A robot distributed cooperative work data collection management system and method

Abstract

This invention relates to the field of data processing technology and discloses a robot distributed collaborative operation data acquisition and management system and method. The system includes: a data acquisition module to acquire node operating status; a local communication module to aggregate and generate a local state set; an edge perception computing module to calculate a collaborative tensor, and when a threshold is reached, to perform state interpolation alignment, aggregating and generating spatial topological features and multi-entity aggregated transactions; a gateway routing control module to use spatial topological features to calculate and predict the coordinate domain, and to modify the logical routing mapping table in memory when crossing boundaries; a logical mapping module for gateway addressing and querying; data storage nodes to write multi-entity aggregated transactions; and an index building module to extract time anchors and topological hash strings and concatenate them to generate a composite primary key. This invention solves the problems of difficult multi-node data alignment, cross-grid routing addressing delay, and difficulty in retrieving collaborative historical data.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a data acquisition and management system and method for distributed collaborative operations of robots. Background Technology

[0002] In multi-robot industrial manufacturing and warehousing logistics applications, data acquisition and management for collaborative multi-robot operations are fundamental to system scheduling. Existing robot data acquisition solutions often employ a model where each node independently reports and writes data to the server. However, when multiple robots are engaged in close-range collaborative operations, differences in the system clocks of individual nodes and communication delays in wireless local area networks make it difficult for the spatial status data received by the server from each node to be aligned in terms of time reference. This independent reporting mode can lead to concurrent write conflicts during intensive multi-node operations, causing the system to be unable to obtain the collaborative status of the robot group within a local area.

[0003] To manage continuous physical work areas, existing management systems typically divide the space into grids and assign corresponding routes and storage nodes to different grids. Robots, while moving within the physical space, will cross the boundaries of these grids. Existing routing mechanisms are mostly passively triggered; route reallocation and storage node switching only occur when the system detects a robot crossing a boundary and entering a new grid. This cross-domain switching process introduces addressing delays when multiple entities are moving simultaneously, leading to interruptions in data communication.

[0004] In terms of persistent data storage, conventional solutions typically use a single device identifier or timestamp as the database primary key. For multi-robot collaborative operation scenarios, historical data backtracking depends on the time dimension of individual machines and also requires reconstructing the spatial distribution characteristics of the group at that time. Existing technologies fail to establish a binding relationship between time state and spatial topology in the underlying data structure. This leads to increased computational consumption when performing collaborative scenario backtracking and historical state-oriented retrieval, requiring the system to perform cross-table queries, making it difficult to form a closed-loop storage structure for historical retrieval data. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a robot distributed collaborative operation data acquisition and management system and method, which solves the technical problems of difficulty in aligning local spatial state data, concurrent write conflicts, and high cross-grid routing addressing latency in existing robots during multi-node collaborative operations.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The first aspect of this invention provides a robot distributed collaborative operation data acquisition and management system, comprising: The data acquisition module is used to acquire the position coordinates and running speed of the mobile robot node in the physical work site and output the basic operating status. The local communication module is used to broadcast the basic operating status in the local area network and aggregate the broadcast information of neighboring nodes to form a local state set. An edge-aware computing module is used to calculate a collaborative tensor based on the local state set and generate multi-entity aggregated transactions and associated spatial topological features through time-window aggregation. The gateway routing control module, which is deployed in the routing gateway server, is used to receive the spatial topology features and modify the logical routing mapping table in memory to output the updated routing mapping table. The logical mapping module is used to reside in and maintain the updated routing mapping table for gateway addressing queries during data flow. The data storage node, deployed in the database cluster, is used to receive and write the multi-entity aggregation transaction according to the addressing and routing instructions of the updated routing mapping table; The index building module is used to extract the internal feature fields of the written multi-entity aggregation transaction and concatenate them to generate a composite primary key to form an index structure that can be retrieved historically.

[0007] Preferably, the gateway routing control module extracts the feature boundaries of the physical work site to generate a discrete reference point set; the gateway routing control module performs polygon subdivision on the continuous physical space based on the discrete reference point set and divides it into two-dimensional spatial grid identifiers; the gateway routing control module uses a consistent hashing algorithm to construct the initial mapping allocation relationship between the two-dimensional spatial grid identifiers and the data storage nodes, so that the logical mapping module in the system instantiates and generates and resides in the logical routing mapping table in memory.

[0008] Preferably, when the edge-aware computing module calculates the collaborative tensor, it performs a weighted summation of the spatial exclusion degree component generated by the relative distance and the motion interference degree component generated by the relative velocity to quantitatively assess the overall interference risk level in the local area where the current node is located.

[0009] Preferably, when the calculated cooperative tensor reaches a preset interference threshold, the edge-aware computing module generates an internal blocking control command to suspend the current discrete data direct writing mode and triggers the local communication module to perform a dynamic handshake to build a data synchronization group; within the data synchronization group, node comparison is performed to elect a master node; the master node extracts its local system clock to generate a relative time anchor to avoid global clock synchronization errors, and sends the relative time anchor to all slave nodes within the data synchronization group; wherein, the preset interference threshold is comprehensively calibrated by combining the spatial safety tolerance of the physical work site and the node communication delay bandwidth.

[0010] Preferably, the edge-aware computing module performs a state interpolation alignment algorithm based on the relative time anchor point to uniformly interpolate and align the acquisition states of the remaining nodes in the data synchronization group to the relative time anchor point, thereby obtaining the alignment position coordinates and alignment running speed. The edge-aware computing module calculates and generates a topological momentum vector based on the alignment running speed. During the calculation, the motion vectors of all nodes in the data synchronization group are linearly superimposed according to the corresponding virtual mass weights to reflect the comprehensive movement direction and speed of the entity group within the local risk area.

[0011] Preferably, the edge-aware computing module constructs a relatively spatially connected graph reflecting spatial distribution relationships and generates a corresponding adjacency matrix, using the alignment coordinates of each node as graph vertices and the relative distances between nodes as graph edges; the edge-aware computing module flattens the adjacency matrix into a one-dimensional numerical sequence and performs a hash operation on the one-dimensional numerical sequence to generate a topological hash string; the edge-aware computing module combines and encapsulates the previously generated topological momentum vector with the topological hash string to generate the spatial topological features.

[0012] Preferably, the routing gateway server receives the spatial topology features and extracts the topological momentum vector from the spatial topology features; the gateway routing control module calculates the predicted centroid coordinates of the group based on the topological momentum vector, the calculation process is based on the mass-weighted average to derive the initial centroid coordinates of the group, and the overall position offset generated within the prediction time step driven by the topological momentum vector; the gateway routing control module constructs a circular envelope region centered on the predicted centroid coordinates of the group as the prediction coordinate domain for the next period.

[0013] Preferably, the gateway routing control module executes spatial conflict determination logic for crossing the current grid boundary and performs a spatial intersection test between the geometric contour of the next-cycle predicted coordinate domain and the preset grid boundary within the system. When it is determined that the next-cycle predicted coordinate domain crosses the current grid boundary, the gateway routing control module extracts the set of local spatial coordinates where the next-cycle predicted coordinate domain overlaps with adjacent grids. The gateway routing control module dynamically generates high-priority temporary addressing entries for the set of local spatial coordinates in the logical routing mapping table, and accordingly modifies the logical routing mapping table at the memory level, temporarily mapping and associating the expected adjacent grid local areas to the data storage node accessed by the current master node at the logical addressing level, and finally outputs the updated routing mapping table.

[0014] Preferably, the master node in the system serializes and encapsulates the aligned data to generate the multi-entity aggregated transaction and transmits it upwards; the data storage node in the system receives and completes the writing of the multi-entity aggregated transaction according to the addressing and routing instructions of the updated routing mapping table; the index building module extracts the multi-entity aggregated transaction that has been written to disk, and obtains the topology hash string and the relative time anchor point contained in the feature fields of the multi-entity aggregated transaction, concatenates them, uses the relative time anchor point representing the time dimension as a prefix, uses the topology hash string representing the spatial structure dimension as a suffix, and introduces a preset text separator in the middle to connect them, thereby generating the composite primary key.

[0015] A second aspect of the present invention provides a method for data acquisition and management of distributed collaborative operations of robots, comprising the following steps: The gateway routing control module extracts the feature boundaries of the physical work site to divide a two-dimensional spatial grid identifier and stores the initial logical routing mapping table in memory; The data acquisition module obtains the position coordinates and running speed of the mobile robot node in the physical work site and outputs the basic operating status. The local communication module broadcasts the basic operating status in the local area network and aggregates the broadcast information of neighboring nodes to form a local state set; The edge-aware computing module calculates a collaborative tensor based on the local state set. When the collaborative tensor reaches a preset interference threshold, it triggers data synchronization interpolation and alignment. It then uses the aligned data to generate a topological momentum vector and a topological hash string, which are then aggregated through a time window to generate multi-entity aggregated transactions and associated spatial topological features. The gateway routing control module receives the spatial topology features, uses the topological momentum vector in the spatial topology features to calculate the next cycle prediction coordinate domain, and modifies the logical routing mapping table in memory to output the updated routing mapping table. The logical mapping module resides in and maintains the updated routing mapping table for gateway addressing queries during data flow; The data storage node receives and writes the multi-entity aggregation transaction according to the addressing and routing instructions of the updated routing mapping table; The index building module extracts the relative time anchor and topological hash string from the internal feature fields of the multi-entity aggregated transaction that have been written, and concatenates them to generate a composite primary key to form an index structure that can be retrieved historically. The preset interference threshold is comprehensively calibrated by combining the spatial safety tolerance of the physical work site with the node communication delay bandwidth.

[0016] This invention provides a data acquisition and management system and method for distributed collaborative robot operations. It has the following beneficial effects: 1. This invention calculates a collaborative tensor using an edge-aware computing module. When a preset interference threshold is reached in a local area, the independent writing of each node is stopped, triggering the construction of a data synchronization group and the master node issuing a relative time anchor. Based on this relative time anchor, the system performs interpolation alignment on the collected states of each node, and then generates a multi-entity aggregated transaction for unified uploading. This solves the problems of spatial state data alignment difficulties and concurrent write conflicts caused by clock differences and network latency in multi-node collaborative operations.

[0017] 2. This invention utilizes a gateway routing control module to extract the topological momentum vector from spatial topological features, calculates the next-cycle predicted coordinate domain of the group, and performs a spatial intersection test between the contour of this coordinate domain and the grid boundary. When it is determined that the predicted coordinate domain crosses the grid boundary, the system dynamically generates a high-priority temporary addressing entry in the logical routing mapping table in memory, enabling the data expected to cross the domain to maintain association and be mapped to the original data storage node. This solves the problems of high routing latency and data communication interruption caused by the robot crossing spatial grid boundaries.

[0018] 3. This invention utilizes an index building module to extract internal features of written multi-entity aggregated transactions, concatenating relative time anchors representing the time dimension with topological hash strings representing the spatial distribution structure to generate a composite primary key. This mechanism associates the temporal states and spatial topological relationships of multiple robot groups in the underlying data structure, providing fundamental support for targeted retrieval of historical data in the system and forming a data processing flow from collection and alignment to storage and traceability. Attached Figure Description

[0019] Figure 1 This is a structural block diagram of the robot distributed collaborative operation data acquisition and management system of the present invention; Figure 2 This is a logical architecture diagram of the robot distributed collaborative operation data acquisition and management system of the present invention; Figure 3 This is a flowchart of the robot distributed collaborative operation data acquisition and management method of the present invention; Figure 4 This is a hardware structure block diagram of the electronic device of the present invention; Figure 5 This is a comparison chart showing the cross-node transaction lock conflict rate as a function of the number of nodes, according to the present invention. Figure 6 This is a comparison chart showing how the data write latency of this invention varies with the number of nodes. Detailed Implementation

[0020] The technical solutions in 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] See attached document Figure 1 This invention provides a robot distributed collaborative operation data acquisition and management system and method. The robot distributed collaborative operation data acquisition and management system includes: The data acquisition module is used to acquire the position coordinates and running speed of the mobile robot node in the physical work site and output the basic operating status. The local communication module is used to broadcast the basic operating status in the local area network and summarize the broadcast information of neighboring nodes to form a local state set. The edge-aware computing module is used to calculate the collaborative tensor based on the local state set and generate multi-entity aggregated transactions and associated spatial topological features through time window aggregation. The gateway routing control module, deployed on the routing gateway server, is used to receive spatial topology features and modify the logical routing mapping table in memory to output the updated routing mapping table. The logical mapping module is used to reside in and maintain the updated routing mapping table for gateway addressing queries during data flow. Data storage nodes, deployed in the database cluster, are used to receive and write multi-entity aggregation transactions based on the addressing and routing instructions of the updated routing mapping table. The index building module is used to extract the internal feature fields of the written multi-entity aggregated transactions and concatenate them to generate a composite primary key to form an index structure that can be retrieved historically.

[0022] See attached document Figure 2 At the logical functional level, the robot distributed collaborative operation data acquisition and management system is divided into an edge-aware computing layer, a dynamic gateway routing layer, and a distributed persistent storage layer. The edge-aware computing layer is responsible for performing state interaction, synchronization group construction, temporal aggregation and packaging, and spatial topology feature extraction operations; the dynamic gateway routing layer is responsible for maintaining the logical addressing mapping table between the physical space grid and data storage nodes and performing memory-level routing changes based on predicted trajectories; the distributed persistent storage layer is responsible for receiving aggregated data blocks, performing disk writes, and constructing composite primary keys to support historical retrieval tasks.

[0023] See attached document Figure 3 This invention provides a method for data acquisition and management of distributed collaborative robot operations, comprising the following steps: S1, the system initializes logical routing. The gateway routing control module obtains the two-dimensional spatial grid identifiers of the physical work site feature boundary and builds a logical routing mapping table in memory corresponding to the spatial grid identifiers and the underlying data storage nodes. S2, edge node state broadcasting and tensor calculation: the mobile robot node collects position coordinates and running speed and broadcasts them in the local area network. The edge perception calculation module calculates the cooperative tensor between entities in the local space based on the position coordinates and running speed. S3, Data synchronization group negotiation and anchor point extraction: When the calculated cooperative tensor reaches the preset interference threshold condition, multiple mobile robot nodes in the local area stop the discrete data direct writing mode and build a data synchronization group in the local area network. The master node elected in the data synchronization group extracts the system clock record of the local machine as the relative time anchor point. S4, relative time window alignment and topology feature extraction: the master node uniformly interpolates and aligns the acquisition status of the other nodes in the data synchronization group to the relative time anchor point, and calculates and generates the topology momentum vector and topology hash string based on the aligned position coordinates and running speed. S5, based on momentum-based memory-level route drift, the routing gateway server receives the topological momentum vector and calculates the predicted coordinate domain for the next cycle. When it is determined that the predicted coordinate domain crosses the current grid boundary, the gateway routing control module dynamically modifies the logical route mapping table in memory to map and associate the expected adjacent grid local area to the data storage node accessed by the current master node. S6, aggregated transaction persistence and composite index construction: the master node encapsulates the aligned data into a multi-entity aggregated transaction and uploads it. The routing gateway server forwards the multi-entity aggregated transaction to the corresponding data storage node according to the modified logical routing mapping table. The index building module synchronously generates a composite primary key based on the topological hash string and the relative time anchor point to complete the storage loop.

[0024] To fully explain the technical details and implementation logic of this invention, the following provides a detailed description of each core working step in the robot distributed collaborative operation data acquisition and management method and its execution mechanism in specific system modules.

[0025] In this embodiment, regarding step S1, the system initializes logical routing. The gateway routing control module obtains the two-dimensional spatial grid identifiers representing the physical work site's characteristic boundaries and constructs a logical routing mapping table in memory corresponding to the spatial grid identifiers and the underlying data storage nodes. To build the basic addressing architecture supporting subsequent high-concurrency writes, this process is specifically implemented through the following sub-steps: S101, the gateway routing control module extracts the feature boundaries of the physical work site to generate a discrete set of reference points. At the basic principle level, since mobile robot nodes are usually distributed in a continuous and unbounded physical space, if the underlying data is sliced ​​directly using the original coordinates, it is easy to cause a lot of cross-shard lock contention in the distributed database. Therefore, the system needs to transform the disordered continuous physical space into discrete blocks that can be accurately addressed by the gateway logic layer.

[0026] As a preferred approach, the gateway routing control module reads a pre-stored global two-dimensional environment map and extracts the coordinates of the static obstacle outlines and the extreme physical boundary coordinates of walkable areas. Based on this, by integrating the aforementioned edge data, the gateway routing control module extracts a discrete set of reference points by combining the coordinates of the channel intersections and shelf endpoints, according to the coordinates of the static obstacle outlines and the extreme physical boundary coordinates of walkable areas.

[0027] S102, the gateway routing control module performs polygon subdivision on the continuous physical space based on the discrete reference point set and divides it into two-dimensional spatial grid identifiers. The gateway routing control module uses the two-dimensional Thiessen polygon algorithm to perform spatial gridding on the physical work site containing the discrete reference point set. In this partitioning operation, the gateway routing control module divides the globally continuous two-dimensional physical space into multiple non-overlapping Thiessen polygon regions. This partitioning logic ensures that for any coordinate within any Thiessen polygon region, the Euclidean distance to the corresponding reference point in that region is less than the Euclidean distance to any other reference point in the discrete reference point set.

[0028] For the specific geometric plane segmentation and point set bounding box generation calculations of the Thiessen polygon algorithm, those skilled in the art can implement it using existing Delaunay triangulation dual logic. The underlying geometric segmentation calculation principle is well-known in the field and will not be elaborated upon here. To facilitate logical addressing management, the gateway routing control module assigns a corresponding string number to each segmented Thiessen polygon region. This string number constitutes the two-dimensional spatial grid identifier defined in this invention.

[0029] S103, the gateway routing control module establishes the initial mapping and allocation relationship between two-dimensional spatial grid identifiers and data storage nodes. In the distributed architecture of this embodiment, the distributed persistent storage layer within the database cluster is configured with multiple independently deployed data storage nodes. The gateway routing control module uses a consistent hashing algorithm, taking each global two-dimensional spatial grid identifier as an input key, to calculate the hash value slot corresponding to each two-dimensional spatial grid identifier.

[0030] Based on the calculation results, the gateway routing control module assigns a corresponding physical connection address to each two-dimensional spatial grid identifier according to the ring topology relationship between the hash value slot and the data storage node. To ensure the integrity and fault tolerance of the algorithm logic when fluctuations occur in the underlying storage network, when any data storage node is detected to be offline or added / expanded, the gateway routing control module only performs local remapping calculations on the two-dimensional spatial grid identifiers of adjacent slots on the consistent hash ring, without triggering a global mapping address reallocation. This physical connection address includes the network interconnection protocol address and access connection port of the corresponding data storage node.

[0031] S104, the logical mapping module instantiates and maintains a logical routing mapping table in memory. To achieve high-speed data routing distribution, the gateway routing control module declares and initializes a key-value dictionary data structure in the memory space of the routing gateway server. The gateway routing control module uses the two-dimensional spatial grid identifier as the key field of the dictionary structure and the physical connection address of the corresponding data storage node as the value field, and writes them in batches into the key-value dictionary data structure. After completing this batch write operation, the key-value dictionary data structure is instantiated as the logical routing mapping table.

[0032] The logical routing mapping table is detached from the physical disk storage medium and resides in the resident memory area allocated by the logical mapping module. This memory-resident mechanism allows the gateway server to complete the underlying storage addressing and distribution operations through memory queries when subsequent data write requests from the edge side arrive at the routing gateway server, thereby reducing the additional latency caused by disk read and write.

[0033] In this embodiment, regarding the aforementioned step S2, edge node state broadcasting and tensor calculation, the mobile robot node collects position coordinates and running speed and broadcasts them in the local area network, and the edge perception calculation module calculates the cooperative tensor between entities in the local space based on the position coordinates and running speed.

[0034] To achieve state perception and risk quantification within a local area, this process is implemented through the following sub-steps: S201, the data acquisition module obtains the position coordinates and running speed of the mobile robot node in the physical work area and outputs its basic operating status. During the execution of distributed tasks, the data acquisition module reads data from the vehicle-mounted sensors at set time intervals. As a preferred method, the data acquisition module uses the fusion data of LiDAR and wheeled odometer to calculate its current two-dimensional planar spatial coordinates and uses them as its position coordinates; Simultaneously, the current linear velocity and angular velocity are calculated based on the wheel speed data fed back by the motor driver, and used as the running speed. For the method of pose estimation through multi-sensor data fusion, those skilled in the art can implement it using existing extended Kalman filter algorithms; its underlying derivation logic is well-known in the field and will not be elaborated here. Combining the information read above, the data acquisition module serializes and encapsulates the position coordinates and running speed along with the current system timestamp, thereby generating the basic running state corresponding to the current cycle.

[0035] S202, the local communication module broadcasts its basic operating status within the local area network and aggregates the broadcast information from neighboring nodes to form a local state set. Based on the individual operating data acquired above, in order to further understand the dynamic changes of surrounding entities, the local communication module sends data packets containing its own basic operating status to all directions at a preset broadcast frequency through the vehicle's wireless communication hardware. Combined with the setting of signal transmission characteristics, this local area network physically defines an effective communication coverage area.

[0036] Using this network mechanism, the local communication module monitors the channel within its communication range in real time, receiving status data packets from other mobile robot nodes. After receiving the data, the local communication module parses and caches the received status data packets belonging to different neighboring nodes, thereby forming a local state set reflecting the dynamic operation of multiple entities within the local spatial environment.

[0037] S203, the edge-aware computing module calculates the cooperative tensor between entities in the local space based on the local state set. After collecting the local states, the system needs to further evaluate the collision and interference risks. In the physical principle of multi-robot swarm operations, evaluating whether two robots interfere depends not only on the spatial distance between them but also on their relative motion state. If two nodes are close but run in the same direction and at similar speeds, their relative motion amplitude is small, and the interference risk is relatively low; if two nodes are running in opposite directions, crossing directions, or with a large speed difference, their interference risk is relatively high. Based on the above considerations, this embodiment combines relative position and relative speed difference to quantify the interference risk. The edge-aware computing module reads various parameters from the local state set. To avoid invalid traversal when the local state set is empty, when the local state set is empty, the edge-aware computing module directly assigns the cooperative tensor of the current node to zero; when the local state set is not empty, the edge-aware computing module calculates the cooperative tensor using the following cooperative tensor calculation formula: ; in: Represents a node The collaborative tensor; Indicates that it is located at node The set of neighboring nodes within the communication range; This represents the index of a neighbor node in the set of neighbor nodes; The summation operator represents the summation of all neighbor nodes in the set of neighbor nodes; This represents the position coordinates of the current node in two-dimensional vector form, obtained by the data acquisition module. Represents the neighbor nodes extracted from the local state set. The neighbor's location coordinates in two-dimensional vector form; This represents the Euclidean distance between the current node's coordinates and the coordinates of its neighbors. This represents the squared value of the Euclidean distance. This represents a preset positive real constant used to prevent calculation errors caused by a zero denominator when the distance is too close; its value range is set between 0.001 and 0.01. This represents the spatial displacement component calculated based on the inverse distance relationship; This represents the distance attenuation weight that adjusts the magnitude of the spatial displacement component. Its value is greater than zero and is calibrated according to the safety distance standards of the physical work site. This represents the running speed in the form of a two-dimensional vector of the current node, obtained by the data acquisition module. Represents the neighbor nodes extracted from the local state set. The speed of neighbor operation in two-dimensional vector form; This represents the square of the relative speed difference between the current node and its neighboring nodes; This represents the motion interference component calculated based on the relative velocity difference; The speed weight represents the adjustment of the magnitude of the motion interference component. Its value is greater than zero and is calibrated in conjunction with the node's maximum design speed.

[0038] The above formula, by weighting and summing the spatial exclusion component generated by relative distance and the motion interference component generated by relative velocity difference, can quantitatively assess the overall interference risk level within the local area where the current node is located. When two nodes are moving in similar directions and at similar speeds, the relative velocity difference is small, and the motion interference component is small; when two nodes are moving in opposite directions, crossing each other, or with a large velocity difference, the relative velocity difference increases, and the motion interference component increases accordingly, thereby increasing the node's interference risk. The collaborative tensor is calculated by the edge-aware computing module and serves as the core data basis for the system to determine whether to trigger the collaborative networking management logic in the subsequent step S3.

[0039] The above formula, by weighting and summing the spatial exclusion component generated by relative distance and the motion interference component generated by relative velocity, can quantitatively assess the overall interference risk level within the local area where the current node is located. In the operation logic of the velocity inner product, when the motion directions of two nodes tend to be the same, the inner product value is positive, which increases the cooperative tensor of node i; when the motion directions of two nodes tend to be perpendicular or opposite, the inner product value decreases or becomes negative, thereby reducing the cooperative tensor of node i. The edge-aware computing module outputs the calculated cooperative tensor, which serves as the core data basis for the system to determine whether to trigger the cooperative networking management logic in subsequent step S3.

[0040] In this embodiment, based on the state data obtained in the preceding steps, the system proceeds to step S3: data synchronization group negotiation and anchor point extraction. When the calculated cooperative tensor reaches a preset interference threshold, multiple mobile robot nodes within the local area cease discrete data direct writing mode and construct a data synchronization group within the local area network. The master node elected within the data synchronization group extracts its own system clock record as the relative time anchor point. To ensure data consistency in high-density collaborative scenarios, this process is specifically implemented through the following sub-steps: S301, the edge-aware computing module performs interference threshold condition determination and terminates the discrete data direct write mode. In the underlying data interaction principle of a multi-agent system, when multiple nodes are in a close-range intensive operation state, the state data they generate has a lot of spatial overlap. If they continue to maintain independent asynchronous writing, it is easy to cause a lot of read-write lock conflicts and timing alignment errors at the underlying layer of the distributed database. Based on this consideration, the edge-aware computing module extracts the cooperative tensor obtained from the previous calculation and compares it with the preset interference threshold set internally by the system.

[0041] As a preferred approach, the preset interference threshold is comprehensively calibrated by combining the spatial safety tolerance of the physical work site with the node communication delay bandwidth, with a specific value range set between 1.5 and 3.0. When the cooperative tensor is determined to be less than the preset interference threshold, the edge perception computing module maintains the current discrete data direct writing mode of the mobile robot nodes, allowing each node to independently send its own status data to the database cluster through the routing gateway server. Conversely, when the cooperative tensor is determined to be greater than or equal to the preset interference threshold, it indicates a high risk of spatial interference between entities in the local area. The edge perception computing module generates an internal blocking control command, suspends the current discrete data direct writing mode, and pauses the forwarding of independent status data to the external network.

[0042] In step S302, the local communication module performs a dynamic handshake within the local area network to establish a data synchronization group and elect a master node. After completing the aforementioned interrupt command, to establish a unified local coordination mechanism, the local communication module initiates a multi-node networking process within the local area network. Mobile robot nodes exceeding a preset interference threshold send networking request data packets carrying their own device identifier and coordination tensor values ​​via broadcast or multicast through the local communication module. Neighboring nodes receiving the networking request data packet return confirmation information via unicast response. The master node candidate nodes determine the scope of synchronization group members based on the received confirmation information, thereby forming a logically closed data synchronization group within the physical collision risk area.

[0043] After node discovery is complete, reliable transmission sessions are established between nodes within the data synchronization group to transmit time anchors, alignment status data, and control commands. As a preferred approach, the node discovery phase employs a broadcast discovery and unicast response mechanism, while the session establishment phase utilizes TCP connections or other reliable transmission protocols. The underlying communication connection principles are standard techniques in this field and will not be elaborated upon here.

[0044] To avoid prolonged waiting times due to network fluctuations or physical obstructions at individual nodes during network formation, the local communication module includes a timeout timestamp when initiating a network formation request. If no response is received within a preset time limit, the node is automatically considered to have failed to form a network and is skipped. After the data synchronization group is established, the nodes within the group interact and compare their respective collaboration tensor values. Based on the comparison results, the system sorts the nodes within the group in descending order of their collaboration tensor values, designating the mobile robot node with the largest collaboration tensor value as the master node, and automatically identifying the remaining nodes as slave nodes. If equal values ​​are found during the comparison process, the lexicographical order of the device identifier codes is further compared to determine the master node's affiliation.

[0045] S303, the master node extracts its local system clock to generate a relative time anchor to avoid global clock synchronization errors. After establishing the master node role, the system needs to establish a unified timing reference benchmark for the data synchronization group. In this embodiment, in a conventional distributed cluster architecture, devices typically rely on a global network time protocol for clock synchronization, but wireless transmission delay jitter in mobile network environments can easily cause local synchronization errors to accumulate. To avoid such errors, this embodiment adopts a local relative time reference mechanism.

[0046] Specifically, the processing unit inside the master node directly reads the current count value of the local hardware clock and records the local system timestamp at that instant as a relative time anchor. After extraction, the local communication module encapsulates the relative time anchor into a synchronization command and sends it to all slave nodes in the data synchronization group. Meanwhile, to prevent the master node from unexpectedly crashing before sending the synchronization command, causing slave nodes to wait for an extended period, the system configures a heartbeat detection and timeout reset mechanism for the slave node's listening module. If the relative time anchor is not received within a given heartbeat period, the data synchronization group is disbanded and a new network election is triggered. This time anchoring mechanism does not require slave nodes to tamper with or calibrate their own hardware system clocks; instead, it provides a unified time reference system for the local network, ensuring that subsequent smooth alignment and interpolation operations can be performed on heterogeneous state data generated by differences in sampling frequencies between different nodes under this unified timing reference.

[0047] In this embodiment, after completing network negotiation and extracting the time reference base, the system proceeds to step S4, which involves relative time window alignment and topology feature extraction. The master node uniformly interpolates and aligns the acquisition status of the remaining nodes in the data synchronization group to the relative time anchor point, and calculates and generates a topology momentum vector and a topology hash string based on the aligned position coordinates and running speed. To eliminate data deviations caused by asynchronous sampling and extract macroscopic features of multiple nodes, this process is specifically implemented through the following sub-steps: S401, the edge perception computing module performs a state interpolation alignment algorithm based on relative time anchor points. Because each mobile robot node's internal data acquisition module has an independent operating frequency and sampling clock, the timestamps corresponding to the position coordinates and running speeds uploaded from each slave node to the master node are different. Due to differences in the clock crystal oscillators of the underlying physical sensor hardware and random delays in the local area network communication link, there are timing deviations in the time slices of the actual state data captured by each mobile robot node. If these are directly superimposed and fused in space without processing, it will lead to errors in the spatial position of entities in motion.

[0048] To address the error issues caused by asynchronous sampling, the edge-aware computing module within the master node uses the relative time anchor point extracted in step S3 as a unified time reference. For any node within the data synchronization group, the edge-aware computing module extracts the two most recent historical sampling data points of that node before and after the relative time anchor point from the cache pool. Based on this, the edge-aware computing module uses a linear interpolation algorithm to calculate the fitted state parameters of the node at the relative time anchor point, using these parameters as the node's alignment position coordinates and alignment running speed, respectively.

[0049] To avoid algorithm dead zone errors caused by insufficient historical data due to network outages or cache clearing, if the edge-aware computing module retrieves only one historical sampling data point in the cache pool, it directly uses that single data point as the fitting state parameter. In a more extreme case, if no historical sampling data point is retrieved, the node's data is deemed invalid and removed from the data synchronization group. For linear interpolation algorithms that represent one-dimensional data changing over time, those skilled in the art can implement them using existing linear interpolation mathematical formulas. The interpolation calculation principle is a well-known technology in this field and will not be elaborated here.

[0050] S402, the edge-aware computing module calculates the topological momentum vector based on alignment state parameters. In the principle of physical collision risk prediction for cluster systems, the overall impact damage potential depends not only on the travel speed but also on the direct constraint of the entity's mass. Combining this physical characteristic, introducing mass weights to calculate momentum can more objectively restore the physical motion trend of the local cluster.

[0051] As a preferred approach, to characterize the overall macroscopic motion trend of multiple entities within a local region, the edge perception computing module combines the results of the aforementioned interpolation calculations and uses the topological momentum calculation formula for computation. The topological momentum calculation formula is as follows: ; in, This represents the topological momentum vector of the data synchronization group; This represents the set of aligned nodes within the data synchronization group, including the master node and all slave nodes. Indicates the node index in the alignment node set; This is a summation operator that iterates through and sums all nodes in the aligned node set. Its value is set based on the sum of the mobile robot node's base weight and the physical mass of the currently loaded cargo, with a range of 50 to 500 kilograms, representing the weight allocated to the node. The virtual mass weight; where the node's basic self-weight can be the equipment's factory configuration parameters, the physical mass of the currently loaded cargo can be measured by the vehicle-mounted weighing sensor, or obtained from the load parameters issued by the task management system; when real-time weighing data is unavailable, the nominal load mass recorded in the task order can be used as a substitute value; Represents a node Alignment speed obtained from interpolation calculations in the form of a two-dimensional vector; This represents the product of the virtual quality weight and the alignment speed, i.e., the node's... The individual momentum component.

[0052] This formula linearly superimposes the motion vectors of all nodes within the group according to their mass weights. The topological momentum vector calculated by the above formula can reflect the comprehensive direction and speed of the entity group within the local risk area, thus providing basic data input support for subsequent gateway route prediction.

[0053] S403, the edge-aware computing module performs dimensionality reduction on the relative spatial connectivity graph and generates a topological hash string. While recording the movement trend of the multi-machine group, recording the relative spatial structure at the corresponding time provides a data foundation for subsequent historical retrieval and state backtracking. In this embodiment, the edge-aware computing module constructs a relative spatial connectivity graph reflecting spatial distribution relationships, using the alignment coordinates of each node as graph vertices and the relative distances between nodes as graph edges.

[0054] To ensure stable and consistent hash results under the same spatial topology, the edge-aware computing module first performs a fixed sorting of nodes within the data synchronization group according to preset rules before generating the adjacency matrix. As a preferred method, the fixed sorting rule is to arrange the nodes in ascending order by their device identifier codes; alternatively, it can be arranged in ascending order by the polar angle of each node relative to the group centroid. After the node sorting is completed, the edge-aware computing module generates the corresponding adjacency matrix according to the fixed order. In the judgment logic for constructing the adjacency matrix, if the distance between two nodes is less than the preset connection distance, the corresponding element in the matrix is ​​recorded as 1; otherwise, it is recorded as 0. The preset connection distance is set to a range of 0.5 meters to 2.0 meters, determined by the outer contour envelope radius of the mobile robot node and the safety buffer distance. After the matrix is ​​constructed, the edge-aware computing module flattens the adjacency matrix into a one-dimensional numerical sequence in row-major order.

[0055] Based on the extraction results of the aforementioned one-dimensional numerical sequence, the edge-aware computing module uses the SHA-256 hash algorithm to perform a hash operation on the one-dimensional numerical sequence, generating a fixed-length string of characters, which is used as the topological hash string. The specific implementation process of the SHA-256 hash algorithm can be implemented by those skilled in the art based on existing publicly available algorithms, and will not be elaborated here. The generated topological hash string can correspond to and represent the spatial topological arrangement state of the current data synchronization group at a relative time anchor point, and provides structured feature support for subsequent composite primary key construction and historical retrieval.

[0056] In this embodiment, following the extraction results of spatial topology features, the system proceeds to step S5, which involves memory-level route drift based on momentum. The routing gateway server receives the topological momentum vector and calculates the predicted coordinate domain for the next cycle. When it is determined that the predicted coordinate domain crosses the current grid boundary, the gateway routing control module dynamically modifies the logical routing mapping table in memory to map and associate the expected adjacent grid local regions to the data storage node accessed by the current master node. This memory-level dynamic drift mechanism aims to address the distributed data write association problem when cluster entities cross grids, and is specifically implemented through the following sub-steps: In S501, the routing gateway server receives the topological momentum vector and calculates the predicted coordinate domain for the next cycle. During the data flow from the data synchronization group to the external network, the routing gateway server extracts the received topological momentum vector. In physical scenarios of multi-agent collaborative operations, groups of entities in close formation or with high interference risks often exhibit macroscopic cluster inertia. Projecting their spatial distribution over future time periods based on this cluster inertia can provide preliminary judgment criteria for underlying routing rules.

[0057] Based on the above considerations, the gateway routing control module extracts the group coordinate distribution of the data synchronization group at the relative time anchor point. As a preferred method, the gateway routing control module uses the group centroid prediction formula to calculate the predicted centroid coordinates of the group in the next period. The group centroid prediction formula is as follows: ; in, This represents the predicted centroid coordinates of the group in two-dimensional vector form for the next period, calculated from the previous data. This represents the set of aligned nodes within the data synchronization group, including the master node and all slave nodes. Indicates the node index in the alignment node set; Indicates assignment to a node Virtual quality weights; Represents a node The alignment position coordinates are obtained in the form of a two-dimensional vector after interpolation calculation; This represents the initial centroid coordinates of the group at the current time anchor point, calculated using a mass-weighted average. This represents the topological momentum vector of the data synchronization group, calculated beforehand. This indicates the prediction time step set by the system for calculating the next cycle, and its value range is set between 1.0 second and 3.0 seconds; This represents the overall positional shift of the group as a whole within the prediction time step, driven by topological momentum.

[0058] Based on the calculated predicted centroid coordinates of the group, the gateway routing control module uses these predicted centroid coordinates as the geometric center and the straight-line distance from the farthest node within the data synchronization group to the centroid as the radius to construct a circular envelope region. This circular envelope region is defined as the prediction coordinate domain for the next cycle. To avoid multiple overlapping nodes within the data synchronization group causing the expansion radius to be zero and thus triggering a prediction dead zone, the system introduces a lower limit value for this envelope region. If the calculated radius is less than the preset minimum safety compensation radius, the circular envelope region is directly constructed using this minimum safety compensation radius, which ranges from 1.5 meters to 2.5 meters.

[0059] S502, the gateway routing control module executes the spatial conflict determination logic across the current grid boundary. During the system initialization logic routing phase, the physical work site has been divided into multiple physical spatial grids, each with its own grid boundary. Based on the aforementioned calculation results, the gateway routing control module performs a spatial intersection test between the geometric contour of the predicted coordinate domain and the preset grid boundary within the system.

[0060] When the predicted coordinate domain's envelope outline does not intersect with the current grid's boundary and is contained within the current grid, the gateway routing control module determines there is no cross-region behavior and maintains the existing routing forwarding rules. In another scenario, when the predicted coordinate domain's envelope outline intersects with the current grid's boundary, or when a portion of the region extends beyond the current grid boundary into adjacent grids, it indicates that within the future routing prediction time step, the activity range of nodes within the data synchronization group will enter the adjacent grid region. Since different grids correspond to different data storage nodes in the underlying logical mapping, if the cluster crosses boundaries, the associated state data generated by the same data synchronization group will be distributed and routed to different storage nodes, thus triggering cross-node synchronization lock contention in distributed transactions.

[0061] Based on the above phenomena, the gateway routing control module determines that a spatial crossing conflict exists and triggers a route drift operation. For the two-dimensional intersection determination calculation of the geometric envelope region and boundary line segments, those skilled in the art can use existing polygon clipping or line segment intersection detection algorithms. The underlying geometric calculation logic is well-known in the field and will not be elaborated upon here.

[0062] In S503, the gateway routing control module performs lossless, memory-level dynamic modifications to the logical routing mapping table. Upon determining a spatial crossover conflict, to ensure the cohesion of writing multi-entity aggregated data, the gateway routing control module does not alter the actual boundary division of the physical grid, but instead adjusts the data addressing logic in real-time at the routing layer. To address this adjustment requirement, the gateway routing control module extracts local regions where the predicted coordinate domain overlaps with adjacent grids and discretizes these local regions into local sub-region identifiers. After obtaining these local sub-region identifiers, the gateway routing control module locates the logical routing mapping structure residing in the logical mapping module's memory and establishes a high-priority temporary drift sub-table within it. The temporary drift sub-table uses the local sub-region identifier as the key and the physical connection address of the data storage node currently connected to the master node as the value.

[0063] During route lookups, the gateway routing control module prioritizes matching the temporary drift sub-table; if the temporary drift sub-table is not found, it then matches the default logical route mapping table. This allows for temporary redirection of predicted crossing areas without altering the mapping relationship between the default grid identifier and data storage nodes. This temporary redirection operation and the insertion of temporary entries are completed in the memory of the routing gateway server, without needing to refresh the underlying disk configuration file. To prevent routing table pollution caused by memory update conflicts under multi-threaded concurrency, the gateway routing control module uses atomic updates to write to the temporary drift sub-table. Through this logical route boundary adaptive drift mechanism, the expected adjacent grid local areas are temporarily associated with the data storage node corresponding to the current grid at the logical addressing level. This mechanism ensures that data uploaded by the data synchronization group during boundary crossings can be routed and persisted to the same data storage node, thereby reducing cross-node transaction overhead in the distributed architecture and maintaining the continuity of underlying data flow.

[0064] When the temporary drift areas corresponding to two or more data synchronization groups overlap, the gateway routing control module determines the final effective temporary addressing entry according to preset conflict resolution rules. As a preferred approach, the conflict resolution rules may include at least one of the following: prioritizing the temporary entry generated earlier, or prioritizing the one with the smaller dictionary order of the master node device identifier. Ineffective temporary entries are placed in a queue for recycling and are cleaned up uniformly after the corresponding synchronization group's write process is completed. By setting conflict resolution rules, routing ambiguity can be avoided when multiple synchronization groups perform temporary drift on the same area.

[0065] In this embodiment, following the dynamic modification results at the logical routing level, the system proceeds to step S6: aggregation transaction persistence and composite index construction. The master node encapsulates the aligned data into a multi-entity aggregation transaction and uploads it. The routing gateway server forwards the multi-entity aggregation transaction to the corresponding data storage node based on the modified logical routing mapping table. Simultaneously, the index construction module generates a composite primary key based on the topological hash string and relative time anchor points to complete the storage loop. This process aims to achieve cohesive storage and retrieval support for multi-agent related data, specifically implemented through the following sub-steps: S601, the master node performs serialization and encapsulation of time-aligned data and generates a multi-entity aggregated transaction. After completing the time-series alignment of multiple nodes within the local space, the master node extracts the alignment position coordinates and alignment speed of relevant members within the data synchronization group. As a preferred method, to adapt to the bandwidth transmission limitations of the local area network and the external network and to maintain the structural consistency of multi-node data, the master node uses a serialization framework to package the aligned state data along with the relative time anchor and topology hash string.

[0066] Those skilled in the art can use the Protobuf binary protocol or JSON data structure to define a serialization format, encapsulating discrete node state sets into a structured data block. During the encapsulation process, if it is detected that the alignment state data of individual nodes in the data synchronization group is missing due to underlying memory overflow or process abnormality, the master node will skip the data of that node and serialize the remaining valid nodes; if the number of valid nodes is less than 50% of the initial total number of nodes in the group, the master node will actively stop the current encapsulation operation to reduce the probability of generating incomplete transactions lacking collaborative reference. The structured data block after the above serialization process is defined as a multi-entity aggregation transaction.

[0067] In step S602, the routing gateway server performs aggregate transaction forwarding and disk write based on the modified logical routing mapping table. At the entry point for data flow to the external network, the routing gateway server receives a multi-entity aggregate transaction uploaded by the master node. For this flow request, the gateway routing control module queries the logical routing mapping table residing in the memory of the logical mapping module. If a boundary adaptive drift mechanism was triggered in the preceding steps, the target route pointer for the relevant coordinate area has been temporarily overwritten with the data storage node currently connected to by the master node. Based on the query result, the routing gateway server forwards the multi-entity aggregate transaction to the target data storage node. Upon receiving the transaction, the data storage node calls the write interface of the underlying database system to perform data write to disk.

[0068] In this embodiment, to prevent data corruption caused by partial writes, the data storage node employs transaction-level atomic write operations to maintain data integrity. In the event of a transaction rollback, the data storage node returns a write failure flag to the routing gateway server. The gateway routing control module has a built-in retry counter; upon receiving a failure flag, it re-initiates the write request a preset number of times. If all retries fail to write to disk, the multi-entity aggregated transaction is transferred to a local error log queue and an operational alert is triggered, thereby preventing persistent deadlocks caused by network anomalies.

[0069] S603, the index building module concatenates a composite primary key based on a topological hash string and a relative time anchor. In historical retrieval operations recording multi-entity interference events, the system typically needs to perform bidirectional positioning based on time intervals and spatial topological features. Based on this, the index building module extracts the topological hash string and relative time anchor from the multi-entity aggregation transactions already written to disk. After obtaining the above feature fields, the index building module uses string concatenation logic, using the relative time anchor representing the time dimension as a prefix and the topological hash string representing the spatial structure dimension as a suffix, connecting them with a preset text separator to generate a unique identifier, which is the composite primary key.

[0070] In the underlying principles of database indexing, storage structures such as B+ trees typically follow the leftmost prefix matching principle. Using the time dimension as a prefix allows for range pruning during retrieval by leveraging the monotonically increasing nature of time series data. Subsequently, a hash suffix is ​​used for spatial structure feature matching, thereby reducing the number of I / O reads from the underlying disk. The above textual description suffixes the implementation logic of string serialization and concatenation, and no further mathematical formulas will be introduced here. This composite primary key serves as the underlying data index key value of the database, providing a data anchoring basis for subsequent queries.

[0071] S604, the gateway routing control module performs route mapping state recovery and supports closed-loop storage retrieval. After the multi-entity aggregation transaction is completed and the composite primary key is constructed, the system determines that the current data write process has ended. To prevent temporary route overwriting from remaining in memory and interfering with the addressing requests of other non-cross-boundary entities, the gateway routing control module triggers its internal state recovery logic.

[0072] Specifically, the gateway routing control module locates the temporarily modified addressing entries in the logical routing mapping table and restores their target route pointers to the initially set default node pointers. To address situations where route overwriting cannot be reset due to server downtime or prolonged disk write failures, the gateway routing control module sets a preset timeout reset threshold for single modification operations, with a value ranging from 5 to 15 seconds. When the time elapsed since the temporary modification operation exceeds this preset timeout reset threshold, the gateway routing control module triggers the aforementioned rollback restoration of the target route pointer. In subsequent business applications, when an external system initiates a traceability command for a specific spatiotemporal state, the index building module receives the target time and topology parameters, reverses the reconstruction of the target composite primary key according to the same rules, and then extracts the corresponding multi-entity aggregated transactions in the distributed database cluster to form an index structure available for historical retrieval.

[0073] In this embodiment, based on the method flow and logical structure disclosed in the foregoing embodiments, the present invention also provides a corresponding virtualization system, physical electronic device, and associated computer-readable storage medium.

[0074] To enable data flow and dynamic addressing in multi-entity collaborative operation scenarios, this invention provides a robot distributed collaborative operation data acquisition and management system.

[0075] The system's functional logic is divided into multiple collaborative modules, each module having a corresponding execution mapping relationship with the aforementioned method steps.

[0076] Specifically, the system includes: The system comprises the following modules: a data acquisition module, which acquires the location coordinates and running speed of the mobile robot node in the physical work area and outputs its basic operating status; a local communication module, which broadcasts the basic operating status in the local area network and aggregates the broadcast information of neighboring nodes to form a local state set; an edge-aware computing module, which calculates the collaborative tensor based on the local state set and generates multi-entity aggregated transactions and associated spatial topological features through time-window aggregation; a gateway routing control module, deployed in the routing gateway server, which receives spatial topological features and modifies the logical routing mapping table in memory to output the updated routing mapping table; a logical mapping module, which resides in and maintains the updated routing mapping table for gateway addressing queries during data flow; a data storage node, deployed in the database cluster, which receives and writes multi-entity aggregated transactions based on the addressing and routing instructions of the updated routing mapping table; and an index building module, which extracts the internal feature fields of the written multi-entity aggregated transactions and concatenates them to generate a composite primary key to form an index structure available for historical retrieval.

[0077] See attached document Figure 4Based on the actual physical deployment scenario, the functions of each module of the above system can be carried in the corresponding electronic device. As a preferred approach, this electronic device can be configured as the on-board computing unit of a mobile robot node in the physical scenario, or as the hardware carrier of a routing gateway server on the data center side. This electronic device includes a processor, memory, and network interface components coupled to each other via a communication bus.

[0078] Regarding the underlying hardware form factor of the processor, it can be implemented using chips with logic operation capabilities such as a central processing unit (CPU), a digital signal processor (DSP), or a field-programmable gate array (FPGA). The processor is responsible for reading computer program instructions from memory and executing them to complete the computational tasks described in the preceding embodiments, such as data alignment, topology momentum calculation, routing table overwriting, and composite primary key concatenation. To avoid single-point operational failures causing device stagnation or routing control blockage, the processor employs a multi-core heterogeneous architecture for primary and backup thread isolation. When the underlying watchdog timer detects that the main working thread is blocked or times out, the hardware interrupt mechanism migrates the task context to the backup thread for continued execution, thus maintaining the continuity of the computational process.

[0079] In terms of memory structure, it includes volatile random access memory and non-volatile storage media. The volatile memory is mainly used to provide high-speed memory residency space for the logic mapping module to support routing lookups and temporary pointer modifications; the non-volatile storage media is used to store the underlying operating system kernel and the business application of this invention.

[0080] Considering the volatile nature of memory-level route overwrite operations, an asynchronous persistent queue is also configured in the memory to periodically flush the pointer modification logs that occur in memory to non-volatile storage media, so that the routing state can be restored by log replay after the device encounters an abnormal power outage, preventing irreversible loss of address mapping.

[0081] The network interface component provides a physical channel for data interconnection between electronic devices. In practical applications, when an electronic device acts as a mobile robot node, its network interface uses a Wi-Fi or 5G wireless communication module that supports low-latency broadcasting. To prevent the loss of synchronization data due to the disconnection of a single wireless network, the network interface component is equipped with a multi-NIC dual-link hot standby mechanism. When the primary link signal is lower than the preset reception strength, the hardware level triggers link switching logic to switch to the backup channel for data transmission.

[0082] When an electronic device acts as a routing gateway server, its network interface uses a 10 Gigabit fiber optic Ethernet card to handle high-concurrency data forwarding and disk write tasks. The processor's underlying instruction pipeline scheduling and the network interface's underlying send / receive handshake protocol can be implemented by those skilled in the art based on common computer architecture standards. Their microarchitectural operating mechanisms are well-known technologies in this field and will not be elaborated upon here.

[0083] Based on the aforementioned hardware execution environment, this invention also provides a computer-readable storage medium. In this embodiment, the computer-readable storage medium stores a computer program or an executable instruction set. When the computer program is read and executed by the processor of an electronic device, the electronic device is able to independently or collaboratively execute the method steps described in the foregoing embodiments of this invention, such as distributed routing drift and aggregated storage. As a specific implementation of the lower-level features, the computer-readable storage medium can take the form of a read-only memory, a portable flash drive (USB flash drive), a mobile hard drive, a magnetic storage disk, or an optical read-only optical disc, etc.

[0084] To aid in understanding the operational logic and data flow process of this invention in a real physical environment, this embodiment provides specific implementation data based on a large-scale warehousing and logistics sorting operation scenario with an area of ​​2,000 square meters. Forty mobile robot nodes are deployed within this physical operation site, the underlying database cluster contains four data storage nodes, and the routing gateway server has divided the physical operation site into fifty two-dimensional spatial grid identifiers during the initialization phase and established an initial logical routing mapping table.

[0085] During a specific work period, the system detected two devices performing rack handling tasks. The location coordinates of the first device were set as follows: The running speed is The location coordinates of the second device are: The running speed is The system sets distance attenuation weights. Zero constant Speed ​​Co-weight The system extracts the aforementioned state data and substitutes the specific values ​​into the formula for calculating the cooperative tensor. The specific substitution and calculation process is as follows: ; ; ; ; The above calculations yield a collaborative tensor of approximately 3.971 for the first device. The system's preset interference threshold is 3.0, and this collaborative tensor value exceeds this threshold. Based on this, the system blocks the discrete data direct write mode of both devices, forcing them to establish a data synchronization group via the local area network. The system further compares the values ​​according to the group's master node election rules: when the collaborative tensor values ​​of the two devices are different, the device with the larger collaborative tensor value is selected as the master node; when the collaborative tensor values ​​of the two devices are the same, the master node is determined according to the dictionary order of the device identifier codes. In this embodiment, the first device meets the master node election rules, therefore, the system timestamp of the first device is used as the relative time anchor. The system then proceeds to the alignment and routing prediction stage, setting the total virtual mass weight of the first device and its loaded cargo. The virtual quality weight of the second device The system substitutes the aforementioned mass parameters and alignment velocity into the topological momentum calculation formula. The specific substitution and calculation process is as follows: ; ; ; Through linear superposition, the topological momentum vector of this data synchronization group is found to have a horizontal component of 75° and a vertical component of 15°. Based on this topological momentum, the predicted centroid coordinates of the group are calculated, and the system sets the prediction time step. The above state data is then substituted into the group centroid prediction formula. The specific substitution and calculation process is as follows: ; ; ; ; The predicted centroid coordinates of this group in the next period are obtained from the above calculations. The system further calculates the distance of each node within the data synchronization group relative to the group centroid, and takes the straight-line distance from the node farthest from the centroid as the initial envelope radius. In this embodiment, the calculated initial envelope radius is less than the system's preset minimum safety compensation radius of 2.0 meters, so 2.0 meters is taken as the radius of the predicted coordinate domain, and the predicted coordinate domain is constructed with the predicted centroid coordinates as the center. The gateway routing control module detects that the predicted coordinate domain has crossed the boundary of the currently belonging twelfth grid and extended to the adjacent thirteenth grid. Based on this determination, the gateway routing control module establishes a temporary drift addressing entry in memory for the out-of-bounds local area, so that the area is temporarily mapped to the first data storage node currently accessed by the master node at the logical addressing level. Through this memory-level dynamic modification, the data generated by these two entities when they cross the boundary is cohesively encapsulated into a multi-entity aggregate transaction and uniformly written to the first data storage node.

[0086] To verify the technical effectiveness of this invention, a concurrent data write test was conducted in the aforementioned warehousing environment. The test involved mobile robot nodes increasing from ten to one hundred, recording the system's underlying data write performance when handling high concurrency and close-range interactions between multiple entities. The test introduced a traditional discrete direct-write routing strategy as a benchmark, and extracted two core indicators for statistical analysis: cross-node transaction lock conflict rate and data write latency.

[0087] See attached document Figure 5 ,exist Figure 5 In the diagram, the black solid line with dots represents the traditional discrete direct-write routing strategy, while the dark gray solid line with squares represents the dynamic drift routing strategy based on topological momentum. As the number of concurrent mobile robot nodes increases, the lock contention rate corresponding to the traditional discrete direct-write routing strategy shows a significant upward trend. This is because, in high-density collaborative areas, multiple nodes, due to their dispersed physical locations spanning multiple fixed spatial grids, cause their state data to be routed to different underlying data storage nodes, thus triggering distributed cross-node write lock contention. After applying the solution of this invention, even with one hundred nodes, the cross-node transaction lock contention rate corresponding to the dark gray solid line with squares remains at a low level. This technical effect confirms the effectiveness of the gateway routing control module's coordinate domain prediction mechanism based on topological momentum vector calculation. By overwriting the target pointer in the logical routing mapping table in memory, the state data of densely collaborative entity groups is converged and cohesed into a single data storage node during cross-regional processes, solving the distributed lock contention problem between heterogeneous database nodes.

[0088] See attached document Figure 6 ,exist Figure 6In the diagram, black solid lines with dots represent traditional discrete direct-write routing strategies, while dark gray solid lines with squares represent dynamic drift routing strategies based on topology momentum. Test data shows that in high-concurrency collaborative scenarios, the data write latency of traditional strategies increases non-linearly. Numerous cross-node synchronization waits and fragmented writes cause input / output performance bottlenecks. This embodiment employs a dynamic drift routing strategy based on topology momentum to control data write latency within a stable range. Based on the time-window aggregation mechanism described in the previous embodiment, the system uniformly aligns the discrete state data of multiple neighboring entities within the data synchronization group and packages them into multi-entity aggregated transactions. This multi-entity aggregated encapsulation, combined with dynamic drift adjustment of routing addressing, reduces the number of random handshake interactions between the underlying network interface components and the disk array, improving the data throughput of the distributed cluster system. Simultaneously, the memory-level route overwrite mechanism replaces the read / write modification of disk configuration files, shortening the latency between command issuance and data flow.

Claims

1. A robot distributed collaborative operation data acquisition and management system, characterized in that, include: The data acquisition module is used to acquire the position coordinates and running speed of the mobile robot node in the physical work site and output the basic operating status. The local communication module is used to broadcast the basic operating status in the local area network and aggregate the broadcast information of neighboring nodes to form a local state set. An edge-aware computing module is used to calculate a collaborative tensor based on the local state set and generate multi-entity aggregated transactions and associated spatial topological features through time-window aggregation. The gateway routing control module, which is deployed in the routing gateway server, is used to receive the spatial topology features and modify the logical routing mapping table in memory to output the updated routing mapping table. The logical mapping module is used to reside in and maintain the updated routing mapping table for gateway addressing queries during data flow. The data storage node, deployed in the database cluster, is used to receive and write the multi-entity aggregation transaction according to the addressing and routing instructions of the updated routing mapping table; The index building module is used to extract the internal feature fields of the written multi-entity aggregation transaction and concatenate them to generate a composite primary key to form an index structure that can be retrieved historically.

2. The robot distributed collaborative operation data acquisition and management system according to claim 1, characterized in that, The gateway routing control module extracts the feature boundaries of the physical work site to generate a discrete set of reference points. The gateway routing control module performs polygon subdivision on the continuous physical space based on the discrete reference point set and divides the space into two-dimensional spatial grid identifiers; The gateway routing control module uses a consistent hashing algorithm to construct the initial mapping and allocation relationship between the two-dimensional spatial grid identifier and the data storage node, so that the logical mapping module in the system instantiates and generates a logical routing mapping table in memory.

3. The robot distributed collaborative operation data acquisition and management system according to claim 1, characterized in that, The edge-aware computing module calculates the collaborative tensor using the collaborative tensor calculation formula; The cooperative tensor calculation formula weights and sums the spatial exclusion component generated by relative distance and the motion interference component generated by relative velocity to quantitatively assess the overall interference risk level in the local area where the current node is located.

4. The robot distributed collaborative operation data acquisition and management system according to claim 3, characterized in that, When the calculated cooperative tensor reaches the preset interference threshold, the edge-aware computing module generates an internal blocking control command to stop the current discrete data direct writing mode and triggers the local communication module to perform a dynamic handshake to build a data synchronization group. Within the data synchronization group, node comparisons are performed to elect a master node; The master node extracts the local system clock to generate a relative time anchor point to avoid global clock synchronization errors, and sends the relative time anchor point to all slave nodes in the data synchronization group. The preset interference threshold is comprehensively calibrated by combining the spatial safety tolerance of the physical work site with the node communication delay bandwidth.

5. A robot distributed collaborative operation data acquisition and management system according to claim 4, characterized in that, The edge-aware computing module performs a state interpolation alignment algorithm based on the relative time anchor point to uniformly interpolate and align the acquisition status of the other nodes in the data synchronization group to the relative time anchor point, thereby obtaining the alignment position coordinates and alignment running speed. The edge-aware computing module calculates and generates a topological momentum vector based on the aligned running speed using the topological momentum calculation formula. The topological momentum calculation formula linearly superimposes the motion vectors of all nodes in the data synchronization group according to their corresponding virtual mass weights to reflect the comprehensive direction and speed of the entity group within the local risk area.

6. The robot distributed collaborative operation data acquisition and management system according to claim 5, characterized in that, The edge-aware computing module calculates the collaborative tensor based on the local state set and generates the multi-entity aggregated transaction and associated spatial topological features through time-window aggregation, including: The edge-aware computing module uses the alignment position coordinates of each node as the graph vertices and the relative distance between nodes as the graph edges to construct a relatively spatially connected graph that reflects the spatial distribution relationship and generate the corresponding adjacency matrix. The edge-aware computing module flattens the adjacency matrix into a one-dimensional numerical sequence and performs a hash operation on the one-dimensional numerical sequence to generate a topological hash string. The edge-aware computing module combines and encapsulates the previously generated topological momentum vector with the topological hash string generated in this step, and finally fuses them to generate the spatial topological feature.

7. A robot distributed collaborative operation data acquisition and management system according to claim 6, characterized in that, The routing gateway server receives the spatial topology features and extracts the topological momentum vector from the spatial topology features; The gateway routing control module calculates the predicted centroid coordinates of the group based on the topological momentum vector using the group centroid prediction formula. The group centroid prediction formula is based on the initial centroid coordinates of the group calculated by mass-weighted average, and the overall position offset generated within the prediction time step driven by the topological momentum vector. The gateway routing control module constructs a circular envelope region centered on the group prediction centroid coordinates as the prediction coordinate domain for the next cycle.

8. A robot distributed collaborative operation data acquisition and management system according to claim 7, characterized in that, The gateway routing control module receives the spatial topology features and modifies the logical routing mapping table in memory accordingly to output the updated routing mapping table, including: The gateway routing control module executes spatial conflict determination logic across the current grid boundary and performs spatial intersection test between the geometric contour of the next cycle predicted coordinate domain and the preset grid boundary in the system. When it is determined that the next cycle prediction coordinate domain crosses the current grid boundary, the gateway routing control module extracts the set of local spatial coordinates where the next cycle prediction coordinate domain overlaps with the adjacent grid. The gateway routing control module dynamically generates high-priority temporary addressing entries for the local spatial coordinate set in the logical routing mapping table. Based on this, it performs memory-level modifications to the logical routing mapping table, temporarily mapping and associating the expected adjacent grid local areas to the data storage node currently accessed by the master node at the logical addressing level, and finally outputs the updated routing mapping table.

9. A robot distributed collaborative operation data acquisition and management system according to claim 8, characterized in that, The master node in the system serializes and encapsulates the aligned data to generate the multi-entity aggregated transaction and transmits it upwards; The data storage node in the system receives and completes the writing of the multi-entity aggregation transaction according to the addressing and routing instructions of the updated routing mapping table; The index building module extracts the internal feature fields of the written multi-entity aggregation transaction and concatenates them to generate the composite primary key to form an index structure available for historical retrieval, including: The index building module extracts the multi-entity aggregated transaction that has been written to disk, and obtains the topological hash string and the relative time anchor point contained in the internal feature fields of the multi-entity aggregated transaction. The relative time anchor point, which represents the time dimension, is used as a prefix, and the topological hash string, which represents the spatial structure dimension, is used as a suffix. A preset text separator is introduced in the middle for connection, and finally the composite primary key is generated.

10. A method for data acquisition and management of distributed collaborative robot operations, characterized in that, The robot distributed collaborative operation data acquisition and management system according to any one of claims 1-9 includes the following steps: The gateway routing control module extracts the feature boundaries of the physical work site to divide a two-dimensional spatial grid identifier and stores the initial logical routing mapping table in memory; The data acquisition module obtains the position coordinates and running speed of the mobile robot node in the physical work site and outputs the basic operating status. The local communication module broadcasts the basic operating status in the local area network and aggregates the broadcast information of neighboring nodes to form a local state set; The edge-aware computing module calculates a collaborative tensor based on the local state set. When the collaborative tensor reaches a preset interference threshold, it triggers data synchronization interpolation and alignment. It then uses the aligned data to generate a topological momentum vector and a topological hash string, which are then aggregated through a time window to generate multi-entity aggregated transactions and associated spatial topological features. The gateway routing control module receives the spatial topology features, uses the topological momentum vector in the spatial topology features to calculate the next cycle prediction coordinate domain, and modifies the logical routing mapping table in memory to output the updated routing mapping table. The logical mapping module resides in and maintains the updated routing mapping table for gateway addressing queries during data flow; The data storage node receives and writes the multi-entity aggregation transaction according to the addressing and routing instructions of the updated routing mapping table; The index building module extracts the relative time anchor and topological hash string from the internal feature fields of the written multi-entity aggregation transaction, concatenates them end to end, and finally generates a composite primary key to form an index structure that can be retrieved historically. The preset interference threshold is comprehensively calibrated by combining the spatial safety tolerance of the physical work site with the node communication delay bandwidth.

Images