A method and system for information and data processing in high-concurrency scenarios

Abstract

This invention relates to the field of network communication and collaborative control technology, and particularly to an information data processing method and system applied to high-concurrency scenarios. An information data processing method for high-concurrency scenarios includes the following steps: S1: Obtaining the initial and final information reception times of each cluster member in a robot cluster under high concurrency; and extracting the command and instruction sending times for the robot cluster; S2: Determining the leader device, candidate devices, and follower devices based on the initial information reception time, final information reception time, and command and instruction sending times. This invention constructs a dynamic scoring and clustering model based on time-series differences and multi-dimensional state data to achieve adaptive robot role partitioning; by monitoring role conversion rates and optimizing parameters, it overcomes the defects of static hierarchical classification, and improves the stability, fault tolerance, and response efficiency of cluster collaboration under high concurrency.

Description

Technical Field

[0001] This invention relates to the field of network communication and collaborative control technology, and in particular to an information data processing method and system applied in high-concurrency scenarios. Background Technology

[0002] In the field of distributed collaborative computing, high-concurrency scenarios refer to intensive operating environments where a large number of nodes simultaneously initiate data requests or execute commands. With the development of IoT and swarm intelligence technologies, existing methods have made significant progress in data distribution efficiency and node collaboration by improving communication protocols and load balancing mechanisms, providing technical support for real-time response in multi-agent systems.

[0003] However, traditional processing methods based on fixed thresholds and static hierarchies are insufficient in the face of high-concurrency environments with dynamic topologies and bursty traffic. Existing technologies lack the ability to jointly perceive network latency jitter and node state transitions, making it difficult to cope with system oscillations caused by out-of-order packet arrivals and resource contention. This leads to unstable node role allocation and impaired command propagation consistency, affecting the overall control accuracy of the cluster system.

[0004] Therefore, there is an urgent need to establish a collaborative mechanism that integrates communication quality assessment and dynamic role adjustment in order to improve the adaptive capability and operational stability of distributed systems in high-concurrency environments. Summary of the Invention

[0005] To overcome the drawback of unstable character states, this invention provides an information data processing method and system applicable to high-concurrency scenarios.

[0006] The technical implementation of this invention is: an information data processing method applied in high-concurrency scenarios, comprising the following steps:

[0007] S1: Obtain the initial and final information reception times of each cluster member in the robot cluster under high concurrency scenarios; and extract the command and instruction sending times to the robot cluster.

[0008] S2: Determine the leader device, candidate device, and follower device based on the initial information reception time, the final information reception time, and the command / instruction sending time;

[0009] S3: Determine the demotion rate of the leader device, the promotion rate of the candidate device, and the promotion rate of the follower device in each information data processing cycle under high concurrency scenarios based on the leader device, candidate device, and follower device;

[0010] S4: Adjust the information data processing flow for the next information data processing cycle based on the aforementioned leadership equipment downgrade rate, candidate equipment promotion rate, and follower equipment promotion rate.

[0011] Preferably, the step of obtaining the initial and final information reception times of each cluster member in a high-concurrency scenario, and extracting the command and instruction sending times to the robot cluster, includes:

[0012] The time difference between the command instruction sending time and the initial information reception time of each cluster member in the robot cluster is used as the first time difference;

[0013] The first time difference is used to characterize the delay time from the issuance of a command instruction to the initial receipt of the command by a cluster member, reflecting the initial response speed of the cluster member to the command;

[0014] The time difference between the command and instruction sending time and the final information reception time of each cluster member in the robot cluster is used as the second time difference;

[0015] The second time difference is used to characterize the delay time from when a cluster member issues a command instruction to when it finally receives the command, reflecting the reliability of the cluster member's complete reception of the command.

[0016] Preferably, the step of determining the leader device, candidate device, and follower device based on the initial information reception time, the final information reception time, and the command / instruction transmission time includes:

[0017] The leader device is determined based on the first time difference, combined with the location information, remaining energy, and positioning accuracy of each member in the robot cluster.

[0018] Define the members of the robot cluster, excluding the leader device, as candidate members;

[0019] The candidate device is determined based on the second time difference and the relative position information of the candidate member with respect to the leadership device.

[0020] The leader device and other cluster members in the robot cluster, besides the candidate devices, are designated as follower devices.

[0021] Preferably, determining the leader device based on the first time difference combined with the position information, remaining energy, and positioning accuracy of each member in the robot swarm includes:

[0022] The strategic value of the location information within the cluster is scored and normalized to obtain a normalized location score. The remaining energy and positioning accuracy are also normalized to obtain normalized remaining energy and normalized positioning accuracy. An initial score is obtained using an initial scoring formula. Based on the initial score, candidate leadership devices are selected. The initial scoring formula is as follows:

[0023] ;

[0024] in, For the initial score, For cluster members The first time difference It is the minimum of the first time differences among all cluster members. The maximum value among the first time differences of all cluster members. For cluster members Normalized position score, For cluster members Normalized residual energy, For cluster members Normalized positioning accuracy, , , and Let be the weighting coefficient, satisfying + + + =1;

[0025] Based on the initial score, cluster members that exceed the preset initial score threshold will be selected as candidate leader devices.

[0026] Preferably, the step of selecting cluster members with scores greater than a preset initial score threshold as candidate leader devices based on the initial score includes:

[0027] The logical position ranking of the candidate leader devices in the cluster is determined based on the structural value of the devices in the network topology and the cost of leadership switching. Second time difference data of all candidate leader devices are collected, and a folding coefficient is calculated for each candidate leader device. The formula for the folding coefficient is:

[0028] ;

[0029] in, The folding factor is... For current cluster members The order of leadership, and , "Time" indicates the highest priority. When indicates the lowest priority, This represents the total number of devices in the cluster. For summing loop variables, For cluster members The second time difference, The minimum of the second time differences among all candidate leadership devices. The maximum value among the second time differences of all candidate leadership devices. The sensitivity coefficient, , Let be the weighting coefficient, satisfying + =1;

[0030] The folding coefficient is used to quantify the combined risk of system convergence delay and switching failure caused by communication competition and delay due to data flooding when switching cluster leadership to a backup leader device in high-concurrency scenarios.

[0031] The final score for each candidate leadership device is obtained based on the folding coefficient using the final scoring formula, which is:

[0032] ;

[0033] in, For the final score, For adjustment factors;

[0034] Based on the final score, candidate leadership devices that exceed the preset final score threshold are identified as leadership devices.

[0035] Preferably, determining the candidate device based on the second time difference and the relative position information of the cluster members (excluding the leader device) relative to the leader device includes:

[0036] The relative position distance set is obtained by calculating the relative position distance between each cluster member (excluding the leader device) and the leader device in the robot cluster.

[0037] K-means clustering was performed on N high-concurrency scenarios based on their cycle duration and suddenness to obtain the first clustering result;

[0038] The second time difference data of each cluster member in the robot cluster, excluding the leader device, is obtained and defined as cluster data. The cluster data is then subjected to DBSCAN clustering to obtain the second clustering result.

[0039] Based on the first clustering result, the cluster members in the relative position distance set are connected according to the shortest distance between devices to obtain the first topological connection relationship;

[0040] Based on the second clustering result, cluster members in the relative location distance set are connected according to the shortest distance between devices to obtain a second topological connection relationship;

[0041] Candidate devices are determined based on the first topology connection relationship and the second topology connection relationship.

[0042] Preferably, determining the candidate device based on the first topology connection relationship and the second topology connection relationship includes:

[0043] The absolute difference is obtained by calculating the absolute difference of the second time difference between neighboring cluster members in the first and second topological connection relationships.

[0044] For each pair of neighboring devices, if the absolute difference result is less than the preset absolute difference threshold, the cluster member with the largest second time difference in the pair of neighboring devices is removed to obtain the removal result.

[0045] If the absolute difference result is greater than or equal to the preset absolute difference threshold, then the cluster members in the neighboring device pair are retained to obtain the retention result;

[0046] The first topological connection relationship is updated based on the elimination result and the retention result to obtain the third topological connection relationship;

[0047] The second topological connection relationship is updated based on the elimination result and the retention result to obtain the fourth topological connection relationship;

[0048] Extract shared cluster members from the third and fourth topology connection relationships;

[0049] The shared cluster members are connected according to the shortest distance between devices to construct the candidate device topology connection relationship;

[0050] Cluster members in the candidate device topology connection relationship are identified as candidate devices.

[0051] Preferably, determining the demotion rate of the leader device, the promotion rate of the candidate device, and the promotion rate of the follower device based on each information data processing cycle in a high-concurrency scenario includes:

[0052] Monitor the role status changes of all cluster members in the robot cluster, and collect the number of times the leader device becomes a candidate device, the number of times a candidate device becomes a leader device, the number of times a follower device becomes a candidate device, and the total number of role transitions;

[0053] The leadership device downgrade rate is obtained by dividing the number of times a leadership device is converted to a candidate device by the total number of role conversions.

[0054] The candidate device promotion rate is obtained by dividing the number of times a candidate device is converted into a leadership device by the total number of role conversions.

[0055] The promotion rate of the follower device is obtained by dividing the number of times the follower device is converted into a candidate device by the total number of role conversions.

[0056] Preferably, the information data processing flow for adjusting the next information data processing cycle based on the leadership equipment downgrade rate, candidate equipment promotion rate, and follower equipment promotion rate includes:

[0057] When the degradation rate of leadership equipment exceeds the preset degradation threshold, the preset final score threshold will be increased in the next information data processing cycle.

[0058] When the candidate device promotion rate is lower than the preset promotion threshold, the preset final score threshold will be lowered in the next information data processing cycle.

[0059] When the follower device's advancement rate exceeds the preset disturbance threshold, the neighborhood radius parameter and minimum sample number of the DBSCAN clustering algorithm are tightened in the next information data processing cycle, and the preset absolute difference threshold is reduced.

[0060] An information data processing system for high-concurrency scenarios includes:

[0061] The time difference acquisition and delay assessment module is used to obtain the command sending and receiving time, calculate the first time difference and the second time difference, and evaluate the initial response speed and complete reception reliability of the cluster members respectively.

[0062] The three-layer device dynamic decision-making module determines the leader device, candidate devices, and follower devices based on time difference, location, energy, and positioning accuracy data through a scoring mechanism and topology analysis.

[0063] The role conversion rate monitoring and analysis module monitors the number of device role conversions, calculates the demotion rate of leading devices, the promotion rate of candidate devices, and the promotion rate of following devices, and triggers parameter adjustment instructions based on thresholds.

[0064] The parameter adaptive optimization execution module dynamically adjusts the scoring threshold, clustering parameters, and absolute difference standard based on the conversion rate analysis results, thereby optimizing the data processing strategy for the next processing cycle.

[0065] Beneficial Effects: This invention collects command and instruction transmission times, initial information reception times, and final information reception times of robot swarms under high-concurrency scenarios, constructing a first time difference and a second time difference to quantify the initial response speed and complete reception reliability of swarm members. Based on time differences, location information, remaining energy, and positioning accuracy data, a scoring mechanism and topology analysis are used to dynamically determine leader devices, candidate devices, and follower devices, effectively overcoming the instability problem of role allocation in traditional static hierarchical methods under dynamic topologies and sudden traffic. Furthermore, by monitoring the demotion rate of leader devices, the promotion rate of candidate devices, and the promotion rate of follower devices, a quantitative evaluation result of role conversion rate is formed, and the scoring threshold, clustering parameters, and absolute difference standard are dynamically adjusted accordingly to achieve adaptive parameter optimization. Ultimately, while improving information data processing efficiency, this invention significantly enhances the collaborative stability, fault tolerance, and overall adaptive performance of robot swarms in high-concurrency environments. Attached Figure Description

[0066] Figure 1This is a flowchart of the information data processing method of the present invention applied to high-concurrency scenarios;

[0067] Figure 2 This is a structural diagram of the information data processing system of the present invention applied to high-concurrency scenarios. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] Example 1: An information data processing method applied in high-concurrency scenarios, such as... Figure 1 As shown, it includes the following steps:

[0070] S1: Obtain the initial and final information reception times of each cluster member in the robot cluster under high concurrency scenarios; and extract the command and instruction sending times to the robot cluster.

[0071] The time difference between the command instruction sending time and the initial information reception time of each cluster member in the robot cluster is used as the first time difference;

[0072] The first time difference is used to characterize the delay time from the issuance of a command instruction to the initial receipt of the command by a cluster member, reflecting the initial response speed of the cluster member to the command;

[0073] The time difference between the command and instruction sending time and the final information reception time of each cluster member in the robot cluster is used as the second time difference;

[0074] The second time difference is used to characterize the delay time from when a cluster member issues a command instruction to when it finally receives the command, reflecting the reliability of the cluster member's complete reception of the command.

[0075] It should be noted that in traditional distributed collaborative control systems, static threshold-based processing mechanisms are ill-suited to network latency fluctuations in high-concurrency environments, often leading to oscillations in node role allocation and a lack of consistency in command propagation. This solution collects the initial and final information reception times of each member in the robot swarm and combines this with command / instruction transmission times for collaborative analysis. The initial information reception time refers to the instant when a swarm member successfully parses the command data packet for the first time, such as the moment when the first drone perceives the path update command in an emergency obstacle avoidance scenario. The final information reception time corresponds to the termination moment when a member completes the reception and verification of all command data, such as the time when the last robot in the formation confirms the completion of the command execution. The command / instruction transmission time is defined as the initial timestamp of the control node initiating command transmission. By calculating the first time difference, i.e., the interval between the transmission time and the initial reception time, the initial response agility of members to commands can be quantified, providing a key basis for dynamic leader election. Simultaneously, the second time difference, by measuring the difference between the transmission time and the final reception time, accurately reflects the integrity and reliability of command propagation within the swarm, thereby effectively identifying tail delay risks. This dual time difference measurement mechanism overcomes the shortcomings of traditional methods in terms of insufficient perception of network jitter, and significantly improves the collaborative stability and adaptive capability of the cluster under data flood through real-time latency assessment.

[0076] S2: Determine the leader device, candidate device, and follower device based on the initial information reception time, the final information reception time, and the command / instruction sending time;

[0077] The leader device is determined based on the first time difference, combined with the location information, remaining energy, and positioning accuracy of each member in the robot cluster.

[0078] The strategic value of the location information within the cluster is scored and normalized to obtain a normalized location score. The remaining energy and positioning accuracy are also normalized to obtain normalized remaining energy and normalized positioning accuracy. An initial score is obtained using an initial scoring formula. Based on the initial score, candidate leadership devices are selected. The initial scoring formula is as follows:

[0079] ;

[0080] in, For the initial score, For cluster members The first time difference It is the minimum of the first time differences among all cluster members. The maximum value among the first time differences of all cluster members. For cluster members Normalized position score, The normalized residual energy of the cluster members. For cluster members Normalized positioning accuracy, , , and Let be the weighting coefficient, satisfying + + + =1;

[0081] Based on the initial score, cluster members that exceed the preset initial score threshold will be selected as candidate leader devices.

[0082] It should be noted that in distributed collaborative control systems, traditional node role allocation mechanisms based on static thresholds often lead to frequent fluctuations in the leader device election results due to a lack of adaptability to the dynamic characteristics of the network. To address this issue, this step establishes a leader device election mechanism by integrating multi-dimensional dynamic parameters. The first time difference reflects the agility of the device's command response, and its calculation depends on the difference between the command sending timestamp and the initial receiving timestamp. Location information is collected through satellite positioning or inertial navigation systems and is used to assess the communication hub value of the device in the cluster topology. Remaining energy data is sourced from power management unit monitoring and is directly related to the device's continuous operating time. Positioning accuracy is determined by the sensor error range and affects the reliability of spatial decision-making. The strategic value score quantifies the contribution of the device's location to the cluster communication efficiency; for example, devices located at the center of the topology typically have higher communication hub value or link quality advantages. Normalization converts heterogeneous parameters into dimensionless values, eliminating the impact of magnitude differences on the comprehensive evaluation. The initial scoring formula integrates the four major elements of response speed, location value, energy state, and positioning accuracy through a weighted synthesis method, where the weight coefficients are determined through expert experience, analytic hierarchy process, or regression models based on historical data. For example, w1=0.4, w2=0.2, w3=0.2, w4=0.2. The preset initial scoring threshold is dynamically adjusted based on historical cluster performance data to form an adaptive screening threshold, thereby ensuring the quality of the election while avoiding frequent role switching caused by fluctuations in a single parameter.

[0083] The logical position ranking of the candidate leader devices in the cluster is determined based on the structural value of the devices in the network topology and the cost of leadership switching. Second time difference data of all candidate leader devices are collected, and a folding coefficient is calculated for each candidate leader device. The formula for the folding coefficient is:

[0084] ;

[0085] in, The folding factor is... For current cluster members The order of leadership, and , "Time" indicates the highest priority. When indicates the lowest priority, This represents the total number of devices in the cluster. For summing loop variables, For cluster members The second time difference, The minimum of the second time differences among all candidate leadership devices. The maximum value among the second time differences of all candidate leadership devices. The sensitivity coefficient, , Let be the weighting coefficient, satisfying + =1;

[0086] It should be noted that the summation variable From 1 to The leadership sequence number is used to iterate through all positions from the highest priority to the current device priority.

[0087] The folding coefficient is used to quantify the combined risk of system convergence delay and switching failure caused by communication competition and delay due to data flooding when switching cluster leadership to a backup leader device in high-concurrency scenarios.

[0088] It should be noted that the folding coefficient is a key risk prediction indicator, and its value directly quantifies the systemic risks caused by leadership transitions. A higher folding coefficient means a greater overall risk associated with selecting the candidate device. This is mainly reflected in two aspects: First, the structural restructuring risk quantified by leadership order predicts the cost of topology reconfiguration. For example, if a mobile robot at the rear of the platoon is promoted to leader, the entire cluster undergoes significant formation adjustments to rebuild the optimal communication link, inducing collisions between devices, temporary formation chaos, and additional energy consumption. Second, the communication instability risk quantified by the second time difference predicts the reliability of command transmission. For example, in autonomous driving platooning, if a candidate vehicle with a high second time difference becomes the lead vehicle, it cannot reliably receive and forward commands under high-concurrency command flows, leading to asynchronous responses from following vehicles, loss of platoon spacing, or even mission interruption. This formula, by integrating these two types of risks, achieves a forward-looking assessment of the stability of candidate devices under extreme pressure.

[0089] The final score for each candidate leadership device is obtained based on the folding coefficient using the final scoring formula, which is:

[0090] ;

[0091] in, For the final score, For adjustment factors;

[0092] Based on the final score, candidate leadership devices that exceed the preset final score threshold are identified as leadership devices.

[0093] It should be noted that traditional distributed systems, due to their use of static evaluation criteria, struggle to adapt to the dynamic characteristics of networks under high concurrency during leader election, often leading to frequent role switching and system oscillations. To address this deficiency, this solution introduces a logical position ranking mechanism. The logical position ranking is determined based on the structural value of a device in the network topology and the cost of leader switching. The structural value is quantified by the device's network centrality index, and the leader switching cost is assessed by the communication path reconstruction overhead required to switch to that device. Specifically, in a mesh communication cluster, the higher a device's direct connection count (i.e., degree centrality), the greater its structural value; simultaneously, the fewer communication links that need to be updated or rebuilt to switch to that device, the lower its switching cost. The logical position ranking is obtained by comprehensively evaluating the degree centrality and path reconstruction overhead of all candidate devices.

[0094] Leadership order ( The number represents the logical position ranking. For example, in a mesh cluster consisting of five devices (A, B, C, D, E), device C has direct connections to all other devices (highest degree centrality), and switching to C requires no reconstruction of any existing links (lowest switching cost), therefore its ranking is [number missing]. Device A is only connected to two devices (low degree centrality), and switching to A would require rebuilding multiple core links (high switching costs), therefore its priority is... Leadership hierarchy is used to quantify the structural adjustment costs during leadership transitions: the higher the hierarchy (…). The smaller the value, the lower the topology reconstruction overhead required to switch to that device, and the faster the system converges; conversely, the higher the value, the lower the priority ( The higher the value, the higher the risk of switching.

[0095] The core advantage of the folding coefficient design lies in the use of logarithmic functions. Characteristics: When in order Hours (ranked high), large rate of change of function value, can significantly amplify top-tier alternative equipment (such as...) and The minute difference in priority between the two ensures that the system strongly favors selecting the optimal device; when the priority... When the ranking is high (lower-ranked), the function value changes gradually, reducing the differences between low-ranked devices and effectively avoiding the impact of random fluctuations in tail-ranked devices on the evaluation results. This nonlinear mapping accurately reflects the engineering logic of "prioritizing the best while weakening the worst," mathematically strengthening the system's preference for high-ranked devices.

[0096] The second part of the formula uses an exponential function to perform a nonlinear transformation on the normalized second time difference. The core advantage of this design lies in the exponential function. Sensitivity amplification effect: by adjusting the sensitivity coefficient The curvature of the control function. When When the value is greater than 0, the function will normalize the delay. The linear growth of risk is transformed into a non-linear spike in risk value. This means that for a delay close to the optimal value ( Devices with latency close to 0 exhibit slow risk growth and good system latency tolerance; however, once the latency exceeds a certain critical point ( (If the latency increases), the risk value will rise sharply, creating a "cliff effect," thus quickly eliminating alternative devices whose communication reliability is already at a critical level. This design enables the system to keenly identify and avoid high-risk handover failures caused by high latency.

[0097] The final scoring formula achieves a balanced assessment of the overall capability and risk of candidate equipment by subtracting the product of the folding coefficient and the adjustment factor from the initial capability score. The adjustment factor is used to coordinate the weighting relationship between individual equipment performance and overall system stability.

[0098] The preset final score threshold is set as the average (or median) of the final scores of the candidate leader devices over the past M periods. When N candidate devices simultaneously exceed the threshold, the node with the highest final score is automatically selected as the sole leader, thus avoiding multi-leader decision conflicts. In a specific application scenario, in a vehicle-to-everything (V2X) collaborative control system, a vehicle with a stable communication link but a slightly slower initial response is more suitable for the leader role than a vehicle with a fast response but volatile network. This selection mechanism significantly enhances the overall resilience of the cluster system under high concurrency pressure.

[0099] Define the members of the robot cluster, excluding the leader device, as candidate members;

[0100] The candidate device is determined based on the second time difference and the relative position information of the candidate member with respect to the leadership device.

[0101] The relative position distance set is obtained by calculating the relative position distance between each cluster member (excluding the leader device) and the leader device in the robot cluster.

[0102] K-means clustering was performed on N high-concurrency scenarios based on their cycle duration and suddenness to obtain the first clustering result;

[0103] The second time difference data of each cluster member in the robot cluster, excluding the leader device, is obtained and defined as cluster data. The cluster data is then subjected to DBSCAN clustering to obtain the second clustering result.

[0104] Based on the first clustering result, the cluster members in the relative position distance set are connected according to the shortest distance between devices to obtain the first topological connection relationship;

[0105] Based on the second clustering result, cluster members in the relative location distance set are connected according to the shortest distance between devices to obtain a second topological connection relationship;

[0106] Candidate devices are determined based on the first topology connection relationship and the second topology connection relationship.

[0107] It should be noted that, to overcome the shortcomings of traditional candidate device selection mechanisms in responding insufficiently to network dynamic characteristics under high-concurrency environments, this solution proposes a candidate device determination method based on dual-cluster fusion. Relative position information is used to calculate a set of Euclidean distances through coordinate transformation, forming the basic dataset of spatial relationships between devices. Secondary time difference data focuses on the critical impact of tail delay on system response, combining it with relative position information to screen candidate devices that possess both communication reliability and physical proximity. Cycle duration: refers to the length of time a high-concurrency scenario lasts from start to finish. Its specific quantification method is: within a time window T, the total duration for which the received data request rate continuously exceeds a preset high load threshold (such as 150% of the system's average load). For example, a traffic peak lasting 5 seconds has a cycle duration of 5 seconds. Cycle suddenness: refers to the degree of drastic increase in load rate at the beginning of a high-concurrency scenario. Its specific quantification method is: calculating the instantaneous rate of change of the data request rate per unit time, usually expressed as the first derivative of the request rate within the time window or the absolute value of the slope between adjacent sampling points. For example, a request rate spike from 1000 requests / second to 5000 requests / second within 100 milliseconds is far more sudden than the same increase occurring within one second. In practical calculations, the standard deviation of the request rate within a sliding window or a specific percentile (such as the 95th percentile) is often used to approximate the drastic nature of this change.

[0108] In the scene feature analysis phase, the K-means clustering algorithm was used to process historical periodic data. The complete process included randomly selecting initial centroids, calculating the Euclidean distance between samples, assigning samples to the nearest cluster, recalculating cluster centroids, iteratively optimizing intra-cluster distances, and finally obtaining the first clustering result based on scene features. This result reveals the typical operating state of the system under different load modes.

[0109] For communication reliability assessment, DBSCAN clustering is performed on the second time difference data, executing the following steps: setting neighborhood radius parameters, calculating sample point density, identifying core objects and boundary points, expanding density-connected clusters, and classifying noise points, ultimately establishing device groups with similar delay characteristics. This processing method pays particular attention to the distribution pattern of tail delay, effectively identifying high-latency-risk areas.

[0110] The topology connections built based on the first clustering result prioritize physical proximity to form an optimal communication backbone network; the connections built based on the second clustering result ensure the balance of latency characteristics, forming a reliable subnetwork. Through cross-validation of the dual topology relationships, a set of candidate devices that can both meet the physical topology optimization requirements and effectively control tail latency is finally selected. This fusion mechanism significantly improves the overall response performance of the system in high-concurrency scenarios and establishes a high-quality backup node echelon for cluster collaborative control.

[0111] Example: Suppose K-means clustering is performed on 5 historical periods to obtain two typical scenario categories. When the system identifies the current period as belonging to the "short-term high-burst" category, a predefined tightly coupled topology strategy will be adopted: For devices in the relative distance set {A:0.2m, B:0.5m, D:0.3m, E:0.4m}, the nearest distance connection method is executed. This method selects the single nearest neighbor node for each device, thus forming a device pair AD (distance 0.141m) and BE (distance 0.141m) as the first topological connection relationship.

[0112] It is important to note that the first clustering result establishes a mapping relationship between scene features and topology strategies. Different topology optimization objectives are adopted under different load modes: in the "short-term, high-burst" scenario, the primary goal of topology optimization is to minimize communication latency, thus employing a tightly coupled strategy to form multiple independent device pairs through the nearest connection, reducing communication hops; while in the "long-term, low-burst" scenario, topology optimization focuses more on system stability and fault tolerance, using fully connected or mesh topologies to provide redundant paths. This topology construction strategy, directly determined by the first clustering result, then guides subsequent operations on eliminating and retaining neighboring device pairs by setting different preset absolute difference thresholds, ultimately achieving dynamic control over the size of the candidate device set. This scenario-based adaptive topology mechanism ensures optimal system performance under different load conditions.

[0113] The absolute difference is obtained by calculating the absolute difference of the second time difference between neighboring cluster members in the first and second topological connection relationships.

[0114] For each pair of neighboring devices, if the absolute difference result is less than the preset absolute difference threshold, the cluster member with the largest second time difference in the pair of neighboring devices is removed to obtain the removal result.

[0115] If the absolute difference result is greater than or equal to the preset absolute difference threshold, then the cluster members in the neighboring device pair are retained to obtain the retention result;

[0116] The first topological connection relationship is updated based on the elimination result and the retention result to obtain the third topological connection relationship;

[0117] The second topological connection relationship is updated based on the elimination result and the retention result to obtain the fourth topological connection relationship;

[0118] Extract shared cluster members from the third and fourth topology connection relationships;

[0119] The shared cluster members are connected according to the shortest distance between devices to construct the candidate device topology connection relationship;

[0120] Cluster members in the candidate device topology connection relationship are identified as candidate devices.

[0121] The leader device and other cluster members in the robot cluster, besides the candidate devices, are designated as follower devices.

[0122] It should be noted that this scheme introduces an absolute difference threshold mechanism to optimize the composition of the candidate device set, balancing communication redundancy and topology diversity in a high-concurrency environment. When the absolute difference of the second time difference between neighboring device pairs is lower than a preset threshold, it indicates that the devices have highly similar latency characteristics, are in the same network congestion area, or share a communication bottleneck link. Retaining such functionally redundant devices would lead to an overblown candidate set, increasing the signaling overhead for state synchronization and heartbeat detection. Furthermore, it could trigger a "split-brain" risk when the leader device fails, causing multiple similar candidate devices to compete simultaneously, thus prolonging the system convergence time. Therefore, this scheme effectively eliminates functional redundancy, optimizes resource utilization, and reduces unnecessary communication contention by removing members with large second time differences from device pairs, without significantly impacting system reliability.

[0123] Conversely, when the absolute difference is greater than or equal to a preset threshold, it reflects a significant difference in latency performance between devices, stemming from different network environments or communication paths. For example, one device might be located on a core network link while another is in an edge region. Although the absolute performance of a high-latency device is poor, the topological diversity it represents provides heterogeneous communication paths for the system. In extreme cases where high concurrency leads to congestion or interruption of the main path, this device becomes a critical backup connection point. Therefore, this solution chooses to retain all members in the device pair to maintain topological diversity and avoid a decrease in fault tolerance due to over-reliance on a single communication path.

[0124] The preset absolute difference threshold is dynamically set based on the standard deviation of the historical second time difference dataset multiplied by an empirical coefficient η (η ranges from 1.0 to 2.0). The value of this empirical coefficient η is determined by the first clustering result to achieve adaptive adjustment in conjunction with the topology strategy. Specifically, when identified as a "short-term high-burst" scenario and employing a tightly coupled strategy, a small η value (e.g., η≈1.0) is selected to generate a small absolute difference threshold, facilitating aggressive redundancy removal. Conversely, when identified as a "long-term low-burst" scenario and employing a non-tightly coupled strategy, a large η value (e.g., η≈2.0) is selected to generate a large absolute difference threshold, supporting the retention of more nodes to maintain topology diversity.

[0125] The two types of topological connections obtained after absolute difference filtering carry different optimization objectives. The third topological connection originates from the optimization of the first topology. Its construction is based on the K-means clustering results of the periodic characteristics (including duration and suddenness) of high-concurrency scenarios. The core optimization objective is the proximity of physical locations, forming a physical backbone network with the lowest transmission latency. After eliminating redundant nodes, the third topology ensures that each node retained is both critical and non-redundant in terms of physical layout.

[0126] Accordingly, the fourth topology evolved from the second topology, based on DBSCAN's clustering analysis of the second time difference data. The optimization focus is on the reliability of communication delay, selecting a group of devices with stable communication quality. After correction through the absolute difference screening process, the fourth topology eliminates internal redundancy in the communication reliability dimension, while also taking into account a moderate diversity of delay characteristics, enhancing the overall robustness of the candidate echelon.

[0127] Shared cluster members refer to device nodes that exist in both optimized topologies, such as core devices in an industrial robot platoon that satisfy both optimal physical location and stable communication capabilities. These shared members are reconnected with the nearest-distance topology, forming a set of candidate devices that combines physical proximity and communication reliability. N candidate devices are selected to establish a hierarchical backup mechanism, enabling rapid role switching when the leader device fails. Follower devices act as command execution terminals, responsible for completing specific operational tasks under the leader device's scheduling, ensuring the system maintains basic functionality even under sudden traffic surges. This multi-level device architecture, through precise latency difference analysis and topology optimization, significantly improves the fault tolerance and operational efficiency of distributed systems in high-concurrency scenarios.

[0128] S3: Determine the demotion rate of the leader device, the promotion rate of the candidate device, and the promotion rate of the follower device in each information data processing cycle under high concurrency scenarios based on the leader device, candidate device, and follower device;

[0129] Monitor the role status changes of all cluster members in the robot cluster, and collect the number of times the leader device becomes a candidate device, the number of times a candidate device becomes a leader device, the number of times a follower device becomes a candidate device, and the total number of role transitions;

[0130] The leadership device downgrade rate is obtained by dividing the number of times a leadership device is converted to a candidate device by the total number of role conversions.

[0131] The candidate device promotion rate is obtained by dividing the number of times a candidate device is converted into a leadership device by the total number of role conversions.

[0132] The promotion rate of the follower device is obtained by dividing the number of times the follower device is converted into a candidate device by the total number of role conversions.

[0133] It should be noted that in high-concurrency distributed systems, traditional static configuration mechanisms often lead to disordered system state switching due to a lack of quantitative perception of dynamic changes in node roles. This step establishes an information data processing cycle mechanism to construct a complete role transition monitoring system. This cycle is defined as a fixed time window for the system to collect role status and evaluate performance, continuously monitoring the role state migration of cluster members within each cycle. The leader device degradation rate is obtained by statistically analyzing the proportion of times a leader device is converted to a candidate device in the total number of role transitions, reflecting the stability of the master node; the candidate device promotion rate is obtained by calculating the proportion of candidate devices that successfully convert to leader devices, characterizing the system's fault tolerance capability; the follower device promotion rate is obtained by statistically analyzing the frequency of follower devices being upgraded to candidate devices, reflecting the activity of underlying nodes. By recording the three key transition behaviors—leader to candidate, candidate to leader, and follower to candidate—and the total number of transitions, a quantitative evaluation result of the role transition rate is formed. Each threshold setting is based on statistical analysis of historical operating data: the preset downgrade threshold is taken as the average downgrade rate of historical leading equipment plus two standard deviations; the preset promotion threshold is set as 90% of the candidate equipment promotion rate in the benchmark test; and the preset disturbance threshold is set as the average promotion rate of historical following equipment plus one standard deviation. This multi-dimensional conversion rate monitoring mechanism provides accurate data support for subsequent dynamic parameter adjustments.

[0134] S4: Adjust the information data processing flow for the next information data processing cycle based on the aforementioned leadership equipment downgrade rate, candidate equipment promotion rate, and follower equipment promotion rate.

[0135] When the degradation rate of leadership equipment exceeds the preset degradation threshold, the preset final score threshold will be increased in the next information data processing cycle.

[0136] When the candidate device promotion rate is lower than the preset promotion threshold, the preset final score threshold will be lowered in the next information data processing cycle.

[0137] When the follower device's advancement rate exceeds the preset disturbance threshold, the neighborhood radius parameter and minimum sample number of the DBSCAN clustering algorithm are tightened in the next information data processing cycle, and the preset absolute difference threshold is reduced.

[0138] It should be noted that traditional distributed systems, lacking effective closed-loop adjustment mechanisms, struggle to cope with dynamic characteristic changes under high concurrency scenarios. This step, based on the conversion rate metric obtained in stage S3, implements precise parameter adjustments in the next information data processing cycle. When the leader device degradation rate exceeds the preset degradation threshold, it indicates a lenient tendency in the current leader election criteria. The selection criteria for leader nodes are strengthened by increasing the preset final scoring threshold in the next cycle. When the candidate device promotion rate consistently falls below the preset promotion threshold, it indicates insufficient vitality of the candidate mechanism. The selection range of candidate devices is expanded by lowering the scoring threshold. When the follower device promotion rate exceeds the preset disturbance threshold, it signifies abnormal fluctuations in the underlying topology. Clustering stability is improved by tightening the neighborhood radius parameter and minimum sample size of the DBSCAN clustering algorithm, while the preset absolute difference threshold is lowered to enhance the reliability verification of topology connections. This dynamic parameter optimization mechanism based on the cycle conversion rate forms a complete closed-loop control system, solving the problem of system parameter mismatch under sudden traffic surges in traditional static configuration methods and significantly improving the adaptive capability of distributed systems under dynamic loads.

[0139] Example 2: Based on Example 1, an information data processing system applied to high-concurrency scenarios, such as... Figure 2 As shown, it includes:

[0140] The time difference acquisition and delay assessment module is used to obtain the command sending and receiving time, calculate the first time difference and the second time difference, and evaluate the initial response speed and complete reception reliability of the cluster members respectively.

[0141] The three-layer device dynamic decision-making module determines the leader device, candidate devices, and follower devices based on time difference, location, energy, and positioning accuracy data through a scoring mechanism and topology analysis.

[0142] The role conversion rate monitoring and analysis module monitors the number of device role conversions, calculates the demotion rate of leading devices, the promotion rate of candidate devices, and the promotion rate of following devices, and triggers parameter adjustment instructions based on thresholds.

[0143] The parameter adaptive optimization execution module dynamically adjusts the scoring threshold, clustering parameters, and absolute difference standard based on the conversion rate analysis results, thereby optimizing the data processing strategy for the next processing cycle.

[0144] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An information data processing method applied in high-concurrency scenarios, characterized in that, Includes the following steps: S1: Obtain the initial information reception time and final information reception time of each cluster member in the robot cluster under high concurrency scenarios. The initial information reception time refers to the instant when the cluster member successfully parses the command data packet for the first time, and the final information reception time corresponds to the termination time when the member completes the reception and verification of all command data. Extract the command instruction sending time of the robot cluster, including: calculating the time difference between the command instruction sending time and the initial information reception time of each cluster member in the robot cluster as a first time difference; the first time difference is used to characterize the delay time from the issuance of the command instruction to the initial receipt of the command by the cluster member, reflecting the initial response speed of the cluster member to the command; calculate the time difference between the command instruction sending time and the final information reception time of each cluster member in the robot cluster as a second time difference; the second time difference is used to characterize the delay time from the issuance of the command instruction to the final receipt of the command by the cluster member, reflecting the complete reception reliability of the command by the cluster member. S2: Determining the leader device, candidate devices, and follower devices based on the initial information reception time, the final information reception time, and the command / instruction transmission time includes: determining the leader device based on the first time difference combined with the position information, remaining energy, and positioning accuracy of each cluster member in the robot cluster; defining cluster members in the robot cluster other than the leader device as candidate members; determining candidate devices based on the second time difference combined with the relative position information of the candidate members relative to the leader device; and designating the leader device and cluster members in the robot cluster other than the candidate devices as follower devices. S3: Determine the demotion rate of the leader device, the promotion rate of the candidate device, and the promotion rate of the follower device in each information data processing cycle under high concurrency scenarios based on the leader device, candidate device, and follower device; S4: Adjust the information data processing flow for the next information data processing cycle based on the aforementioned leadership equipment downgrade rate, candidate equipment promotion rate, and follower equipment promotion rate.

2. The information data processing method applied to high-concurrency scenarios according to claim 1, characterized in that, The step of determining the leader device based on the first time difference combined with the position information, remaining energy, and positioning accuracy of each member in the robot swarm includes: The strategic value of the location information within the cluster is scored and normalized to obtain a normalized location score. The remaining energy and positioning accuracy are also normalized to obtain normalized remaining energy and normalized positioning accuracy. An initial score is obtained using an initial scoring formula. Based on the initial score, candidate leadership devices are selected. The initial scoring formula is as follows: ； in, For the initial score, For cluster members The first time difference It is the minimum of the first time differences among all cluster members. The maximum value among the first time differences of all cluster members. For cluster members Normalized position score, For cluster members Normalized residual energy, For cluster members Normalized positioning accuracy, , , and For the weighting coefficients, satisfying + + + =1; Based on the initial score, cluster members that exceed the preset initial score threshold will be selected as candidate leader devices.

3. The information data processing method applied in high-concurrency scenarios according to claim 2, characterized in that, The step of selecting cluster members with scores greater than a preset initial score threshold as candidate leader devices based on the initial score includes: The logical position ranking of the candidate leader devices in the cluster is determined based on the structural value of the devices in the network topology and the cost of leadership switching. Second time difference data of all candidate leader devices are collected, and a folding coefficient is calculated for each candidate leader device. The formula for the folding coefficient is: ； in, The folding factor is... For current cluster members The order of leadership, and , "Time" indicates the highest priority. When indicates the lowest priority, This represents the total number of devices in the cluster. For summing loop variables, For cluster members The second time difference, The minimum of the second time differences among all candidate leadership devices. The maximum value among the second time differences of all candidate leadership devices. The sensitivity coefficient, , For the weighting coefficients, satisfying + =1; The folding coefficient is used to quantify the combined risk of system convergence delay and switching failure caused by communication competition and delay due to data flooding when switching cluster leadership to a backup leader device in high-concurrency scenarios. The final score for each candidate leadership device is obtained based on the folding coefficient using the final scoring formula, which is: ； in, For the final score, For adjustment factors; Based on the final score, candidate leadership devices that exceed the preset final score threshold are identified as leadership devices.

4. The information data processing method for high-concurrency scenarios according to claim 1, characterized in that, The step of determining candidate devices based on the second time difference and the relative position information of cluster members (excluding the leader device) relative to the leader device includes: The relative position distance set is obtained by calculating the relative position distance between each cluster member (excluding the leader device) and the leader device in the robot cluster. K-means clustering was performed on N high-concurrency scenarios based on their cycle duration and suddenness to obtain the first clustering result; The second time difference data of each cluster member in the robot cluster, excluding the leader device, is obtained and defined as cluster data. The cluster data is then subjected to DBSCAN clustering to obtain the second clustering result. Based on the first clustering result, the cluster members in the relative position distance set are connected according to the shortest distance between devices to obtain the first topological connection relationship; Based on the second clustering result, cluster members in the relative location distance set are connected according to the shortest distance between devices to obtain a second topological connection relationship; Candidate devices are determined based on the first topology connection relationship and the second topology connection relationship.

5. The information data processing method for high-concurrency scenarios according to claim 4, characterized in that, The step of determining candidate devices based on the first topology connection relationship and the second topology connection relationship includes: The absolute difference is obtained by calculating the absolute difference of the second time difference between neighboring cluster members in the first and second topological connection relationships. For each pair of neighboring devices, if the absolute difference result is less than the preset absolute difference threshold, the cluster member with the largest second time difference in the pair of neighboring devices is removed to obtain the removal result. If the absolute difference result is greater than or equal to the preset absolute difference threshold, then the cluster members in the neighboring device pair are retained to obtain the retention result; The first topological connection relationship is updated based on the elimination result and the retention result to obtain the third topological connection relationship; The second topological connection relationship is updated based on the elimination result and the retention result to obtain the fourth topological connection relationship; Extract shared cluster members from the third and fourth topology connection relationships; The shared cluster members are connected according to the shortest distance between devices to construct the candidate device topology connection relationship; Cluster members in the candidate device topology connection relationship are identified as candidate devices.

6. The information data processing method applied in high-concurrency scenarios according to claim 1, characterized in that, The step of determining the demotion rate of the leader device, the promotion rate of the candidate device, and the promotion rate of the follower device based on each information data processing cycle in a high-concurrency scenario includes: Monitor the role status changes of all cluster members in the robot cluster, and collect the number of times the leader device becomes a candidate device, the number of times a candidate device becomes a leader device, the number of times a follower device becomes a candidate device, and the total number of role transitions; The leadership device downgrade rate is obtained by dividing the number of times a leadership device is converted to a candidate device by the total number of role conversions. The candidate device promotion rate is obtained by dividing the number of times a candidate device is converted into a leadership device by the total number of role conversions. The promotion rate of the follower device is obtained by dividing the number of times the follower device is converted into a candidate device by the total number of role conversions.

7. The information data processing method applied to high-concurrency scenarios according to claim 1, characterized in that, The information data processing flow for adjusting the next information data processing cycle based on the leadership equipment downgrade rate, candidate equipment promotion rate, and follower equipment promotion rate includes: When the degradation rate of leadership equipment exceeds the preset degradation threshold, the preset final score threshold will be increased in the next information data processing cycle. When the candidate device promotion rate is lower than the preset promotion threshold, the preset final score threshold will be lowered in the next information data processing cycle. When the follower device's advancement rate exceeds the preset disturbance threshold, the neighborhood radius parameter and minimum sample number of the DBSCAN clustering algorithm are tightened in the next information data processing cycle, and the preset absolute difference threshold is reduced.

8. An information data processing system for high-concurrency scenarios, used to implement the information data processing method for high-concurrency scenarios as described in any one of claims 1-7, characterized in that, include: The time difference acquisition and delay assessment module is used to obtain the command sending and receiving time, calculate the first time difference and the second time difference, and evaluate the initial response speed and complete reception reliability of the cluster members respectively. The three-layer device dynamic decision-making module determines the leader device, candidate devices, and follower devices based on time difference, location, energy, and positioning accuracy data through a scoring mechanism and topology analysis. The role conversion rate monitoring and analysis module monitors the number of device role conversions, calculates the demotion rate of leading devices, the promotion rate of candidate devices, and the promotion rate of following devices, and triggers parameter adjustment instructions based on thresholds. The parameter adaptive optimization execution module dynamically adjusts the scoring threshold, clustering parameters, and absolute difference standard based on the conversion rate analysis results, thereby optimizing the data processing strategy for the next processing cycle.

Images