Multi-terminal distributed data cooperative acquisition and synchronization method and system

Abstract

The application discloses a kind of multi-terminal distributed data cooperative collection and synchronization method and system, it is related to electric digital data processing technical field, comprising: first acquisition node responds to the collection completion event of target data object, generates the collection trace of target data object;According to the structured data in collection trace, calculate predicted moving path, push collection trace to the second acquisition node on predicted moving path;Second acquisition node responds to the arrival event of target data object, reads collection trace from local cache, and performs differential data collection operation based on collection trace to target data object;Second acquisition node updates predicted moving path and collection trace after completing differential data collection operation.This application is used to solve the problems of lack of real-time data exchange and cooperative mechanism of collection behavior between each acquisition node in prior art, and lack of active prediction mechanism for collection process.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing technology, and more specifically, to a method and system for multi-terminal distributed data collaborative acquisition and synchronization. Background Technology

[0002] With the widespread deployment of distributed multi-node data acquisition systems, each acquisition node is typically deployed in different physical locations, undertaking data acquisition tasks across different dimensions or stages. In actual data flow, the same target data object usually needs to pass through multiple acquisition nodes sequentially, with each node collecting different attribute fields. Therefore, how to achieve real-time sharing and collaborative acquisition of target data object information among multiple acquisition nodes, avoiding redundant acquisition and data silos, has become a key issue in improving the processing efficiency and resource utilization efficiency of distributed acquisition systems.

[0003] For example, Chinese invention patent application document with publication number CN112000826B discloses a method and system for personnel management. Its technical solution uses a personnel management server, a face recognition algorithm server, a client, and a personnel information collection server to uniformly aggregate personnel information collected by various immigration management agencies into a central database, so as to realize functions such as early warning of persons involved in cases and identity verification and confirmation. The technical logic of collaborative collection in this solution is that each terminal collects independently and then uploads and aggregates the data afterward. That is, after each collection node completes its collection work independently, it uploads the data to the central server for unified archiving and integration.

[0004] However, the above method has the following shortcomings: (1) Each collection node is isolated from each other at the moment of collection, and lacks a real-time data exchange and collaboration mechanism, which may result in the same target data object being repeatedly collected in the same information dimension on different terminals, resulting in resource waste and reduced customs clearance efficiency; (2) There is a lack of proactive prediction mechanism for the collection process, which cannot proactively push the relevant information of the target data object that has been collected to the terminal that is about to perform collection before the collection behavior occurs. The terminal that is about to perform collection can only passively wait for the target data object to arrive before collecting or querying, which affects the efficiency of information collaborative collection between terminals.

[0005] To address the aforementioned shortcomings, there is an urgent need for a distributed data collaborative acquisition and synchronization method that can achieve real-time data exchange based on multiple acquisition nodes and has an active prediction mechanism oriented towards the acquisition process. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a multi-terminal distributed data collaborative acquisition and synchronization method and system to solve the problems of the lack of real-time data exchange and collaboration mechanisms between acquisition nodes and the lack of proactive prediction mechanisms for the acquisition process in the prior art.

[0007] To achieve the above objectives, this application provides the following technical solution: The first aspect of this invention provides a multi-terminal distributed data collaborative acquisition and synchronization method, comprising: a first acquisition node responding to a acquisition completion event of a target data object, generating an acquisition trace of the target data object, wherein the acquisition trace is a structured data packet containing an acquisition status label, a completeness vector identifier, and a data version identifier; analyzing candidate paths of the target data object based on the structured data in the acquisition trace, calculating a predicted movement path based on the candidate paths, pushing the acquisition trace to a second acquisition node on the predicted movement path, and writing the acquisition trace to the local cache of the second acquisition node; the second acquisition node responding to an arrival event of the target data object, reading the acquisition trace from the local cache, and performing a differentiated data acquisition merging operation on the target data object according to a preset data acquisition synchronization mode based on the completeness vector identifier; after completing the differentiated data acquisition merging operation, the second acquisition node updates the predicted movement path and the acquisition status label and data version identifier in the acquisition trace, and pushes the acquisition trace to the acquisition node on the updated predicted movement path.

[0008] In a preferred embodiment, the data fields of the structured data packet include: an identity token field for storing an anonymous identity identifier of the target data object, wherein the identity token is generated by one-way hash encryption of the identity identifier of the target data object; a context metadata field group for storing collection timestamp, node location, and data flow direction identifier; a collection status label for storing the collection status label of the target data object during the collection process, wherein the collection status label includes normal status, data doubtful status, key verification status, and supplementary collection status; a completeness vector identifier for storing the completeness vector identifier of the information dimensions of the target data object that have been collected; a version number field for storing a data version identifier, wherein the data version identifier increments with each collection trace update; and a signature information field for storing a digital signature generated by the private key of the first collection node.

[0009] In a preferred embodiment, the candidate path analysis of the target data object based on the structured data in the acquisition traces includes: obtaining the node positions and data flow direction identifiers in the acquisition traces, and loading a pre-set directed terminal topology graph of the distributed system with the node position of the first acquisition node as the root node and the data flow direction identifier as the search direction constraint; obtaining candidate paths in the directed terminal topology graph based on a preset path node hop count, with the node position of the first acquisition node as the starting point and each exit node in the directed terminal topology graph as the ending point, and constructing a candidate path set.

[0010] In a preferred embodiment, calculating the predicted mobile path based on candidate paths includes: reading historical collection trajectory records within a preset statistical period from a historical database, performing hierarchical statistics according to data object type, data flow direction, and transmission time period, calculating the pass frequency ratio of each candidate path under each hierarchical condition, and obtaining the prior probability weight of each candidate path; obtaining the real-time load parameters of each collection node in the candidate path, the real-time load parameters including the current queue number and average processing time; calculating the standardized load index of each collection node according to the real-time load parameters, and calculating the path weight of the candidate path according to the prior probability weight of the candidate path and the standardized load index of each collection node in the candidate path; normalizing the path weight of the candidate path, selecting the candidate path with the highest normalized path weight as the predicted mobile path, and generating an ordered terminal sequence according to the node position order of the collection nodes on the predicted mobile path.

[0011] In a preferred embodiment, pushing the acquisition traces to a second acquisition node on the predicted movement path includes: obtaining the online status of each acquisition node in the ordered terminal sequence; removing acquisition nodes with online status of offline or faulty from the ordered terminal sequence, and reconnecting the subsequent terminals of the removed terminals to the previous terminals to generate an updated ordered terminal sequence; serializing the acquisition traces into transmission data packets, and adaptively selecting a transmission protocol from a preset protocol set according to the network environment between each acquisition node; introducing an acknowledgment mechanism with timeout retransmission, and sending the transmission data packets to the second acquisition node in the updated ordered terminal sequence according to the acknowledgment mechanism and the transmission protocol.

[0012] In a preferred embodiment, writing the acquisition trace to the local cache of the second acquisition node includes: after the second acquisition node receives and verifies the transmission data packet of the acquisition trace, it writes the acquisition trace as pre-synchronization data to the local cache of the second acquisition node; based on the integrity vector identifier and acquisition status label in the acquisition trace, the data acquisition synchronization mode of the second acquisition node is pre-determined before the target data object arrives, the data acquisition synchronization mode includes full skip mode, incremental supplementary acquisition mode and deep verification mode.

[0013] In a preferred embodiment, the data acquisition synchronization mode of the second acquisition node is predetermined before the target data object arrives, including: reading the integrity vector identifier in the acquisition trace; if the integrity vector identifier indicates that all preset information dimensions of the target data object have been acquired and the acquisition status label is normal, then the data acquisition synchronization mode is determined to be the full skip mode; if the integrity vector indicates that at least one preset information dimension has not been acquired and the acquisition status label is normal, then the data acquisition synchronization mode is determined to be the incremental supplementary acquisition mode; if the acquisition status label is either the data doubtful state or the key inspection state, then the data acquisition synchronization mode is determined to be the deep verification mode.

[0014] In a preferred embodiment, a differentiated data acquisition and merging operation is performed on the target data object based on the integrity vector identifier according to a preset data acquisition synchronization mode. This includes: the second acquisition node acquiring the on-site identity data of the target data object, generating an on-site identity token through one-way hash encryption, and performing a key-value query in the local cache using the on-site identity token as the retrieval key; if the query is successful, the corresponding pre-synchronized data is read from the local cache, the integrity vector identifier and acquisition status label of the target data object are acquired, and the data acquisition synchronization mode is determined; if the query is unsuccessful, a real-time query request is sent to the central server to acquire the most recent acquisition trace of the target data object, and the most recent acquisition trace is written into the local cache before performing the differentiated data acquisition and merging operation.

[0015] In a preferred embodiment, updating the predicted movement path and the acquisition status label and data version identifier in the acquisition trace includes: after the second acquisition node completes the differentiated data acquisition and merging operation, it merges the newly acquired data fields with the pre-synchronized data in the local cache at the field level, updates the acquisition timestamp, node location information, integrity vector identifier, and acquisition status label in the acquisition trace, and increments the data version identifier in the acquisition trace; compares the updated acquisition trace with the preset data synchronization termination condition; if the updated acquisition trace does not meet the preset data synchronization termination condition, the predicted movement path is recalculated starting from the node location of the second acquisition node; if the updated acquisition trace meets the preset data synchronization termination condition, the acquisition trace is marked as terminated, and the update of the acquisition trace is stopped.

[0016] A second aspect of the present invention provides a multi-terminal distributed data collaborative acquisition and synchronization system, comprising: a first acquisition module, wherein a first acquisition node, in response to a acquisition completion event of a target data object, generates an acquisition trace of the target data object, the acquisition trace being a structured data packet containing an acquisition status label, a completeness vector identifier, and a data version identifier; a path prediction module, configured to analyze candidate paths of the target data object based on the structured data in the acquisition trace, calculate a predicted movement path based on the candidate paths, push the acquisition trace to a second acquisition node on the predicted movement path, and write the acquisition trace to the local cache of the second acquisition node; a second acquisition module, wherein the second acquisition node, in response to an arrival event of the target data object, reads the acquisition trace from the local cache, performs a differentiated data acquisition and merging operation on the target data object according to a preset data acquisition synchronization mode based on the completeness vector identifier; and an information update module, configured to update the predicted movement path and the acquisition status label and data version identifier in the acquisition trace after the second acquisition node completes the differentiated data acquisition and merging operation, and push the acquisition trace to the acquisition node on the updated predicted movement path.

[0017] The beneficial effects of this invention are: 1. After the first acquisition node completes the information acquisition of the target data object, the present invention constructs the acquisition trace of the target data object based on the first acquisition node. Through the real-time generation and push of the acquisition trace, each acquisition node can share the acquisition status information of the target data object at the moment the acquisition behavior occurs, realizing real-time collaboration between data acquired by multiple terminals.

[0018] 2. Based on the structured data analysis in the acquisition traces, candidate paths for the target data object are analyzed, and a predicted movement path is calculated based on the candidate paths. The acquisition traces are then pushed to the second acquisition node on the predicted movement path, and the acquisition traces are written to the local cache of the second acquisition node. In response to the arrival event of the target data object, the second acquisition node reads the acquisition traces from the local cache and performs differentiated data acquisition and merging operations on the target data object according to the preset data acquisition synchronization mode based on the integrity vector identifier. Through the predicted movement path driven by the acquisition traces, dynamic and on-demand information pre-synchronization between acquisition nodes is realized. Through the differentiated data acquisition and merging operation, relevant information is proactively pushed to the terminal that may be about to perform acquisition before the acquisition behavior occurs, thereby improving information acquisition efficiency. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating a multi-terminal distributed data collaborative acquisition and synchronization method. Figure 2 This is a schematic diagram of the method for calculating and predicting mobile paths based on candidate paths provided in this embodiment; Figure 3 This is a flowchart illustrating the method for pushing the collected traces to the second collection node on the predicted movement path provided in this embodiment. Figure 4 This is a schematic diagram of the method for writing acquisition traces to the local cache of the second acquisition node provided in this embodiment; Figure 5 This is a flowchart illustrating the method for performing differentiated data collection and merging operations on target data objects based on integrity vector identifiers according to a preset data collection synchronization mode, as provided in this embodiment. Figure 6 This is a schematic diagram of a multi-terminal distributed data collaborative acquisition and synchronization system. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1, please refer to Figure 1 , Figure 1 This invention provides a multi-terminal distributed data collaborative acquisition and synchronization method, comprising the following steps: A first acquisition node, in response to a target data object acquisition completion event, generates an acquisition trace of the target data object. The acquisition trace is a structured data packet containing an acquisition status label, a completeness vector identifier, and a data version identifier. Based on the structured data in the acquisition trace, candidate paths for the target data object are analyzed, and a predicted movement path is calculated based on the candidate paths. The acquisition trace is then pushed to a second acquisition node on the predicted movement path, and the acquisition trace is written to the local cache of the second acquisition node. In response to a target data object arrival event, the second acquisition node reads the acquisition trace from its local cache and performs a differentiated data acquisition and merging operation on the target data object according to a preset data acquisition synchronization mode based on the completeness vector identifier. After completing the differentiated data acquisition and merging operation, the second acquisition node updates the predicted movement path and the acquisition status label and data version identifier in the acquisition trace, and pushes the acquisition trace to the acquisition node on the updated predicted movement path.

[0022] It should be noted that with the rapid development of information technology, the number of distributed multi-node data collection systems deployed in various application fields continues to grow. For example, in application fields such as entry and exit information collection, public security, intelligent manufacturing, and the Internet of Things, the same business object (such as entry and exit personnel, production equipment, and Internet of Things terminals) usually needs to pass through multiple collection nodes in sequence. Each collection node undertakes data collection tasks of different dimensions or different stages. This invention can be specifically applied to application fields such as entry and exit information collection, public security, intelligent manufacturing, and the Internet of Things.

[0023] For example, in one specific embodiment, due to the continuous increase in the flow of inbound and outbound personnel in the field of entry and exit information collection, various ports and border inspection stations have deployed more and more personnel information collection nodes, including self-service clearance gates, manual inspection counters, and mobile inspection terminals, to collect document information, facial images, fingerprint information, health declaration information, etc. of inbound and outbound personnel. In the actual clearance process, an inbound or outbound person usually needs to pass through multiple collection nodes in sequence, and each terminal undertakes information collection tasks of different dimensions or different stages. Therefore, how to achieve real-time sharing and collaborative collection of personnel information among multiple collection nodes, and avoid duplicate collection and information silos, has become a key issue in improving port clearance efficiency and management level. This invention can be applied to multi-terminal collaborative information collection in the field of entry and exit information collection. In this specific embodiment, the first and second collection nodes in this invention are specifically the collection terminal nodes in the port of entry and exit, and the target data object is specifically the personnel who need to enter and exit the country.

[0024] S1. The first acquisition node responds to the acquisition completion event of the target data object and generates an acquisition trace of the target data object. The acquisition trace is a structured data packet containing an acquisition status label, a completeness vector identifier, and a data version identifier. Once the first acquisition node completes the information acquisition operation of the target data object and obtains at least one piece of valid acquisition information of the target data object, it automatically triggers the acquisition trace generation process. When the present invention is applied to the field of entry and exit information acquisition, the valid acquisition information can specifically be the document information, facial image, fingerprint information, travel information and other information of the entry and exit personnel.

[0025] It should be noted that in existing technologies, after each acquisition node completes information collection, it only stores the data locally or uploads it to a central database. The information generated by the acquisition activity is not broadcast externally in a standardized format, causing other acquisition nodes to be unable to perceive the occurrence of the acquisition event. The purpose of this step is to transform a discrete, local acquisition activity into standardized collaborative signaling, i.e., acquisition traces, that can be recognized and used by other acquisition nodes, providing a data foundation for subsequent multi-acquisition node collaboration. The structured data packets of the acquisition traces specifically include the following data fields: S11, Identity Token field, used to store the anonymous identity identifier of the target data object. The identity token is generated by one-way hash encryption of the identity identifier of the target data object. The unique identification field of the target data object is extracted from the collection information of the target data object collected by the first collection node. If the present invention is applied to the field of entry and exit information collection, the unique identification field may specifically include the passport number, ID card number and travel permit number of the entry and exit personnel. In order to ensure that the sensitive information of the target data object is not leaked during the terminal transmission process, the identification field is encrypted by one-way hash (such as SHA-256) to generate a fixed-length anonymized identification. The anonymized identification is used as the unique identification index of the target data object in the multi-terminal collaborative collection process.

[0026] It should be noted that since the original unique identification fields (such as passport numbers) are highly sensitive data, there is a serious risk of information leakage if they are transmitted directly in plaintext between collection nodes. Therefore, each collection node needs a unified and tamper-proof method to identify the target data object. This step performs one-way hash desensitization processing on the unique identification field of the target data object to obtain an identity token. The identity token is used to achieve the uniqueness of the target data object and data security in the multi-terminal collaborative collection process.

[0027] S12, Context Metadata Fields group, used to store the acquisition timestamp, node location, and data flow direction identifier; The collection timestamp refers to the time when the first collection node collects information from the target data object, which can be accurate to the millisecond level. The node location refers to the location of the first collection node. If this invention is applied to the field of entry and exit information collection, the node location can be "XX Port - Arrival Hall - B3 Channel - Gate 2", the target data object is entry and exit personnel, and the data flow direction is specifically the clearance direction identifier of entry and exit personnel, which includes entry, exit, and transit. If this invention is applied to the field of intelligent manufacturing, the node location can be the location of processing equipment, detection terminals, or sensor nodes, the target data object is the production object, and the data flow direction is specifically the production process direction of the production object. If this invention is applied to the field of the Internet of Things, the node location can be the location of the Internet of Things terminal, the target data object is the mobile device, and the data flow direction is specifically the data transmission direction of the mobile device.

[0028] S13. Collection status label, used to store the collection status label of the target data object during the collection process. The collection status label includes normal status, data doubtful status, key inspection status and data needing to be collected. Based on the abnormal detection results during the collection process, the collection traces are assigned collection status labels, including normal status, data doubtful status, key inspection status, and data supplementation status.

[0029] In this embodiment, the assignment logic for the collection status label is as follows: (1) Normal status: All collected data are within the normal threshold range. For example, when this invention is applied to the field of entry and exit information collection, the judgment rule for the normal status is that the facial similarity of the entry and exit personnel is ≥0.85, the document is valid and successfully read, the fingerprint is successfully matched, the body temperature is between 36.0 and 37.3℃, and the image quality meets the standard; (2) Data doubtful status: A certain piece of collected data is not within the normal threshold range but no alarm is triggered. For example, when this invention is applied to the field of entry and exit information collection, the judgment rule for the data doubtful status is that the facial similarity is between 0.70 and 0.85. (2) .85 (low but not completely failed), image quality slightly below standard, body temperature between 37.3~37.5℃ (mild fever); (3) Key inspection status: triggering a clear alarm. For example, when this invention is applied to the field of entry and exit information collection, the key inspection status is determined by the failure to read the document chip, the hit of the warning list, and the body temperature ≥37.5℃; (4) Status requiring supplementary collection: key dimensions are missing and cannot be supplemented in the current collection. For example, when this invention is applied to the field of entry and exit information collection, the status requiring supplementary collection is determined by the successful collection of document number but the failure of facial image collection, and the incomplete collection of fingerprints. In the case of this invention applied to the field of entry and exit information collection, and where the normal threshold range includes normal threshold ranges for facial recognition similarity and fingerprint matching, the normal threshold range for facial recognition similarity can be set to 0.85-1.0. That is, when the facial recognition system performs identity authentication for entry and exit documents, if the similarity between the document photo and the facial image collected on-site is greater than 0.85, then the facial similarity data is considered to be within the normal threshold range. The associated information dimension can be preset according to the actual requirements of the entry and exit port. For example, the collected body temperature data can be used as a key information dimension.

[0030] It should be noted that if an anomaly is detected at a certain acquisition node during the acquisition process (such as a failed face comparison), and this anomaly information cannot be transmitted to subsequent terminals in a timely manner, the target data object may exploit the information delay to evade verification in later stages, creating a security vulnerability. Therefore, this step introduces acquisition status tags. Acquisition status tags enable the real-time marking of abnormal acquisition information during the acquisition process and its transmission along with the acquisition trace, allowing security risks to propagate from the first acquisition node along with the target data object, eliminating the time lag in information transmission. Here, downstream terminals refer to acquisition nodes after the first acquisition node, and upstream acquisition nodes refer to acquisition nodes that have already completed the acquisition of the target data object information.

[0031] S14, Completeness Vector Identifier, used to store the completeness vector identifier of the information dimensions that have been collected from the target data object; The information dimensions of the target data object collected by the first collection node are marked item by item to generate a completeness vector identifier. The information dimensions of the target data object collected are marked item by item based on the preset information dimensions. The preset information dimensions can be preset based on the actual application field. For example, when this invention is applied to the field of entry and exit information collection, the preset information dimensions can include at least document information, facial image, fingerprint information, health declaration information and travel information.

[0032] It should be noted that if downstream terminals cannot know which information dimensions of the target data object have been collected by upstream terminals, they can only perform full collection for each arriving target data object, that is, collect all information of the target data object. This can easily lead to data redundancy in multi-terminal collaborative collection. Therefore, this step introduces a completeness vector identifier. The completeness vector identifier enables downstream terminals to accurately know which information dimensions of the target data object have been collected and which information dimensions are missing before the target data object arrives, providing a decision basis for subsequent differentiated collection.

[0033] S15, Version Number field, is used to store the data version identifier, which is incremented each time the data collection trace is updated; The initial value of the data version identifier is 1. The data version identifier is incremented by 1 after each update of the data collection trace.

[0034] It should be noted that if the data collection traces lack version identifiers after multiple updates, the same data object held by different terminals may be used in a mix of old and new versions, resulting in data inconsistency. Therefore, this step introduces a data version identifier into the data collection traces. The introduction of the data version identifier provides a comparison benchmark for data consistency between multiple terminals, enabling any terminal to determine whether the data collection traces in its local cache are the latest version by using the version number.

[0035] S16, Signature Information Field, used to store the digital signature generated by the private key of the first acquisition node; It should be noted that the data collection traces are transmitted between the various data collection nodes. If there is a lack of anti-tampering mechanisms, malicious attackers may forge or tamper with the trace data (such as changing "key inspection" to "normal"), which may cause significant security risks. Therefore, this step introduces digital signatures into the data collection traces to provide integrity protection and source authentication for the data collection traces, ensuring that the data cannot be tampered with or denied in the transmission link.

[0036] S2. Analyze the candidate paths of the target data object based on the structured data in the collection traces, calculate the predicted movement path based on the candidate paths, push the collection traces to the second collection node on the predicted movement path, and write the collection traces to the local cache of the second collection node. S21. Analyze candidate paths for the target data object based on the structured data in the data collection traces, including: S211. Obtain the node position and data flow direction identifier in the collection trace, and load the pre-set directed terminal topology graph of the distributed system with the node position of the first collection node as the root node and the data flow direction identifier as the search direction constraint. Using the position of the first collection node as the root node, obtain each collection node in the passage direction of the target data object according to the data flow direction identifier, and use each collection node in the passage direction as a node to generate a pre-set directed terminal topology graph of the distributed system. This directed terminal topology graph uses each collection node as a node and the data flow direction identifier as a directed edge, and includes the set of all nodes that can legally flow from the node corresponding to the current first collection node to the next collection node.

[0037] It should be noted that this step uses a directed terminal topology graph to narrow the infinite possible movement paths into a finite and definite set of legal paths, so that the path prediction of the target data object is based on real physical space constraints, ensuring the physical feasibility of the prediction results and avoiding invalid pushes.

[0038] S212. Starting from the node position of the first acquisition node and taking each exit node in the directed terminal topology graph as the endpoint, obtain candidate paths in the directed terminal topology graph based on the preset path node hop count, and construct a candidate path set. In the directed terminal topology graph, starting from the current first data collection node and ending at each exit node in the graph, for example, when this embodiment is applied to the field of entry and exit information collection, each exit node in the directed terminal topology graph can specifically be the exit node of the entry and exit port clearance process. A breadth-first or depth-first algorithm is used to select all feasible paths with no more than a preset number of path node hops, and all feasible paths with no more than the preset number of path node hops are recorded as candidate paths. A candidate path set is then formed based on all candidate paths. The preset number of path node hops can be set to 6, meaning that all feasible paths with no more than six nodes are recorded as candidate paths.

[0039] Please see Figure 2 S22. Calculate the predicted mobile path based on the candidate path, including: S221. Read historical collection trajectory records within a preset statistical period from the historical database, perform hierarchical statistics according to data object type, data flow direction and transmission time period, calculate the pass frequency ratio of each candidate path under each hierarchical condition, and obtain the prior probability weight of each candidate path. For each candidate path in the candidate path set, flow frequency data within a preset statistical period is retrieved from historical clearance data. In this embodiment, the preset statistical period can preferably be the past 30 natural days. The statistical dimensions of the flow frequency data include at least the data object type, data flow direction, and transmission time period. The percentage of passage frequency for each candidate path under each layer dimension is calculated, and the prior probability weight of each candidate path is obtained. Specifically, the flow frequency data of the candidate path under each layer condition is obtained. Based on the statistical dimension of the flow frequency data, the matching dimension of the target data object in the information collection data of the first collection node that is in the same dimension as the statistical dimension is obtained. The historical passage frequency of each candidate path in the matching dimension within the statistical period and the historical total passage frequency of each candidate path within the statistical period are obtained. Based on the ratio of the historical passage frequency of the candidate path in the matching dimension to the historical total passage frequency, the percentage of historical passage frequency of the candidate path is calculated, and the percentage of historical passage frequency is used as the prior probability weight of each candidate path.

[0040] It should be noted that this step obtains the historical pass frequency ratio of candidate paths through multi-dimensional historical statistical dimension matching, and characterizes the prior probability weight of each candidate path based on the historical pass frequency ratio. The prior probability weight is used to filter out candidate paths that the target data object is more likely to pass through, thereby improving the efficiency of path prediction.

[0041] S222. Obtain the real-time load parameters of each collection node in the candidate path. The real-time load parameters include the current queue number and the average processing time. Send load status query requests to the terminal nodes in each candidate path, and receive the returned real-time load parameters. The real-time load parameters include at least the current queue number and average processing time of each terminal node.

[0042] S223. Calculate the standardized load index of each acquisition node based on the real-time load parameters, and calculate the path weight of the candidate path based on the prior probability weight of the candidate path and the standardized load index of each acquisition node in the candidate path. The path weight of candidate paths is calculated by introducing a preset load penalty mechanism. Specifically, the standardized load index of each acquisition node is calculated based on real-time load parameters, and the standardized load index is mapped to the load penalty parameter of each acquisition node. The load penalty parameters of all acquisition nodes in each candidate path are multiplied together and then multiplied by the prior probability weight to obtain the path weight of the candidate path. The physical meaning of the load penalty mechanism is that the higher the load of a node, the lower its prior probability weight is. The prior probability weights of each node in the candidate path are accumulated to obtain the prior probability weight of the candidate path, which includes the following steps: (1) Calculate the standardized load index of the collection nodes corresponding to each node in the candidate path based on the current queue number, average processing time, and online status. Specifically, obtain the maximum parallel queue number and the average standard processing time of the collection nodes corresponding to each node in the candidate path, calculate the ratio of the real-time queue number of the collection node to the maximum parallel queue number, calculate the ratio of the average processing time of the collection node to the average standard processing time, and sum these two ratios to obtain the standardized load index of the collection nodes corresponding to each node in the candidate path. The maximum parallel queue number is set based on the type of the collection node. For example, when this invention is applied to the field of entry and exit information collection, if the type of the collection node is a self-service gate, the maximum parallel queue number is set to 3 people; if the type of the collection node is a manual inspection station, the maximum parallel queue number is set to 6 people. In addition, for collection nodes that are in normal online status, calculate the standardized load index according to the above method. For collection nodes that are offline or faulty online, record their standardized load index as infinity so that the candidate path containing the collection node is automatically excluded in the subsequent correction. (2) The load penalty parameter of the acquisition node corresponding to each node in the candidate path is calculated using an exponentially decaying load penalty function. The specific calculation method is as follows: with the natural constant e as the base, the product of the negative value of the preset load sensitivity constant and the standardized load exponent is used as the exponent for power operation to obtain the load penalty parameter of the acquisition node corresponding to each node in the candidate path. In this embodiment, the preferred value of the preset load sensitivity constant is 2. It should be noted that the exponentially decaying load penalty function used based on the load penalty mechanism aims to map the standardized load exponent of the acquisition node into a load penalty parameter, which is used to correct the prior probability weight of the acquisition node. The specific physical meaning is that the larger the standardized load exponent of the acquisition node, the smaller its corresponding load penalty parameter, and the greater the reduction in its prior probability weight.

[0043] (3) For the kth candidate path in the candidate path set, if the candidate path needs to pass through m nodes after removing the starting point, the load penalty parameters corresponding to all nodes passed through from the 1st to the mth node of the candidate path are multiplied together to obtain the multiplication result. The multiplication result is multiplied with the prior probability weight of the candidate path to obtain the path weight of the candidate path. The path weight of all candidate paths in the candidate path set is calculated based on the above method.

[0044] It should be noted that if only historical statistics are relied upon without considering real-time load status, the target data object may be guided to the collection node with the longest queue, resulting in process congestion. This step introduces real-time load correction to enable the predicted movement path to dynamically avoid congested nodes, thereby improving the process efficiency of the target data object.

[0045] S224. Normalize the path weights of the candidate paths, select the candidate path with the highest normalized path weight as the predicted mobile path, and generate an ordered terminal sequence based on the node position order of the collection nodes on the predicted mobile path. When selecting the candidate path with the highest corrected path weight as the predicted moving path, if there are candidate paths with the same corrected path weight, the candidate path with the fewest total number of nodes is selected as the predicted moving path. The collection nodes on the predicted moving path are arranged into an ordered terminal sequence, that is, the ordered terminal sequence contains all collection nodes on the predicted moving path, and all collection nodes in the ordered terminal sequence are arranged in the order of the actual physical flow of the target data object along the predicted moving path.

[0046] It should be noted that the ordered terminal sequence provides a unique and definite terminal sequence for the subsequent push operation of the collected traces, making the entire terminal collaborative collection process executable and predictable, and avoiding uncertainty in system behavior caused by decision ambiguity.

[0047] Please see Figure 3 S23, Pushing the collected traces to the second collection node on the predicted movement path, including: S231. Obtain the online status of each acquisition node in the ordered terminal sequence; The online status of each acquisition node in the ordered terminal sequence includes three states: normal operation, offline, and fault.

[0048] S232. Remove the acquisition nodes that are offline or faulty from the ordered terminal sequence, and reconnect the subsequent terminals of the removed terminals to the previous terminal to generate an updated ordered terminal sequence. If a terminal in the ordered terminal sequence is currently offline or faulty, it is removed from the ordered terminal sequence, and the subsequent terminals of that terminal on the predicted mobility path are relinked to the previous terminal to ensure that all nodes on the predicted mobility path remain connected.

[0049] S233. Serialize the collected traces into transmission data packets, and adaptively select the transmission protocol from the preset protocol set according to the network environment between each collection node. When encapsulating data collection traces into transmission data packets, priorities can be assigned to the encapsulated transmission data packets. Specifically, the transmission data packets corresponding to data collection traces with the collection status label as "key to inspection" are assigned the highest priority, while the transmission data packets corresponding to data collection traces with other collection status labels are assigned normal priority.

[0050] Adaptively selecting a transmission protocol from a preset protocol set based on the network environment between each data acquisition node. Specifically, the preset protocol set may include direct P2P transmission protocols, transmission protocols relayed by a central server, and broadcast / multicast transmission protocols. For example, if the predicted network environment between the data acquisition nodes on the mobile path is that they are on the same local area network and support P2P communication, then a direct P2P transmission protocol is used to directly establish a point-to-point encrypted channel; if the predicted network environment between the data acquisition nodes on the mobile path is that they are across network segments or have firewall restrictions, then a transmission protocol relayed by a central server is used, with reliable message queue relayed through the central server; if the predicted network environment between the data acquisition nodes on the mobile path requires simultaneous push to a large number of terminals, then a broadcast / multicast transmission protocol is used.

[0051] S234. Introduce an acknowledgment mechanism with timeout retransmission, and send the transmission data packet to the second acquisition node in the updated ordered terminal sequence according to the acknowledgment mechanism and transmission protocol. The confirmation mechanism with timeout retransmission works as follows: after the first acquisition node sends the data packet corresponding to the acquisition trace, it waits for the reception confirmation signal from the second acquisition node on the predicted movement path. If the reception confirmation signal is not received within a specified time window (e.g., 5ms), retransmission is automatically triggered. The maximum number of retransmissions can be preset to 3. The second acquisition node that still does not receive the reception confirmation signal after exceeding the retransmission limit is marked as a push failure terminal, and the terminal ID and other terminal identifiers of the push failure terminal are recorded in the system log for subsequent manual retransmission.

[0052] It should be noted that, due to the risk of packet loss during data transmission, the second acquisition node may not receive the acquisition traces if acknowledgment and retransmission are not performed. This step, through the acknowledgment mechanism with timeout retransmission, can significantly improve the delivery rate of data packets to the second acquisition node and ensure stable transmission of data packets.

[0053] Please see Figure 4S24. Write the acquisition traces to the local cache of the second acquisition node, including: S241. After receiving and verifying the transmission data packet of the acquisition trace, the second acquisition node writes the acquisition trace as pre-synchronization data to the local cache of the second acquisition node. After receiving the data packet, the second acquisition node first verifies the validity of the digital signature in the data packet to prevent data tampering, and verifies the integrity of the data packet. After the verification is successful, it parses out all fields in the acquisition trace.

[0054] It should be noted that, in order to prevent malicious data injection attacks, digital signatures must be verified to ensure that the received collection traces are issued by legitimate terminals and have not been tampered with. For example, when this invention is applied to the field of entry and exit information collection, if the received entry and exit personnel data is not stored in the local high-speed cache of the second collection node, when the entry and exit personnel arrive at the second collection node, they still need to query the central database or wait for data loading, which cannot achieve zero-latency response for data query. This step ensures system security through digital signature verification and prevents forged data intrusion. At the same time, by storing the collection traces as pre-synchronized data in the local high-speed cache of the second collection node, nanosecond-level fast reading of the collection traces is achieved, providing speed assurance for real-time decision-making on site.

[0055] S242. Based on the integrity vector identifier and acquisition status label in the acquisition trace, the data acquisition synchronization mode of the second acquisition node is determined in advance before the target data object arrives. The data acquisition synchronization modes include full skip mode, incremental acquisition mode and depth verification mode. S2421. Read the integrity vector identifier in the collection trace. If the integrity vector identifier indicates that all preset information dimensions of the target data object have been collected and the collection status label is normal, then the data collection synchronization mode is determined to be the full skip mode. The specific collection operation corresponding to the full skip mode is that the second collection node verifies the identity token of the target data object and skips the information collection of the target data object.

[0056] S2422. If the completeness vector indicates that at least one preset information dimension has not been collected, and the collection status label is in normal state, then the data collection synchronization mode is determined to be the incremental supplementary collection mode. The collection operation corresponding to the incremental supplementary collection mode is to collect only the missing information dimension in the collection trace corresponding to the target data object. The missing information dimension can be obtained by comparing the collected information dimension with the preset information dimension based on the completeness vector identifier in the collection trace. For example, when this invention is applied to the field of entry and exit information collection, the preset information dimension includes at least document information, facial image, fingerprint information, health declaration information, and travel information.

[0057] S2423. If the data collection status label is either a data doubtful status or a key inspection status, then the data collection synchronization mode will be determined as the deep verification mode. The specific data collection operation corresponding to the deep verification mode is to trigger the key inspection process and trigger the early warning notification.

[0058] It should be noted that, for example, when this invention is applied to the field of entry and exit information collection, if the data collection synchronization mode of the second collection node for the target data object is analyzed based on the integrity vector identifier only after the entry and exit personnel arrive, the collection waiting time of the target data object will be increased. Therefore, this application adopts a pre-obtained data collection synchronization mode, so that the second collection node can directly perform the collection action based on the data collection synchronization mode when the target data object arrives, instead of needing to make on-site decisions on the data collection synchronization mode, further reducing the customs clearance delay for entry and exit personnel.

[0059] Please see Figure 5 S3, the second acquisition node responds to the arrival event of the target data object by reading the acquisition trace from the local cache and performing differentiated data acquisition and merging operations on the target data object according to the preset data acquisition synchronization mode based on the integrity vector identifier; S31. The second acquisition node acquires the on-site identity data of the target data object, generates an on-site identity token through one-way hash encryption, and performs key-value lookup in the local cache using the on-site identity token as the retrieval key. For example, when this invention is applied to the field of entry and exit information collection, after the target data object arrives at the second collection node, the physical identity collection of the terminal is triggered by scanning the ID card or facial recognition, the on-site identity token of the target data object is extracted, and the on-site identity token is used as an index to search in the pre-synchronized data in the local cache of the second collection node to obtain two search result branches: the query hit branch and the query miss branch.

[0060] S32. If the query is successful, read the corresponding pre-synchronized data from the local cache, obtain the integrity vector identifier and acquisition status label of the target data object, and determine the data acquisition synchronization mode. After a successful query, valid and unexpired collection traces are retrieved from the pre-synchronized data in the local cache. At this point, differentiated collection is performed on the target data object according to the data collection synchronization mode. The specific collection operation of differentiated collection is determined based on the data collection synchronization mode of the target data object.

[0061] S33. If the query does not find the target data object, send a real-time query request to the central server to obtain the most recent collection trace of the target data object. After writing the most recent collection trace to the local cache, perform the differentiated data collection and merging operation. If the query fails and no valid cache is found in the pre-synchronized data of the local cache, the second collection node initiates a real-time query request to the central server at the port of entry / exit to query the most recent collection trace of the target data object. If the central server returns valid data, the valid data is written to the local cache of the second collection node, and differentiated collection is performed on the target data object according to the data collection synchronization mode obtained from the collection trace. If the central server does not return valid data, the second collection node performs full collection on the target data object, that is, collects all the information of the target data object, and generates a new collection trace corresponding to the target data object after the collection is completed.

[0062] It should be noted that the terminal that actually receives the target data object may not receive the pre-synchronized data corresponding to the collection trace due to network packet loss, terminal restart, or personnel being temporarily guided to an unpredictable path. This step implements a fast clearance strategy at the level of "skipping collection" or "collecting only one fingerprint" through the cache hit branch, which reflects the core value of multi-terminal collaborative collection. The cache miss branch provides a dual guarantee mechanism, namely central query and full collection, to ensure that the clearance process will not be interrupted under any extreme circumstances, thereby improving the fault tolerance and robustness of multi-terminal collaborative collection.

[0063] S34. Perform the corresponding data acquisition operation according to the data acquisition synchronization mode. The data acquisition operation includes skipping data acquisition, only acquiring missing dimension data, and triggering a detailed data verification process. The data acquisition synchronization modes include full skip mode, incremental supplementary acquisition mode, and deep verification mode. The data acquisition operations corresponding to the full skip mode, incremental supplementary acquisition mode, and deep verification mode are skip data acquisition, only acquire missing dimension data, and trigger detailed data verification process, respectively. Specifically, skip data acquisition means that the second acquisition node only verifies the identity token of the target data object and skips the acquisition of information of the target data object. Only acquire missing dimension data means that only acquire the missing information dimension in the acquisition trace corresponding to the target data object. The missing information dimension can be obtained by obtaining the acquired information dimension based on the completeness vector identifier in the acquisition trace and comparing the acquired information dimension with the preset information dimension. The preset information dimension includes at least document information, facial image, fingerprint information, health declaration information, and travel information. Trigger detailed data verification process means triggering key inspection process and triggering early warning notification.

[0064] S4. After completing the differential data collection and merging operation, the second collection node updates the predicted movement path and the collection status label and data version identifier in the collection trace, and pushes the collection trace to the collection node of the updated predicted movement path. S41. After completing the differential data collection and merging operation, the second collection node merges the newly collected data fields with the pre-synchronized data in the local cache at the field level, updates the collection timestamp, node location information, integrity vector identifier and collection status label in the collection trace, and increments the data version identifier in the collection trace. Since the new data generated by the differential data collection and merging operation has scattered fields compared to the pre-synchronized data, if it is directly transmitted downwards without merging, the downstream terminal will receive fragmented information and will not be able to accurately determine the complete state of the target data object. Therefore, the new data generated by the differential data collection and merging operation is merged with the pre-synchronized data in the local cache at the field level. Specifically, the pre-synchronized data of the second collection node is used as the basis, and the corresponding fields are covered or filled with the newly collected data of the second collection node to form a complete update record, which is persistently backed up in the local high-speed cache of the second collection node. Through field merging, the unique information data of the target data object is formed, ensuring the consistency and integrity of the data corresponding to the target data object.

[0065] In addition, the collection timestamp is updated to the time when the second collection node completes this collection; the node position is updated to the position of the second collection node; the integrity vector identifier is updated to the latest integrity vector identifier of the collection trace after collection in the data collection synchronization mode; the collection status label is updated based on the abnormal information of the collected data during the collection process of the second collection node; and the data version identifier is updated, specifically by incrementing the data version identifier of the collection trace by 1.

[0066] S42. Compare the updated collection traces with the preset data synchronization termination conditions; The preset data synchronization termination conditions include: the integrity vector indicating that all preset information dimensions have been collected, the collection status label being in a normal state, and the out-degree of the current collection node in the directed acyclic terminal topology graph being zero. The system retrieves the completeness vector identifier, acquisition status label, and topological attributes of the node corresponding to the second acquisition node in the directed terminal topology graph from the updated acquisition trace. It then compares these attributes with preset data synchronization termination conditions. The preset minimum dimension threshold in the data synchronization termination conditions can be set as the information dimension ratio, i.e., the ratio of the acquired information dimension in the acquisition trace to the preset information dimension. For example, the preset minimum dimension threshold can be set to 90% of the information dimension ratio. A current acquisition node having an out-degree of 0 in the directed terminal topology graph means that this acquisition node is a topological convergence node in the directed terminal topology graph, with no subsequent acquisition nodes downstream.

[0067] In addition, if this embodiment is applied to the field of entry and exit information collection, the on-site staff at the entry and exit port can also manually trigger the "forced collection termination" command through the terminal interface. This command is applicable to abnormal situations such as entry and exit personnel being intercepted or transferred to other departments for processing.

[0068] It should be noted that by setting preset data synchronization termination conditions, the system can prevent the prediction and push of the corresponding collection traces after the target data object has left the collection process, thus avoiding resource waste and timely release of system computing and storage resources.

[0069] S421. If the updated acquisition trace does not meet the preset data synchronization termination condition, the predicted movement path is recalculated starting from the node position of the second acquisition node. If the data synchronization termination condition is not met, the predicted movement path is recalculated starting from the location of the second acquisition node. Based on the recalculated and updated ordered terminal sequence, the updated acquisition traces are simultaneously and in parallel pushed to all downstream acquisition nodes on the new predicted movement path and to acquisition nodes that have not yet met in the original predicted movement path. The receiving end determines whether to perform a local cache update by comparing the global version number with the local cache version number. Specifically: If the updated acquisition trace does not meet the preset data synchronization termination condition, the second acquisition node is used as the new starting point, and the predicted movement path is updated through step S2. The updated acquisition trace is then pushed in parallel to all downstream acquisition nodes on the predicted movement path. Steps S2 to S4 are repeated until the updated acquisition trace meets the data synchronization termination condition. Furthermore, when pushing the updated acquisition trace in parallel to all downstream acquisition nodes on the predicted movement path, each downstream acquisition node compares the data version identifier of the acquisition trace in its local cache with the data version identifier of the currently pushed acquisition trace. Only if the data version identifier of the currently pushed acquisition trace is greater than the data version identifier of the acquisition trace in its local cache is the acquisition trace in the downstream acquisition node updated; otherwise, the currently pushed acquisition trace is discarded.

[0070] S422. If the updated data collection trace meets the preset data synchronization termination condition, then mark the data collection trace as terminated and stop updating the data collection trace. If the data synchronization termination condition is met, the acquisition trace will be marked as terminated, the complete data record of the target data object and the complete version sequence of the acquisition trace will be encrypted and uploaded to the preset central database, and a cache release command will be sent to the acquisition node that has received the acquisition trace but has not yet encountered the target data object.

[0071] It should be noted that uploading the collected data to the central database for archiving provides tamper-proof original evidence for post-event auditing, tracing, and data analysis. Furthermore, the cache release command precisely cleans up any remaining expired data collection traces, ensuring efficient utilization of terminal storage resources.

[0072] Example 2, please refer to Figure 6 , Figure 6 This invention discloses a multi-terminal distributed data collaborative acquisition and synchronization system, comprising: a first acquisition module, wherein a first acquisition node, in response to a target data object acquisition completion event, generates an acquisition trace of the target data object, the acquisition trace being a structured data packet containing an acquisition status label, a completeness vector identifier, and a data version identifier; a path prediction module, used to analyze candidate paths of the target data object based on the structured data in the acquisition trace, calculate a predicted movement path based on the candidate paths, push the acquisition trace to a second acquisition node on the predicted movement path, and write the acquisition trace to the local cache of the second acquisition node; a second acquisition module, wherein the second acquisition node, in response to a target data object arrival event, reads the acquisition trace from the local cache, performs a differentiated data acquisition and merging operation on the target data object according to a preset data acquisition synchronization mode based on the completeness vector identifier; and an information update module, used to update the predicted movement path and the acquisition status label and data version identifier in the acquisition trace after the second acquisition node completes the differentiated data acquisition and merging operation, and pushes the acquisition trace to the acquisition node on the updated predicted movement path.

[0073] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0074] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0075] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0076] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0077] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

[0078] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for multi-terminal distributed data collaborative acquisition and synchronization, characterized in that, Includes the following steps: In response to the target data object's acquisition completion event, the first acquisition node generates an acquisition trace of the target data object. The acquisition trace is a structured data packet containing an acquisition status label, a completeness vector identifier, and a data version identifier. Based on the structured data analysis in the acquisition traces, candidate paths for the target data object are analyzed, and a predicted movement path is calculated based on the candidate paths. The acquisition traces are then pushed to the second acquisition node on the predicted movement path, and the acquisition traces are written to the local cache of the second acquisition node. In response to the arrival event of the target data object, the second acquisition node reads the acquisition trace from the local cache and performs differentiated data acquisition and merging operations on the target data object according to the preset data acquisition synchronization mode based on the integrity vector identifier. After completing the differential data collection and merging operation, the second collection node updates the predicted movement path and the collection status label and data version identifier in the collection trace, and pushes the collection trace to the collection node of the updated predicted movement path. After writing the acquisition traces to the local cache of the second acquisition node, the process also includes: Read the integrity vector identifier in the collection trace. If the integrity vector identifier indicates that all preset information dimensions of the target data object have been collected and the collection status label is normal, then the data collection synchronization mode is set to full skip mode. If the completeness vector indicates that at least one preset information dimension has not been collected, and the collection status label is in normal status, then the data collection synchronization mode will be determined as incremental supplementary collection mode. If the data collection status label is either "data in doubt" or "data under key inspection", then the data collection synchronization mode will be set to "deep verification mode".

2. The multi-terminal distributed data collaborative acquisition and synchronization method according to claim 1, characterized in that, The data fields of the structured data packet include: The identity token field is used to store the anonymous identity identifier of the target data object, which is generated by one-way hash encryption of the identity identifier of the target data object. The context metadata field group is used to store the collection timestamp, node location, and data flow direction identifier; The collection status label is used to store the collection status label of the target data object during the collection process. The collection status label includes normal status, data doubtful status, key inspection status, and data needing to be collected. Completeness vector identifier, used to store the completeness vector identifier of the information dimensions that have been collected from the target data object; The version number field is used to store the data version identifier, which is incremented each time the data collection trace is updated; The signature information field is used to store the digital signature generated by the private key of the first acquisition node.

3. The multi-terminal distributed data collaborative acquisition and synchronization method according to claim 1, characterized in that, Based on the structured data analysis in the data collection traces, candidate paths for the target data object are analyzed, including: Obtain the node positions and data flow direction identifiers in the collection traces, and load the pre-set directed terminal topology graph of the distributed system with the node position of the first collection node as the root node and the data flow direction identifier as the search direction constraint. Starting from the node position of the first acquisition node and ending at each exit node in the directed terminal topology graph, candidate paths in the directed terminal topology graph are obtained based on the preset path node hop count, and a candidate path set is constructed.

4. The multi-terminal distributed data collaborative acquisition and synchronization method according to claim 1, characterized in that, Calculate the predicted mobile path based on the candidate path, including: Historical data collection trajectory records within a preset statistical period are read from the historical database. The data are then statistically analyzed in layers according to data object type, data flow direction, and transmission time period. The pass frequency ratio of each candidate path under each layer condition is calculated, and the prior probability weight of each candidate path is obtained. Obtain the real-time load parameters of each collection node in the candidate path, including the current queue number and average processing time; The standardized load index of each acquisition node is calculated based on the real-time load parameters. The path weight of the candidate path is calculated based on the prior probability weight of the candidate path and the standardized load index of each acquisition node in the candidate path. The path weights of the candidate paths are normalized, and the candidate path with the highest normalized path weight is selected as the predicted mobile path. An ordered terminal sequence is generated based on the node position order of the collected nodes on the predicted mobile path.

5. The multi-terminal distributed data collaborative acquisition and synchronization method according to claim 1, characterized in that, Pushing the collected traces to a second collection node on the predicted movement path includes: Obtain the online status of each acquisition node in the ordered terminal sequence; The acquisition nodes that are offline or faulty are removed from the ordered terminal sequence, and the subsequent terminals of the removed terminals are reconnected to the previous terminal to generate an updated ordered terminal sequence. The collected traces are serialized into transmission data packets, and a transmission protocol is adaptively selected from a preset protocol set according to the network environment between each collection node. An acknowledgment mechanism with timeout retransmission is introduced, and the data packets are sent to the second acquisition node in the updated ordered terminal sequence based on the acknowledgment mechanism and transmission protocol.

6. The multi-terminal distributed data collaborative acquisition and synchronization method according to claim 1, characterized in that, The acquisition traces are written to the local cache of the second acquisition node, including: After receiving and verifying the data packet of the acquisition trace, the second acquisition node writes the acquisition trace as pre-synchronization data to the local cache of the second acquisition node.

7. The multi-terminal distributed data collaborative acquisition and synchronization method according to claim 1, characterized in that, Based on the completeness vector identifier, a differentiated data collection and merging operation is performed on the target data object according to a preset data collection synchronization mode, including: The second acquisition node obtains the on-site identity data of the target data object, generates an on-site identity token through one-way hash encryption, and performs key-value lookup in the local cache using the on-site identity token as the retrieval key; If the query hits, read the corresponding pre-synchronized data from the local cache, obtain the integrity vector identifier and collection status label of the target data object, and determine the data collection synchronization mode; If the query fails to find the target data object, a real-time query request is sent to the central server to obtain the most recent collection traces of the target data object. After writing the most recent collection traces to the local cache, a differentiated data collection and merging operation is performed.

8. The multi-terminal distributed data collaborative acquisition and synchronization method according to claim 1, characterized in that, Update the predicted movement path and the acquisition status label and data version identifier in the acquisition trace, including: After completing the differential data collection and merging operation, the second collection node merges the newly collected data fields with the pre-synchronized data in the local cache at the field level, updates the collection timestamp, node location information, integrity vector identifier and collection status label in the collection trace, and increments the data version identifier in the collection trace. The updated data collection traces are compared with the preset data synchronization termination conditions; If the updated acquisition trace does not meet the preset data synchronization termination condition, the predicted movement path is recalculated starting from the node position of the second acquisition node; if the updated acquisition trace meets the preset data synchronization termination condition, the acquisition trace is marked as terminated and the update of the acquisition trace is stopped.

9. A system using the multi-terminal distributed data collaborative acquisition and synchronization method as described in any one of claims 1-8, comprising: The first acquisition module and the first acquisition node respond to the acquisition completion event of the target data object and generate an acquisition trace of the target data object. The acquisition trace is a structured data packet containing an acquisition status label, a completeness vector identifier, and a data version identifier. The path prediction module is used to analyze the candidate paths of the target data object based on the structured data in the collection traces, calculate the predicted movement path based on the candidate paths, push the collection traces to the second collection node on the predicted movement path, and write the collection traces to the local cache of the second collection node. The second acquisition module, in response to the arrival event of the target data object, reads the acquisition traces from the local cache and performs differentiated data acquisition and merging operations on the target data object according to the preset data acquisition synchronization mode based on the integrity vector identifier; The information update module is used to update the predicted movement path and the collection status label and data version identifier in the collection trace after the second collection node completes the differential data collection and merging operation, and push the collection trace to the collection node of the updated predicted movement path.

Images