Search

CN-121985380-A - Mine multisource data fusion transmission method and device based on distributed edge agent cooperation

CN121985380ACN 121985380 ACN121985380 ACN 121985380ACN-121985380-A

Abstract

A mine multisource data fusion transmission method and device based on distributed edge agent cooperation relates to the field of data transmission. The method comprises the steps of collecting heterogeneous data through a plurality of data collection terminals deployed in a mine production environment, generating a data certificate based on the heterogeneous data, splitting a data processing task into task sub-slices of different processing types through a main agent analysis value mark, selecting a slave agent from bidding results of other edge computing nodes aiming at bidding information, constructing the main agent and the slave agent into an edge agent alliance for processing the heterogeneous data, performing real-time interaction of event characteristics in the edge agent alliance, executing cross-mode joint coding on associated heterogeneous data streams based on the event characteristics, and aggregating and uploading the joint coded data to a cloud in a transmission window returned by a 5G base station. By implementing the technical scheme, the mine multi-source heterogeneous data can be effectively processed, and the distribution and the dynamic scheduling of network resources according to the needs can be realized.

Inventors

  • ZHANG LEI
  • Zhu Gaige

Assignees

  • 北京新润通科技有限公司

Dates

Publication Date
20260505
Application Date
20260115

Claims (10)

  1. 1. A mine multisource data fusion transmission method based on distributed edge agent cooperation is characterized by comprising the following steps: Collecting heterogeneous data through a plurality of data acquisition terminals deployed in a mine production environment, generating a value mark for the heterogeneous data through a 5G fusion gateway based on equipment types of data sources and a preset value rule base, calculating a hash value of the heterogeneous data based on the value mark to serve as a digital fingerprint, generating a data certificate based on the digital fingerprint, and broadcasting the data certificate to a distributed account book commonly maintained by a plurality of edge computing nodes deployed in the same area; Determining an edge computing node which verifies and records the data certificate in the distributed account book as a master agent, resolving the value mark by the master agent, splitting a data processing task into task sub-pieces with different processing types, broadcasting bidding information containing resource requirements of the task sub-pieces to other edge computing nodes, selecting a slave agent from bidding results of the other edge computing nodes on the bidding information, and constructing the master agent and the slave agent into an edge agent alliance for processing the heterogeneous data; in the edge agent alliance, each slave agent builds a data processing pipeline according to the logic sequence of the task sub-slices, and the slave agents processing different modal data perform real-time interaction of event characteristics in the pipeline processing process and execute cross-modal joint coding on the associated heterogeneous data streams based on the event characteristics; And predicting the transmission requirement of the data packet to be uploaded based on the locally stored historical task log by the edge agent alliance, generating a transmission request suggestion packet containing predicted flow, priority and suggested transmission parameters, sending the transmission request suggestion packet to a 5G base station, receiving a transmission window returned by the 5G base station, and aggregating and uploading the jointly encoded data to a cloud in the transmission window.
  2. 2. The method of claim 1, wherein the generating value tags for the heterogeneous data via a 5G fusion gateway based on the device type of the data source and a preset value rule base comprises: Analyzing the received heterogeneous data through the 5G fusion gateway, identifying the equipment type and the data type, and extracting a semantic unit corresponding to the data type; Obtaining a unit value density score of the semantic unit based on the equipment type, the content of the semantic unit and the frequency of the historical similar data called and decision-making by the cloud analysis model; Aggregating byte intervals occupied by semantic units with the unit value density scores higher than a preset first threshold value in the heterogeneous data into high-value data segments, aggregating byte intervals occupied by semantic units with the unit value density scores lower than a preset second threshold value in the heterogeneous data into low-value data segments, and the balance being medium-value data segments, generating segment-level value marks for each data segment, and generating a segment mapping table based on the segment-level value marks; And integrating all segment level value marks to generate a global value mark of the heterogeneous data.
  3. 3. The method of claim 2, wherein the computing the hash value of the disparate data based on the value indicia as a digital fingerprint comprises: extracting original data blocks corresponding to each data segment from the original byte stream of the heterogeneous data according to the byte interval corresponding to each data segment in the segment mapping table through the 5G fusion gateway; for each original data block, respectively calculating a content hash as a block-level fingerprint of the data block; Constructing a merck subtree of a target data segment based on the block-level fingerprints of a plurality of data blocks belonging to the target data segment, and taking a root hash of the merck subtree as a segment integrity fingerprint of the target data segment, wherein the target data segment is any one data segment in the heterogeneous data; extracting a segment value mark corresponding to the target data segment, and calculating a hash value of the segment value mark as a segment value abstract fingerprint of the target data segment; Taking segment integrity fingerprints and segment value abstract fingerprints of all data segments as leaf nodes, constructing a global merck tree, and generating a global root hash; And packaging the global root hash, the segment integrity fingerprint, the segment value summary fingerprint and the segment mapping table together to generate the digital fingerprint of the heterogeneous data.
  4. 4. The method according to claim 1, wherein the slave agents processing different modality data perform real-time interaction of event features during pipeline processing, specifically comprising: A first slave agent processing time sequence sensing data, when detecting that the original data value of the monitored target parameter exceeds a preset third threshold value locally, sending a space-time sensing request packet to the master agent, wherein the space-time sensing request packet comprises event characteristics, and the event characteristics comprise a time stamp, an event type and the physical position of the first slave agent; Retrieving and waking up a plurality of geographically or logically related second slave agents from a device topology and semantic relationship graph by the master agent based on the event type and the physical location; Broadcasting cooperative sensing instructions to a plurality of second slave agents through the master agent, and receiving confirmation signals and confidence degrees returned by the plurality of second agents, wherein the cooperative sensing instructions comprise event types and time windows, and the confirmation signals are used for representing that cooperative verification features exist; If the number of the received confirmation signals and the confidence coefficient weighted value of the master agent exceed a preset fourth threshold value in a preset decision time, judging that the cross-mode collaborative perception event is established, generating a fusion event descriptor, and distributing the fusion event descriptor to all target second slave agents returning the confirmation signals, wherein the fusion event descriptor comprises an event type, a fusion timestamp, a trigger source ID, a collaborative verification source ID list and a data value confidence coefficient.
  5. 5. The method of claim 4, wherein performing cross-modal joint coding on the associated heterogeneous data streams based on the event characteristics comprises: In the edge agent alliance, the target second slave agent analyzes the fusion event descriptor to obtain the event type and the data value confidence, available computing resources, predicted bandwidth and current mine production stage information of the target second slave agent are obtained in real time, a dynamic context is formed through fusion, a model is generated through a preset strategy based on the dynamic context, and a personalized coding strategy vector is generated; The target second slave agent issues a virtual resource offer in the edge agent alliance according to the resource demand in the personalized coding strategy vector, and selects a third slave agent based on bidding results of other slave agents in the edge agent alliance; Encoding a video stream by the target second slave agent, positioning associated data fragments which are overlapped in time or logically associated with the event features extracted from the video stream from a non-video mode data stream corresponding to the collaborative evidence source ID list by the third slave agent, encoding the associated data fragments by using encoding parameters higher than a first quality threshold, and compressing non-associated data fragments by using encoding parameters lower than a second quality threshold, wherein the second quality threshold is lower than the first quality threshold; Binding and packaging the video data encoded by the target second slave agent and the non-video data encoded or compressed by the third slave agent according to a unified space-time index to obtain a multi-mode data unit.
  6. 6. The method according to claim 4, wherein the step of predicting, by the edge agent consortium, a transmission demand of a data packet to be uploaded based on a locally stored historical task log, and generating a transmission request suggestion packet including a predicted traffic, a priority, and a suggested transmission parameter, specifically comprises: Performing time sequence mode analysis on the historical task log through the edge agent alliance to obtain a flow base line; Acquiring the number of slave agents currently active in the edge agent alliance, the current calculation load of each slave agent and the data value confidence carried by the fusion event descriptor, mapping the data value confidence into a transmission priority weight, calculating to obtain a real-time collaborative processing capacity coefficient by combining the number of the slave agents currently active and the calculation load, and correcting the flow baseline by utilizing the real-time collaborative processing capacity coefficient to obtain a predicted flow; The predicted flow, the transmission priority weight, the current average network time delay and the packet loss rate are used as query characteristics to be input into a transmission strategy knowledge base of the edge intelligent agent alliance, a suggested transmission parameter set is matched, the transmission parameter set comprises coding redundancy, maximum retransmission times and first packet transmission time delay, and the transmission strategy knowledge base stores mapping relations from various preset network states and task characteristics to an optimal transmission parameter set; And assembling the predicted flow, the transmission priority weight and the transmission parameter set to generate the transmission request suggestion packet.
  7. 7. The method of claim 6, wherein the method further comprises: If the 5G base station receives a plurality of transmission request proposal packets from different edge agent alliances in the same area in the same scheduling period, the 5G base station sorts the plurality of transmission request proposal packets according to the transmission priority weight and the first packet sending time delay in each transmission request proposal packet, and performs preliminary non-overlapping window pre-allocation on a time-frequency resource grid based on the required resource of each transmission request proposal packet to form a plurality of candidate transmission window schemes; The candidate window schemes comprising window starting time, window duration and guaranteed bandwidth corresponding to the windows are respectively fed back to corresponding edge agent alliances through the 5G base station; And if the corresponding edge agent alliance confirms that the candidate window scheme is accepted, the 5G base station locks and reserves a corresponding wireless resource block on a time-frequency resource grid before the starting time of the window, and sends a transmission window confirmation signaling to the corresponding edge agent alliance, wherein the transmission window confirmation signaling comprises a window parameter and a unique transaction identifier reserved at the time and is used for access verification during subsequent data uploading.
  8. 8. A computer system comprising a memory, a processor and a computer program stored on the memory, characterized in that the processor executes the computer program to carry out the steps of the method according to any one of claims 1-7.
  9. 9. A computer readable storage medium having stored thereon a computer program/instruction, which when executed by a processor, implements the steps of the method of any of claims 1-7.
  10. 10. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any of claims 1-7.

Description

Mine multisource data fusion transmission method and device based on distributed edge agent cooperation Technical Field The application relates to the field of data transmission, in particular to a mine multisource data fusion transmission method and device based on distributed edge agent cooperation. Background In the field of mine production, along with the development of digitization and intelligence, fusion transmission of multi-source data has important significance for improving the efficiency, safety and management level of mine production. In order to solve the problem of mine multisource data fusion transmission, the following methods are mainly adopted in the prior art. One is a centralized data processing and transmission mode, that is, all collected data is summarized to a central node, and then the central node processes and transmits the data. Another is a traditional distributed data processing approach that distributes data processing tasks across multiple nodes. The defects of the prior art are that the centralized processing mode can not meet the real-time processing and transmission requirements of a large amount of multi-source heterogeneous data of mines, and the data processing bottleneck is easy to cause. The traditional distributed processing mode lacks effective data marking and cooperative mechanism, so that errors and losses of data are easy to occur in the transmission and processing processes, and the data fusion effect is affected. In addition, although the use of the independent communication link can meet the basic requirement, the problems of resource waste, network congestion, high background data analysis complexity and the like exist, so that the waste of 5G frequency band resources is caused, the risk of network congestion is increased, and the complexity of cloud data analysis is increased due to the non-uniformity of different types of data transmission protocols, so that the high requirement of real-time decision making is difficult to meet. Disclosure of Invention The application provides a mine multi-source data fusion transmission method and device based on distributed edge agent cooperation, which effectively processes mine multi-source heterogeneous data, realizes cross-mode joint coding, reasonably predicts transmission requirements, realizes on-demand distribution and dynamic scheduling of network resources, ensures real-time performance of key data and reduces network congestion risk. In a first aspect, the application provides a mine multisource data fusion transmission method based on distributed edge agent cooperation, which comprises the following steps: Collecting heterogeneous data through a plurality of data acquisition terminals deployed in a mine production environment, generating a value mark for the heterogeneous data through a 5G fusion gateway based on equipment types of data sources and a preset value rule base, calculating a hash value of the heterogeneous data based on the value mark to serve as a digital fingerprint, generating a data certificate based on the digital fingerprint, and broadcasting the data certificate to a distributed account book commonly maintained by a plurality of edge computing nodes deployed in the same area; Determining an edge computing node which verifies and records the data certificate in the distributed account book as a master agent, resolving the value mark by the master agent, splitting a data processing task into task sub-pieces with different processing types, broadcasting bidding information containing resource requirements of the task sub-pieces to other edge computing nodes, selecting a slave agent from bidding results of the other edge computing nodes on the bidding information, and constructing the master agent and the slave agent into an edge agent alliance for processing the heterogeneous data; in the edge agent alliance, each slave agent builds a data processing pipeline according to the logic sequence of the task sub-slices, and the slave agents processing different modal data perform real-time interaction of event characteristics in the pipeline processing process and execute cross-modal joint coding on the associated heterogeneous data streams based on the event characteristics; And predicting the transmission requirement of the data packet to be uploaded based on the locally stored historical task log by the edge agent alliance, generating a transmission request suggestion packet containing predicted flow, priority and suggested transmission parameters, sending the transmission request suggestion packet to a 5G base station, receiving a transmission window returned by the 5G base station, and aggregating and uploading the jointly encoded data to a cloud in the transmission window. By adopting the technical scheme, the source traceability and ownership verifiability of heterogeneous data are ensured through the value marking, the digital fingerprint and the distributed account book mechanism, the dat