of business occurrence and the time of status update to generate pledge status hierarchical information. The system performs object identifier consistency comparison on the physical evidence state, enterprise supervision state, and bank confirmation state in the pledge status layer information, verifies the business nodes and formation time corresponding to each state, merges state fragments with duplicate content, excludes state fragments with time inversion and object mismatch, marks the validity of the retained state and marks the preceding and following positions, and generates an initial pledge status image.

3. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 2, characterized in that, The steps for forming the state propagation gap are as follows: Based on the initial pledged state mirror, and using the pledged object index and the time of business occurrence as the alignment benchmark, the state fragments in the physical evidence state, corporate supervision state, and bank confirmation state are sequentially organized, and the state fragments reflecting the same business change are attached to the same time position, and the difference in state update time is corrected to generate pledged state time-series aligned data. By comparing the differences in the state content, update time, and order of changes in the pledged state time-series alignment data, the location where the existing state has changed but the associated state has not yet been synchronized is located. At the same time, the location where the state update is stopped and the misalignment is marked, forming a state propagation gap.

4. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 1 or 3, characterized in that, The steps for generating pledged debt information are as follows: Based on the state propagation gap, the business attribution of the gap is determined along the connection between the pledged object and the pledge contract, the scope of evidence involved in the gap is determined along the connection between the regulatory equipment and the ownership certificate, the back-transferred debt value is calculated according to the duration of the state propagation gap, the degree of state difference and the scope of related transfer, and the back-transferred debt value is marked to the location of the state propagation gap to generate the initial back-transferred debt distribution; Based on the initial distribution of returned debt, the returned debt value is transferred along the associated direction between the pledged object, pledge contract, monitoring equipment and ownership certificate, and the transfer path pointing to the same pending synchronization matter is merged. The source location, propagation direction, scope of influence and returned debt value of the returned debt are retained to generate pledged returned debt information.

5. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 4, characterized in that, The steps for obtaining the pledged object to be returned are as follows: Extract the source location, debt transfer path and pledge object index from the status propagation gap of pledged debt information, trace the pledge business to which the status change belongs along the connection between the pledge object and the pledge contract, classify the status propagation gap into the pledge object, and generate gap attribution location information. Based on the location information of the gap, and combined with the transfer status of the returned debt on the regulatory equipment and ownership certificate nodes, the pledged objects that have been confirmed by the bank and have not formed a valid status change are excluded, and the pledged objects that need to be supplemented, changed and confirmed in execution status are retained, and the information of the pledged objects to be returned is generated.

6. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 5, characterized in that, The steps for generating RPA return orchestration information are as follows: Based on the information of the pledged object to be returned, extract the state fragments and supporting evidence on both sides of the state propagation gap, convert the state content according to the meaning of the fields on the return page of the target financial institution, establish the linkage between the converted field content and the supporting evidence, and encapsulate it into a pledge return task. The pledge return tasks are arranged in order according to the formation sequence of state propagation gaps and the return debt transition path. Idempotent flags are configured in combination with pledge object index, state update time and task content. Recovery checkpoints are set at field write and return submission nodes to generate RPA return orchestration information.

7. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 6, characterized in that, The steps for performing page positioning, field writing, and postback submission are as follows: Based on the RPA back-to-archive information, the RPA process is started according to the execution order of the pledge back-to-archive task. The page path and field content to be written of the target financial institution are extracted, and the pledge business page is entered. The field filling position is locked according to the page level, field name and control attribute, and page positioning data is generated. Based on the page location data, the field content in the staking return task is written into the locked page control. The page display content is compared with the field content to be written. For page controls whose positions have changed, the field name, input format and adjacent tags are matched to determine the available controls and complete the return submission, generating RPA task submission information.

8. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 1, characterized in that, The steps for obtaining the RPA execution trajectory and current bank receipt information are as follows: Based on RPA task submission information, the execution nodes of page positioning, field writing and return submission are recorded. When network timeout and task interruption occur, the idempotency flag is used to determine the task submission status, and the unfinished field writing and return submission are continued to be executed according to the recovery checkpoint, generating the RPA return execution trajectory. Based on the RPA execution trajectory, extract the task acceptance information and processing feedback returned by the bank's page, bind the processing feedback to the pledge return task according to the idempotency identifier, organize the task submission nodes and bank processing content, and generate the current bank receipt information.

9. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 8, characterized in that, The processing status of the task to extract the pledge return is handled as follows: Extract the idempotency identifier from the RPA back-transmission execution trajectory and the current bank receipt information, attach the current bank receipt information to the pledge back-transmission task with the same idempotency identifier, and organize the execution sequence according to the task submission time and receipt formation time to generate back-transmission task receipt association data. The bank acceptance feedback and processing progress in the data associated with the return task receipts are analyzed, and the pledge return tasks are divided into successfully received tasks, returned tasks, and tasks in process. The receipt content and original execution position of each task are retained, and the processing status information of the return task is generated.

10. The intelligent data feedback system for bank-enterprise pledge based on RPA as described in claim 9, characterized in that, The steps for generating the bank-enterprise pledge data feedback status are as follows: Based on the status information of the back-transmission task, the pledge status in the successfully received task is written into the bank confirmation status of the initial pledge status mirror, the back-transmission debt mark on the status propagation gap is removed, and the return reason is extracted from the receipt content of the returned task. The return reason is attached to the original pledge back-transmission task to generate back-transmission task processing data. The system summarizes the fields to be modified and the evidence to be supplemented in the data of the returned task processing to form the content to be corrected. For tasks in progress, the original returned debt is retained and a receipt confirmation mark is configured. The system collects the task processing status, debt release status and content to be corrected, and generates the data return status of bank-enterprise pledge data.

Images