Sensor cluster management method, system and device based on industrial internet of things

Abstract

The application discloses a kind of sensor cluster management method, system and equipment based on industrial internet of things, it is related to industrial internet of things technical field, method includes: the physical space is divided into multiple subspaces and obtains internal perception node;Theoretical and actual coordinates and attitude angle of each node are extracted;Unit offset value is calculated in combination with sensitivity matrix, and then group offset value and anchoring drift rate representing mechanical stability are obtained;According to the time point of node reporting, a typical time difference is extracted, and a standard time difference is calculated in combination with bottom layer backoff and retransmission delay, to obtain network initial coordination value;According to the difference between actual and theoretical attitude angle, an attitude penalty value is obtained;Anchoring drift rate, attitude penalty value and initial coordination value are integrated to obtain the target coordination value of each subspace;Finally, the total coordination value is calculated in combination with space weight, and the bottom control instruction is triggered accordingly.The application has the advantages of precise quantification, multidimensional fusion and closed-loop intervention.

Description

Technical Field

[0001] This invention relates to the field of industrial Internet of Things (IoT) technology, and specifically to a sensor cluster management method, system, and device based on industrial IoT. Background Technology

[0002] In modern industrial production, a large number of sensor nodes are usually densely deployed on the surface of core equipment such as large rotating machinery to build an Internet of Things (IoT) sensing topology covering the entire working area, thereby collecting equipment operating status data.

[0003] However, due to high-frequency mechanical vibrations and complex electromagnetic interference in industrial production environments, sensors are at risk of fastener loosening during operation. This can cause the sensor's sensing center to deviate from its preset theoretical installation coordinates and undergo spatial orientation deviation. This physical displacement is anisotropic in three-dimensional space; for example, radial offset along a motor has a greater impact on sensing accuracy than axial offset. However, existing cluster management mechanisms mostly rely on static topological coordinates or simple scalar geometric distances, making it difficult to accurately quantify this direction-sensitive mechanical anchoring state. Simultaneously, changes in the sensor's spatial orientation can cause antenna pointing deflection, resulting in cross-polarization attenuation and multipath effects in the radio frequency signal, ultimately degrading the quality of the underlying communication link.

[0004] In communications, the increased asynchronous data transmission and latency in industrial wireless networks are primarily due to channel contention backoff at the medium access control layer and anticipated retransmissions under complex electromagnetic conditions, rather than simply physical transmission distance limitations. Existing methods fail to effectively map the congestion mechanisms at the underlying communication protocol level to data reporting time differences. The lack of a quantitative model that mathematically couples anisotropic displacement caused by mechanical vibration, antenna polarization loss, and network layer communication timing makes it difficult for management centers to accurately assess the effectiveness of multi-node collaborative sensing topologies within a local space. This results in the system's inability to output accurate collaborative performance indicators in a timely manner when faced with hardware malfunctions or local channel congestion. Furthermore, it hinders the ability to trigger closed-loop intervention actions such as redundant sensor takeover or issuing speed-down commands to the programmable logic controller based on quantification results, thus limiting the sensing reliability and equipment control capabilities of the Industrial Internet of Things (IIoT) under complex operating conditions. Summary of the Invention

[0005] To address the technical problems in the background art, the present invention provides a sensor cluster management method, system and device based on the Industrial Internet of Things. By quantifying the correlation between anisotropic physical displacement caused by mechanical vibration of equipment, antenna polarization attenuation and communication timing of the medium access control layer, the collaborative sensing status of multiple nodes is evaluated and control commands are triggered.

[0006] A sensor cluster management method based on the Industrial Internet of Things (IIoT) includes: acquiring a physical space and dividing the physical space into multiple subspaces; acquiring multiple sensing nodes within each subspace; acquiring the theoretical coordinates and theoretical attitude angles of each sensing node, and acquiring the actual coordinates and actual attitude angles of each sensing node; acquiring a sensitivity matrix, and acquiring a unit offset value based on the theoretical coordinates, the actual coordinates, and the sensitivity matrix; acquiring a group offset value based on the unit offset values ​​of all sensing nodes, and acquiring an anchoring drift rate based on the group offset value and a standard offset value; and acquiring the reporting time point of each sensing node. The system obtains the typical time difference for each subspace based on the reported time point; obtains the backoff delay and retransmission delay, and obtains the standard time difference based on the backoff delay and retransmission delay; obtains the initial coordination value for each subspace based on the typical time difference and the standard time difference; obtains the attitude penalty value based on the theoretical attitude angle and the actual attitude angle; obtains the target coordination value for each subspace based on the anchoring drift rate, the attitude penalty value, and the initial coordination value for each subspace; obtains the total coordination value based on the target coordination value for each subspace and the spatial weight; and triggers control commands based on the total coordination value.

[0007] Optionally, obtaining the unit offset value based on the theoretical coordinates, the actual coordinates, and the sensitivity matrix includes: obtaining the axis weight values ​​in the sensitivity matrix; obtaining the coordinate difference between the actual coordinates and the theoretical coordinates; multiplying the squares of each coordinate difference with its corresponding axis weight value and summing the results; and taking the square root of the summation result to obtain the unit offset value.

[0008] Optionally, obtaining the anchoring drift rate based on the group offset value and the standard offset value includes: dividing the group offset value by the standard offset value to obtain the offset ratio; subtracting the offset ratio from the value to obtain the difference result; and extracting the maximum value between the difference result and the value zero as the anchoring drift rate.

[0009] Optionally, obtaining the typical time difference of each subspace based on the reporting time point includes: calculating the absolute value of the difference between the reporting time points of any two sensing nodes in the same subspace; and selecting the absolute value of the difference with the largest value as the typical time difference of that subspace.

[0010] Optionally, obtaining the backoff delay and retransmission delay, and obtaining the standard time difference based on the backoff delay and the retransmission delay, includes: obtaining the number of active nodes and the interference index; determining the backoff delay based on the number of active nodes and the contention window mechanism; determining the retransmission delay based on the interference index; obtaining the buffer delay, and adding the backoff delay, the retransmission delay and the buffer delay to obtain the standard time difference.

[0011] Optionally, obtaining the initial coordination value of each subspace based on the typical time difference and the standard time difference includes: dividing the typical time difference by the standard time difference to obtain the proportion to be processed; subtracting the proportion to be processed from the value to obtain the time difference amount; if the time difference amount is greater than zero, using the time difference amount as the initial coordination value; if the time difference amount is not greater than zero, using the value of zero as the initial coordination value.

[0012] Optionally, obtaining the attitude penalty value based on the theoretical attitude angle and the actual attitude angle includes: calculating the angle difference between the actual attitude angle and the theoretical attitude angle to obtain the attitude offset angle; and when the attitude offset angle exceeds the matching threshold, performing a linear mapping on the attitude offset angle to obtain the attitude penalty value.

[0013] Optionally, obtaining the target coordination value for each subspace based on the anchoring drift rate, the attitude penalty value, and the initial coordination value for each subspace includes: multiplying the compensation coefficient by the anchoring drift rate, adding a value of 1, and subtracting the attitude penalty value to obtain a correction coefficient value; multiplying the correction coefficient value by the initial coordination value to obtain an evaluation result; and extracting the minimum value between the evaluation result and the value of 1 as the target coordination value.

[0014] A sensor cluster management system based on the Industrial Internet of Things (IIoT) is also provided. The system includes a management platform, a sensor network platform, and an object platform connected in sequence. The management platform includes: an acquisition module for acquiring a physical space, dividing the physical space into multiple subspaces; acquiring multiple sensing nodes within each subspace; acquiring the theoretical coordinates and theoretical attitude angles of each sensing node; and acquiring the actual coordinates and actual attitude angles of each sensing node; and a first processing module for acquiring a sensitivity matrix, acquiring a unit offset value based on the theoretical coordinates, the actual coordinates, and the sensitivity matrix; acquiring a group offset value based on the unit offset values ​​of all sensing nodes; and acquiring the anchoring drift based on the group offset value and a standard offset value. The first processing module is used to obtain the reporting time point of each sensing node, and obtain the typical time difference of each subspace based on the reporting time point; obtain the backoff delay and retransmission delay, and obtain the standard time difference based on the backoff delay and retransmission delay; obtain the initial coordination value of each subspace based on the typical time difference and the standard time difference; the second processing module is used to obtain the attitude penalty value based on the theoretical attitude angle and the actual attitude angle; obtain the target coordination value of each subspace based on the anchor drift rate, the attitude penalty value and the initial coordination value of each subspace; obtain the total coordination value based on the target coordination value of each subspace and the spatial weight, and trigger control commands based on the total coordination value.

[0015] An electronic device is also provided, comprising: a memory having a computer program stored thereon; and a processor for executing the computer program in the memory to implement a sensor cluster management method based on the Industrial Internet of Things.

[0016] The beneficial effects of this invention are reflected in: In the entire sensor cluster management method based on the Industrial Internet of Things (IIoT), by dividing the physical space and capturing the actual coordinates and attitude angles of the sensing nodes in real time, the system reduces the errors caused by static topology blind spots and obtains the actual physical state benchmark of the equipment. A mechanical sensitivity matrix is ​​introduced, and anisotropic unit offset distance is calculated by assigning different weights to the three-dimensional coordinate axes, deriving the anchoring drift rate. This mechanism aligns with the physical law that industrial machinery is sensitive to displacement in specific directions, improving the accuracy of assessing hardware loosening. Regarding communication coordination assessment, based on the underlying industrial wireless protocol, a standard time difference is constructed using the channel contention backoff delay and electromagnetic interference expected retransmission delay of the medium access control layer. This standard time difference is then used to measure the actual typical time difference and calculate... The initial coordination value can reflect the congestion and asynchronous state of the local network. This scheme further extracts the attitude penalty factor by the deviation between the theoretical and actual attitude angles, transforms the physical deflection at the hardware level into the radio frequency polarization attenuation, and numerically couples the anchor drift rate, attitude penalty value and initial coordination value, so that the obtained target coordination value combines the mechanical stability state and the communication link quality, realizing the fusion quantification of multiple physical field parameters. Based on the weighted aggregated total coordination value, the system can trigger the wake-up and takeover of redundant nodes or issue control commands such as reducing the speed, thereby reducing the risk of perception blind spots and controlling the operating status of mechanical equipment. It constructs a feedback link from quantitative evaluation to the underlying hardware execution, improving the operational stability of the asset monitoring network and the level of equipment operation and maintenance management. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0018] Figure 1 This is a schematic diagram illustrating the steps of the sensor cluster management method based on the Industrial Internet of Things of the present invention; Figure 2 This is a schematic diagram of some steps in S1 of the sensor cluster management method based on the Industrial Internet of Things of the present invention; Figure 3 This is a schematic diagram of some steps in S2 of the sensor cluster management method based on the Industrial Internet of Things of the present invention; Figure 4This is a schematic diagram of some steps in S3 of the sensor cluster management method based on the Industrial Internet of Things of the present invention; Figure 5 This is a schematic diagram of some steps in S4 of the sensor cluster management method based on the Industrial Internet of Things of the present invention; Figure 6 This is a schematic diagram illustrating the composition of the sensor cluster management system based on the Industrial Internet of Things (IIoT) of this invention. Figure 7 This is a schematic diagram of the optimized industrial Internet of Things (IoT) involved in this invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0021] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0022] This invention provides a sensor cluster management method based on the Industrial Internet of Things, such as... Figure 1 As shown, in one embodiment, the method includes: S1. Obtain the physical space and divide it into multiple subspaces; obtain multiple sensing nodes within each subspace; obtain the theoretical coordinates and theoretical attitude angles of each sensing node; obtain the actual coordinates and actual attitude angles of each sensing node. S2. Obtain the sensitivity matrix. Based on the theoretical coordinates, actual coordinates, and sensitivity matrix, obtain the unit offset value. Obtain the group offset value based on the unit offset values ​​of all sensing nodes. Obtain the anchoring drift rate based on the group offset value and the standard offset value. S3. Obtain the reporting time point of each sensing node, and obtain the typical time difference of each subspace based on the reporting time point; obtain the backoff delay and retransmission delay, and obtain the standard time difference based on the backoff delay and retransmission delay; obtain the initial coordination value of each subspace based on the typical time difference and the standard time difference. S4. Obtain the attitude penalty value based on the theoretical attitude angle and the actual attitude angle; obtain the target coordination value of each subspace based on the anchoring drift rate, the attitude penalty value and the initial coordination value of each subspace; obtain the total coordination value based on the target coordination value and the spatial weight of each subspace; and trigger the control command based on the total coordination value.

[0023] In this embodiment, it should be noted that in S1, the industrial physical space is evenly divided into multiple subspaces of equal volume, and a spatial coordinate system is established. The theoretical coordinates and theoretical attitude angles of each sensing node are obtained from the underlying database, and the actual coordinates and actual attitude angles are obtained through the positioning and calculation module.

[0024] Taking a large rotating motor operating workshop as an example, a coordinate system is established with the southwest corner of the main bearing fixed base as the origin, dividing the monitoring area into three subspaces. In the first subspace, located in the core high-frequency vibration zone, the theoretical coordinates of the sensing node numbered Node_1 are extracted as (150.0, 200.0, 50.0) mm, and the yaw angle in the theoretical attitude angle is 90.0 degrees. During full-load operation of the equipment, due to the slippage of fasteners caused by mechanical vibration, the actual coordinates of this node are acquired in real time as (150.4, 201.2, 50.3) mm, and the actual yaw angle shifts to 98.5 degrees. The reason for adopting this acquisition and spatial division logic is that existing cluster management relies heavily on static preset network topologies, making it difficult to perceive the real physical changes caused by the industrial environment to the hardware equipment. By capturing the three-dimensional coordinate deviation and the angle change of spatial attitude in real time, the problem of static topology blind spots caused by hardware loosening is solved. This step obtains objective baseline data on the operating status of the equipment from the physical source, so that subsequent quantitative evaluations are no longer based on ideal models, providing basic data support for improving the reliability of the overall sensing network.

[0025] In S2, the scattered displacement deviations are transformed into quantifiable mechanical anchoring state indices. This step obtains a sensitivity matrix containing weight coefficients for the three-dimensional coordinate axes, calculates the difference between the actual and theoretical coordinates, and uses a formula to obtain the anisotropic unit offset value.

[0026] Based on application scenario data, a high weight was assigned to the radial vibration of the motor, with the X-axis weight set at 1.0, the Y-axis weight at 2.5, and the Z-axis weight at 0.8. Using the coordinate differences of Node_1 (0.4, 1.2, and 0.3 mm), the weighted summation and square root formula yielded a unit offset value of approximately 1.957 mm. Subsequently, the arithmetic mean of the five nodes in the first subspace (including the remaining nodes' offsets of 1.820, 2.105, 1.750, and 2.168 mm) was calculated, resulting in a group offset value of 1.960 mm. After obtaining the standard offset value of 5.000 mm, the anchorage drift rate was calculated using the maximum value function. Subtracting the ratio of 1.960 to 5.000 from 1 yielded a result of 0.608. This calculation logic, incorporating a sensitivity matrix, is adopted because displacements in different directions within mechanical equipment have varying degrees of interference with the sensing data. Radial offset typically carries a higher risk than axial offset, and simple Euclidean distance is insufficient to reflect this industrial phenomenon. The maximum value function in the formula sets a lower limit for physical tolerance. When the actual average offset exceeds the standard tolerance, the result is truncated to zero to prevent negative outputs from causing calculation errors. This step solves the problem that existing scalar distance methods are difficult to use for directional assessment of mechanical looseness, achieving accurate quantification of the stable state of sensor cluster hardware and making the assessment of physical offset consistent with the actual operating characteristics of industrial equipment.

[0027] In S3, after quantization of the physical domain is completed, the timing coordination at the network communication layer is evaluated, and the initial coordination value is output. This step extracts the data reporting time points of all nodes in the same subspace, calculates the maximum time difference as the typical time difference, constructs a standard time difference based on the underlying parameters of the wireless protocol, and finally calculates the initial coordination state through the ratio.

[0028] In the application scenario, the reporting times of the five nodes in the first subspace are 12.5, 14.0, 28.5, 16.2, and 13.8 milliseconds, respectively, resulting in a maximum time difference, or typical time difference, of 16.0 milliseconds. Subsequently, given that the number of currently active nodes is 5 and the background interference index is high, the backoff delay is calculated to be 6.5 milliseconds using the contention window mechanism of the media access control layer. The expected retransmission delay caused by electromagnetic interference is 11.0 milliseconds, and adding the protocol's inherent buffer delay of 2.5 milliseconds, the sum yields a standard time difference of 20.0 milliseconds. Subtracting the ratio of 16.0 to 20.0 from 1 gives an initial coordination value of 0.200. This logic of calculating the standard time difference based on underlying protocol parameters is adopted because, within the limited space of an industrial workshop, the latency caused by electromagnetic wave transmission distance is relatively low. The real cause of data asynchrony is multi-node contention for the channel and data packet retransmissions caused by background electromagnetic interference. This logic addresses the shortcomings of existing methods that use physical transmission distance and propagation rate to measure wireless communication performance. It accurately maps the root causes of data latency to objective evaluation benchmarks, enabling the calculated collaborative value to truly reflect the congestion level of local wireless networks in industrial interference environments.

[0029] In S4, the aforementioned mechanical anchoring physical state is numerically coupled across dimensions with the network-level communication coordination performance to output the target coordination value. This step first calculates the angle difference between the actual and theoretical attitude angles. If it exceeds the matching threshold, an antenna attitude penalty factor is generated. Then, the compensation coefficient, anchoring drift rate, and penalty factor are combined into a correction coefficient, which is multiplied by the initial coordination value to obtain the target result.

[0030] Based on specific data, the actual yaw angle of Node_1, 98.5 degrees, results in an 8.5-degree attitude offset angle compared to the theoretical 90.0 degrees. Since this exceeds the matching threshold of 5.0 degrees, a penalty weight of 0.05 is added for every 1.0-degree exceedance, resulting in an antenna attitude penalty factor of 0.175. Setting the compensation coefficient to 0.5, the target cooperative value is calculated using the minimum value function formula: the correction coefficient is 1 plus 0.5 multiplied by 0.608, then subtracted by 0.175, yielding 1.129. Multiplying this by the initial cooperative value of 0.200 gives an evaluation result of 0.2258. After upper limit truncation, the target cooperative value in the first subspace is 0.2258. This coupled calculation logic is adopted because the physical flipping of the sensor directly changes the antenna pointing, leading to RF signal cross-polarization loss. This physical change is a direct cause of increased network retransmission and latency. By using physical stability as positive compensation and antenna deflection as a negative penalty factor, the problem of fragmented monitoring data from different dimensions is solved. Meanwhile, the minimum value function prevents mathematical errors caused by excessively large correction coefficients leading to the collaborative value exceeding the maximum score limit. This step establishes a quantitative connection between the physical layer and the network layer, providing a unified evaluation metric that considers both hardware status and communication quality.

[0031] The final step of S4 is to calculate the overall collaborative efficiency of the entire physical space through weighted aggregation calculation, and to trigger closed-loop intervention commands for automated industrial equipment based on the quantification result. This step assigns spatial weights to different subspaces, sums the target collaborative values ​​of each subspace by weight, and executes control operations when the result is lower than a preset safety threshold.

[0032] In the application scenario, the target coordination values ​​for the second and third subspaces were 0.8500 and 0.9200, respectively. Based on the degree of process criticality, the first subspace of the main bearing core area was assigned the highest weight of 0.60, while the second and third subspaces were assigned weights of 0.25 and 0.15, respectively. Through weighted multiplication and summation, the total coordination value was calculated to be 0.48598. Since this value is lower than the industrial safety operating threshold of 0.65000, a control command was immediately triggered, reducing the operating speed of the rotary motor from 3000.0 RPM to 1500.0 RPM via the programmable logic controller, while simultaneously activating the three redundant sensing nodes in this area to take over the task. This spatial weight-based aggregation logic was adopted because in complex industrial environments, the impact of different mechanical areas on overall safe production is not equal. The monitoring weight of the core component area should be higher than that of the peripheral auxiliary areas, and the weighted summation reflects this objective difference.

[0033] In summary, the entire sensor cluster management method based on the Industrial Internet of Things (IIoT) reduces errors caused by static topology blind spots by dividing the physical space and capturing the actual coordinates and attitude angles of sensing nodes in real time, thus obtaining a benchmark of the actual physical state of the equipment. A mechanical sensitivity matrix is ​​introduced, and anisotropic unit offset distance is calculated by assigning different weights to the three-dimensional coordinate axes, deriving the anchoring drift rate. This mechanism aligns with the physical law that industrial machinery is sensitive to displacement in specific directions, improving the accuracy of assessing hardware loosening. Regarding communication coordination assessment, based on the underlying industrial wireless protocol, a standard time difference is constructed using the channel contention backoff delay and electromagnetic interference expected retransmission delay of the medium access control layer, thereby measuring the actual typical time difference. The calculated initial coordination value reflects the congestion and asynchronous state of the local network. This scheme further extracts the attitude penalty factor by the deviation between theoretical and actual attitude angles, transforms the physical deflection at the hardware level into radio frequency polarization attenuation, and numerically couples the anchor drift rate, attitude penalty value, and initial coordination value. This results in a target coordination value that combines mechanical stability and communication link quality, achieving the fusion and quantification of multiple physical field parameters. Based on the weighted aggregated total coordination value, it can trigger the wake-up and takeover of redundant nodes or issue control commands such as reducing speed. This reduces the risk of perception blind spots while controlling the operating status of mechanical equipment, constructing a feedback link from quantitative evaluation to underlying hardware execution, and improving the operational stability of the asset monitoring network and the level of equipment operation and maintenance management.

[0034] like Figure 2 As shown, in one specific embodiment, S1 includes: S11, obtaining the physical space. A spatial coordinate system is established with the core equipment reference installation point within the physical space as the origin.

[0035] S12. Divide the entire physical space into multiple subspaces evenly according to a preset physical volume threshold. The division process ensures that each subspace has the same volume. The specific method for determining the preset physical volume threshold is as follows: based on the measured data of the radio signal coverage range and the effective radius of mechanical high-frequency vibration transmission of a limited number of previous IoT node deployments in the target workshop, the data is processed using a spatial clustering algorithm. Specifically, the maximum effective three-dimensional coverage envelope volume of each sensor node maintaining stable short-range communication (e.g., packet loss rate less than 1%) during historical network testing is extracted, and combined with the transmission space volume of the vibration wave of the same type of equipment attenuated to a safe level in the rigid base material, the average of the historical intersection space volumes of the two is used as the initial threshold, and then rounded off by engineering. For example, by analyzing the historical data of 10 previous deployments in a workshop, it was found that the optimal local cooperative communication envelope volume of the wireless node under electromagnetic interference is about 12 cubic meters, while the effective interference space of strong mechanical vibration waves is about 15 cubic meters. The system extracts their intersection data and, after safety derating, presets the physical volume threshold to 10 cubic meters, thereby accurately dividing the motor area into 3 subspaces of equal volume.

[0036] S13. Obtain the sensor cluster deployed in the physical space, identify and obtain multiple sensing nodes located in each subspace.

[0037] S14. For each sensing node in the sensor cluster, retrieve its installation location and antenna orientation from the database during the design phase, and obtain the theoretical coordinates and theoretical attitude angles of each sensing node. The theoretical attitude angles include pitch angle, yaw angle, and roll angle.

[0038] S15. Through the positioning and attitude calculation module, the actual position and spatial attitude of the probe sensing center of each sensing node in the spatial coordinate system are collected in real time, and the actual coordinates and actual attitude angles of each sensing node are obtained.

[0039] In this embodiment, it should be noted that step S11 aims to address the technical problem of the lack of a unified spatial quantification benchmark in industrial settings. In complex heavy machinery operating environments, without a fixed reference frame, the positional data of sensing nodes loses physical meaning. Therefore, the physical space is first acquired, and a spatial coordinate system is established using the core equipment reference installation point within this physical space as the origin. Taking a large rotating motor operating workshop as an example, the southwest corner of the fixed base of main bearing No. 2 is selected as the origin of the spatial coordinate system, with coordinates defined as (0.0, 0.0, 0.0) millimeters. The logic behind this setting is that the main bearing base is a relatively stable part of the entire mechanical equipment, less susceptible to external interference. Using it as a benchmark can objectively reflect the relative displacement of other parts during equipment operation. By establishing this three-dimensional coordinate system, the originally ambiguous factory physical space is transformed into a digital grid that can be processed by a computer, overcoming the problem in existing management schemes where node positions cannot correspond to specific mechanical structures. The beneficial effect of this step is that it provides a unified spatiotemporal metric for subsequent anisotropic displacement measurement and communication-based collaborative evaluation, making all physical displacements and attitude deflections traceable references and laying the data foundation for the entire multidimensional quantization model.

[0040] For S12, this step mainly addresses the technical problem that the uneven distribution of environmental interference factors in a large and complex industrial space can easily mask local anomalies in the overall assessment. The entire physical space is uniformly divided into multiple subspaces according to a preset physical volume threshold, and the division process ensures that the volume of each subspace remains equal.

[0041] In the aforementioned motor workshop scenario, the entire monitoring area is uniformly divided into three subspaces. The computational logic behind this uniform division is that if the entire workshop undergoes general data processing, stable data from the edge regions might dilute abnormal fluctuations in the core high-frequency vibration area, leading to a delay in response to local risks. Ensuring equal volume ensures that each subspace has an equal spatial scale benchmark when calculating network parameters such as the standard time difference, avoiding errors caused by differences in spatial size. This step deconstructs the massive industrial workshop into multiple independently manageable and assessable logical units, achieving granular tracking of the sensing network's state. The beneficial effect of this step is improved sensitivity to changes in the local environment, enabling the management center to pinpoint specific physical areas with mechanical loosening or channel congestion, providing spatial structural support for the subsequent allocation of differentiated spatial weights.

[0042] For S13, this step aims to address the technical problem of managing heterogeneous nodes in the Industrial Internet of Things (IIoT) and the difficulty in effectively mapping network devices to specific physical process areas. After completing the spatial partitioning, the sensor cluster deployed in the physical space is acquired, and multiple sensing nodes located in each subspace are identified and acquired.

[0043] In the aforementioned scenario, five sensing nodes for monitoring equipment status were identified and deployed within the first subspace of the core high-frequency vibration zone. This grouping and classification logic was adopted because the communication coordination of industrial sensing networks is largely constrained by physical distance and the local electromagnetic environment. Nodes within the same subspace typically face similar mechanical vibration spectra and channel interference models. Assigning sensing nodes to specific subspaces based on their physical coordinates ensures that subsequent time synchronization analysis and mechanical displacement calculations have the same local environmental premise. This step overcomes the problem of existing methods that only manage nodes by logical network segments, ignoring their physical spatial proximity. The beneficial effect of this step is that it establishes a binding relationship between network topology and physical topology, allowing the behavioral characteristics of multiple nodes within a local area to be aggregated into a whole, thus more objectively reflecting the sensing coverage quality and network coordination status of that specific industrial area.

[0044] For S14, the core of this step is to establish a theoretical reference benchmark for sensor deployment, addressing the technical problem of lacking initial comparison parameters when the equipment's operating state changes. For each sensing node in the sensor cluster, its installation location and antenna orientation during the design phase are retrieved from the database to obtain the theoretical coordinates and theoretical attitude angles of each sensing node. The theoretical attitude angles specifically include pitch angle, yaw angle, and roll angle.

[0045] Based on specific data, the theoretical coordinates of the sensing node numbered Node_1 in the first subspace were extracted from the equipment ledger as (150.0, 200.0, 50.0) mm, with a theoretical yaw angle of 90.0 degrees. This information extraction logic stems from the fact that the probe positions and antenna orientations of industrial equipment sensors are precisely set during factory manufacturing or initial installation and calibration to achieve efficient mechanical signal pickup and radio frequency signal transmission. Extracting these theoretical values ​​as static benchmarks provides a mathematical comparison starting point for subsequent assessments of whether mechanical fasteners have loosened and whether antennas have deflected. This step enables historical comparisons of the evolution of equipment health status. The beneficial effect of this step is that it clarifies the expected sensing topology state before environmental interference occurs, allowing any physical deviations caused by high-frequency mechanical vibrations to be identified and quantified.

[0046] For S15, this step aims to address the technical problem of slow physical drift caused by long-term high-frequency vibration of hardware equipment in industrial environments, which existing static management methods cannot detect in a timely manner. Through the positioning and attitude calculation module, the actual position and spatial attitude of the probe sensing center of each sensing node in the spatial coordinate system are collected in real time, obtaining the actual coordinates and actual attitude angles of each sensing node.

[0047] In applications where the motor is operating at full load, due to continuous mechanical vibration, the actual coordinates of Node_1 changed to (150.4, 201.2, 50.3) mm in real time, while its yaw angle shifted to 98.5 degrees. This dynamic real-time acquisition logic is adopted because the vibration of heavy machinery often has a cumulative deformation effect, and the slippage of fasteners is a gradual process; static topology alone cannot reflect the current true physical boundary. Through high-frequency real-time acquisition, the true three-dimensional distribution state of the sensing network at the current moment can be obtained. This step breaks down the data isolation between the network layer and the physical layer, solving the problem that existing IoT platforms easily treat sensor data with positional shifts as valid data. The beneficial effect of this step is that it gives cluster management the ability to monitor physical deformation in real time, providing a dynamic input source for subsequent calculations of anisotropic unit offset values ​​and antenna attitude penalty factors, ensuring the objectivity of collaborative performance evaluation.

[0048] like Figure 3 As shown, in one specific embodiment, S2 includes: S21, obtaining the sensitivity matrix. Extracting the axis weight values ​​set along the three-dimensional coordinate axes from the sensitivity matrix. Based on the actual coordinates and theoretical coordinates, obtaining the absolute value of the numerical difference between each sensing node on the three-dimensional coordinate axes, as the coordinate difference value. Multiplying the squared value of each coordinate difference by its corresponding axis weight value and summing the results, then taking the square root of the summation result to obtain the unit offset value. The formula for calculating the unit offset value is:

[0049] in, Indicates the first The unit offset value of each sensing node; , , These represent the axis weights along the X, Y, and Z axes in the sensitivity matrix, respectively. , , They represent the first The difference between the actual and theoretical coordinates of a sensing node on the X, Y, and Z axes.

[0050] S22. Sum the unit offset values ​​of all sensing nodes in the sensor cluster, divide the sum by the total number of sensing nodes, and calculate the arithmetic mean. Use this arithmetic mean as the group offset value.

[0051] S23. Obtain the standard offset value. The standard offset value is a distance constant preset based on the equipment's safe operating tolerance.

[0052] S24. Obtain the anchoring drift rate based on the group offset value and the standard offset value. Divide the group offset value by the standard offset value to obtain the offset ratio; subtract the offset ratio from the value to obtain the difference; extract the maximum value between the difference and zero as the anchoring drift rate. The formula for calculating the anchoring drift rate is:

[0053] in, Indicates the anchorage drift rate; This represents the maximum value function, used to output the largest item among multiple values ​​within parentheses; Indicates the group offset value; This represents the standard offset value. In this calculation, when the group offset value is greater than or equal to the standard offset value, the difference result is not greater than zero, the maximum value extracted is zero, that is, the anchoring drift rate is zero.

[0054] In this embodiment, it should be noted that step S21 aims to address the technical problem that existing three-dimensional spatial geometric distance calculations cannot reflect the differences in displacement risk in different directions of industrial machinery and equipment. The axis weight values ​​in the sensitivity matrix are obtained, the coordinate difference between the actual coordinates and the theoretical coordinates is calculated, and the square root of the sum of the squared difference multiplied by the corresponding axis weight value is obtained to obtain the unit offset value.

[0055] Specifically, the calculation expression in S21 used to obtain the unit offset value The core of this calculation process lies in solving the technical problem that existing three-dimensional spatial geometric distance calculations cannot truly reflect the differences in displacement risk of industrial machinery and equipment in different directions.

[0056] In conventional IoT spatial location management, the standard Euclidean distance formula is typically used to measure node offset. This calculation method treats the X, Y, and Z axes in space as equally important, meaning that a one-millimeter offset in any direction yields the same evaluation result. However, in real industrial production environments, taking large rotating motors as an example, the stress structure and manufacturing tolerances of mechanical equipment exhibit significant anisotropy. A small physical displacement along the motor shaft (usually defined as the Z-axis) may simply be due to normal thermal expansion and contraction or fall within the design-allowed axial clearance, having limited impact on overall equipment operation and data sensing. However, if an equivalent displacement occurs along a radial direction perpendicular to the shaft (e.g., the Y-axis), it often indicates bearing wear, substantial loosening of fasteners, or severe rotor imbalance. This radial offset can cause sensor probes to deviate from their optimal vibration measurement area, compromising the effectiveness of data acquisition.

[0057] Therefore, this scheme introduces a sensitivity matrix into the expression, utilizing pre-set shaft weight values ​​for specific mechanical structures. , and As a multiplier, it represents the difference between the actual coordinates and the theoretical coordinates on the corresponding axes. , and The squares are multiplied. In this operational logic, squaring the coordinate difference is not only to eliminate the influence of the algebraic signs of the positive and negative directions, but more importantly, to use the non-linear growth characteristics of the square function to mathematically amplify and penalize large single-axis malignant deviations, which conforms to the physical law that mechanical damage increases non-linearly with displacement. The subsequent square root operation is to reduce the dimension of the weighted amplified value and restore it to a linear length dimension (such as millimeters) consistent with the original physical space, ensuring the uniformity of physical units.

[0058] Based on the specific data of the application scenario, the anisotropy sensitivity weight set for this motor is the X-axis. Radial Y-axis , Axis Z-axis For the sensing node numbered Node_1 located in the first subspace of the high-frequency vibration region, the absolute values ​​of the differences between its actual coordinates and theoretical coordinates are respectively... millimeters millimeters and Millimeters. Substituting into the calculation expression, the operation first calculates the weighted square terms for each axis: the X-axis is... The highly sensitive Y-axis is The low-sensitivity Z-axis is Then, the three values ​​are summed arithmetically to obtain... Finally, the square root operation is performed to obtain the unit offset value. Millimeters. If the existing unweighted Euclidean formula is used, the calculation result is only approximately... Millimeters. Through the calculation of this expression, using... The weighting coefficients objectively amplify the physical risk represented by a radial offset of 1.2 mm, resulting in... Millimeters more closely approximate the actual mechanical failure level currently faced by the node. This calculation process transforms the isotropic pure geometric space into an anisotropic industrial risk assessment space, enabling the underlying physical displacement data to reflect the structural sensitivity characteristics of industrial equipment. This helps to detect structural loosening in dangerous directions at an early stage, providing a physically sound underlying quantitative basis for the subsequent overall stability assessment of the cluster.

[0059] The method for determining the weight values ​​of the three-dimensional coordinate axes in the sensitivity matrix is ​​as follows: Obtain historical fault and maintenance data of the target industrial equipment over a set time period (e.g., the past 12 months), extract the historical average unit displacement of the sensor in the X, Y, and Z axes before each fault occurs, and the probability distribution of the fault. Use the probability percentage of each axis causing system shutdown or performance degradation as the initial reference value for the weights. Then, use the Analytic Hierarchy Process (AHP) for normalization to obtain the final axis weight values. For example, historical data shows that the probability of a motor experiencing a 1 mm abnormal displacement in the radial (Y-axis), axial (Z-axis), and lateral (X-axis) directions causing a shutdown is 58%, 19%, and 23%, respectively. After normalization and fine-tuning with the expert matrix, the system ultimately sets these values ​​to 2.5, 0.8, and 1.0, respectively.

[0060] For step S22, this step primarily addresses the technical problem that occasional displacement reading fluctuations of individual sensor nodes may lead to misjudgments of the overall state of a local area. The unit offset values ​​of all sensing nodes in the sensor cluster are summed, and the sum is divided by the total number of sensing nodes to obtain the arithmetic mean, which is then used as the group offset value. In the first subspace of the application scenario, in addition to Node_1's 1.957 mm, the unit offset values ​​of the other four sensing nodes were simultaneously calculated: 1.820, 2.105, 1.750, and 2.168 mm, respectively. These five values ​​are added together and divided by 5 to obtain the group offset value of 1.960 mm for this subspace. This arithmetic averaging calculation logic is adopted because in industrial settings, a single node may exhibit abnormal offset data due to external impacts or its own local installation tolerances, which does not necessarily indicate structural loosening of the entire mechanical base. By calculating the group offset value, the random errors of individual nodes can be smoothed out, and common characteristics reflecting the overall stability of the mechanical structure within that local area can be extracted. The beneficial effect of this step is that it enhances the statistical significance and resistance to environmental interference of the physical displacement quantification data, making the anchoring drift rate derived from it more stable and avoiding inaccurate control actions triggered by abnormal conditions of individual sensors.

[0061] For S23, this step focuses on addressing the technical challenge of setting objective physical boundaries and safety tolerances for the quantitative assessment of hardware displacement. A standard offset value is obtained, which is a distance constant pre-defined based on the equipment's safe operating tolerances. In the application scenario of rotating electric motors, based on the mechanical vibration specifications and equipment maintenance data for this model, the preset standard offset value is 5.000 mm. This logic of directly retrieving a constant as a reference stems from the fact that each type of heavy machinery has strict tolerance and vibration requirements from the initial design stage. When the physical slippage of the sensor probe exceeds this tolerance, it not only means that the monitoring data may lose its reference value, but also indicates that the underlying fasteners may have experienced fatigue failure. The standard offset value acts as the full-scale scale for the physical domain assessment, transforming the relatively abstract geometric offset distance into an engineering parameter with practical industrial significance. Through this step, differentiated tolerance standards can be set for different types of mechanical equipment or different process environments. The beneficial effect of this step is that it provides a reliable mathematical denominator for calculating the anchorage drift rate, enabling the quantitative model to adapt to different scenarios and ensuring that the assessment results align with actual industrial production and operation specifications.

[0062] The specific method for determining the standard offset value is as follows: by collecting the limit displacement dataset of the same type of mechanical equipment within a limited number of historical normal operation cycles, and performing statistical Gaussian distribution fitting calculation; specifically, the maximum physical slip volume records of the sensor probe in the past 50 complete full-load operation cycles are extracted, and after removing the outliers of accidental external impacts using a filter, the upper limit of the confidence level of the remaining historical slip volume data is calculated, and this is used as the standard offset value of the equipment under the premise of no physical damage; for example, if the maximum normal slip volume of a certain rotating motor in 50 historical cycles is mainly concentrated between 3.5 and 4.5 mm, the system calculates its 99% confidence interval upper limit as 4.95 mm, and after engineering rounding and redundancy addition, the standard offset value is finally set to a fixed 5.000 mm.

[0063] For S24, this step aims to address the technical problem of converting engineering-level displacement data into dimensionless evaluation coefficients that reflect the stable state of the hardware. The group offset value is divided by the standard offset value to obtain the offset ratio. This ratio is then subtracted from the numerical value to obtain the difference. The maximum value between this difference and zero is then extracted as the anchorage drift rate.

[0064] Specifically, the calculation expression in S24 used to obtain the anchorage drift rate The core design of this computational process lies in solving the technical problem of how to set a unified physical stability boundary for hardware devices of different specifications and models, and how to transform dimensional physical displacement data into standardized dimensionless evaluation coefficients that can be used for cross-dimensional fusion calculations. This involves calculating the unit offset value of a single node and obtaining the group average offset value. Finally, what is obtained is a physical quantity expressed in units of length (such as millimeters).

[0065] Because different types of industrial equipment have vastly different tolerances for vibration displacement—for example, a 2mm offset on a heavy-duty forging press might be well within safety tolerances, while the same 2mm offset on a precision CNC machine tool's spindle sensor would signify a serious monitoring failure—directly using absolute physical distance as an evaluation metric cannot establish a lateral comparison benchmark between different devices or different subspaces. To overcome this evaluation barrier caused by equipment differences, the expression introduces a standard offset value. As the denominator. Calculate the ratio. In reality, assessing the actual physical slippage of the current group has already consumed a percentage of the preset mechanical safety tolerance; this is a relative proportion stripped of specific physical units. Because in IoT cluster management logic, evaluation metrics typically follow the normalization principle of "the larger the value, the better the state," using constants... Subtracting this ratio reverses the consumption ratio to the remaining stable margin ratio. Under normal operating conditions, the population offset is small, and the result of this subtraction will be close to... This indicates that the mechanical anchoring of the sensor cluster is in good condition; as the vibration intensifies and the offset increases, this value will decrease linearly.

[0066] However, in harsh industrial environments, machinery may experience sudden and severe structural damage or violent collisions, leading to actual group offset values. Far exceeding the preset standard offset value At this point, the ratio will be greater than Direct subtraction results in a negative number. Allowing negative parameters to enter subsequent network performance multiplication operations would affect the mathematical logic of the entire multidimensional evaluation model, potentially causing regions with extremely poor communication conditions to falsely receive high scores due to the positive-negative-positive formula. Therefore, the expression is wrapped with a function that takes the maximum value. This function acts as a logical protection and truncation mechanism for the algorithm model, constantly comparing the internally calculated difference with a constant. The comparison is performed, and the larger one is forcibly output. This means that once the physical slip exceeds the tolerance limit, a negative number will no longer be output; instead, the anchor drift rate will be rigidly locked to a fixed value. From a mathematical perspective, it can be determined that the mechanical stability of this subspace has failed.

[0067] Based on specific application scenario data, the group offset values ​​of the five sensing nodes in the first subspace were calculated. for Millimeters, while extracting the standard offset value of this model of rotary motor. for Millimeters. Substituting this into the calculation expression, the operation first performs division to obtain the offset ratio: This indicates that the area has consumed the equipment's allowable displacement. Next, perform the subtraction operation: This yields the current stable margin. Finally, the maximum value function is executed: because... Greater than If the current state does not breach the fault tolerance threshold, the function outputs a numerical value. As anchoring drift rate This calculation process, through the design of ratio inversion and extreme value truncation, successfully transforms the specific mechanical deformation phenomenon into a closed-boundary, logically unified process. The interval standardization parameters not only effectively quantify the current mechanical stability life cycle of the cluster, but also ensure the mathematical consistency of the data under extreme conditions, eliminating logical risks for subsequent numerical fusion with time-series indicators at the communication layer.

[0068] like Figure 4 As shown, in one specific implementation, S3 includes: S31, for each subspace, obtaining the reporting time points of all sensing nodes within that subspace within the same monitoring period. Calculating the absolute value of the difference between the reporting time points of any two sensing nodes within the same subspace. After traversing all combinations, selecting the absolute value of the largest difference as the typical time difference for that subspace.

[0069] S32. Obtain the number of active nodes in the current subspace and the interference index of the wireless channel. Determine the backoff delay based on the number of active nodes and the contention window mechanism of the medium access control layer; determine the retransmission delay based on the interference index; extract the buffer delay preset by the communication protocol. Arithmetically add the backoff delay, retransmission delay, and buffer delay to obtain the standard time difference. The specific method for determining the default buffer delay in the communication protocol is as follows: It is calculated based on the historical high-concurrency queue processing time statistics of the underlying wireless communication module's hardware processing chip, combined with a queuing theory model in mathematical statistics. Furthermore, it is obtained by extracting 500 data packet transmission and reception logs from the same batch of IoT sensing nodes under high-pressure laboratory testing conditions, recording the physical time taken for each data transmission from the microcontroller to the RF transmission buffer until the media access control layer is ready. After removing discrete extreme values ​​affected by abnormal system interruptions, the expected queuing time at a 95% confidence level is calculated using a Markov chain model and used as the inherent buffer delay of the protocol. For example, analysis of 500 historical high-load stress test data of a certain type of industrial gateway shows that the queuing and encapsulation time of its underlying data packet buffer is mostly concentrated in the range of 2.1 milliseconds to 2.8 milliseconds. The system uses the model to calculate a stable expected time of 2.45 milliseconds at a 95% confidence level, and after rounding upwards, the default buffer delay is set to a constant value of 2.5 milliseconds.

[0070] S33. Calculate the initial coordination value for each subspace. Divide the typical time difference by the standard time difference to obtain the proportion to be processed. Subtract the proportion to be processed from the value to obtain the time difference amount. Perform a numerical judgment: if the time difference amount is greater than zero, use the time difference amount directly as the initial coordination value; if the time difference amount is not greater than zero, use zero as the initial coordination value.

[0071] In this embodiment, it should be noted that step S31 aims to address the technical problem of accurately quantifying the asynchronous nature of local wireless network data reporting and capturing communication tail delays. The reporting time points of all sensing nodes within the subspace within the same monitoring period are obtained. The absolute value of the difference between the reporting time points of any two nodes is calculated, and the absolute value of the largest difference is selected as the typical time difference for that subspace. In the scene data, the reporting timestamps of the five nodes in the first subspace within the monitoring period are 12.5, 14.0, 28.5, 16.2, and 13.8 milliseconds, respectively. The differences are calculated by iterating through and combining the values, and the maximum difference of 16.0 milliseconds between the third node (28.5 milliseconds) and the first node (12.5 milliseconds) is extracted as the typical time difference for this area.

[0072] The calculation logic of extracting the maximum difference is adopted because, in industrial IoT monitoring tasks, the last data packet arriving in a local network often determines the data alignment efficiency of the application layer algorithm. Using only the average time difference would mask timing lags caused by packet loss or interference at certain nodes. The beneficial effect of this step is that it exposes the relatively weak communication links in local wireless sensor networks in industrial environments, allowing the evaluation to directly target timing differences that constrain response capabilities, and providing objective data for measuring the true collaborative state of the network.

[0073] For S32, this step aims to address the technical problem that existing evaluation models, which apply physical distance, lead to errors in electromagnetic wave communication delay calculations and fail to reflect the underlying congestion mechanisms of industrial networks. It involves obtaining the number of active nodes and the interference index of the wireless channel, determining the backoff delay using a contention window mechanism, determining the retransmission delay based on the interference index, and adding the buffer delay. The arithmetic sum of these three factors yields the standard time difference. In the application scenario, with 5 active nodes detected and high-level background interference, the backoff delay calculated using the medium access control layer algorithm is 6.5 milliseconds, the retransmission delay due to interference is 11.0 milliseconds, and the protocol's preset buffer delay is 2.5 milliseconds. The sum of these three factors gives a standard time difference of 20.0 milliseconds.

[0074] This approach, based on the underlying protocol's computational logic, is adopted because the latency caused by electromagnetic wave transmission distance is relatively low within a factory workshop. The asynchronous arrival of node data is due to backoff waiting caused by multiple sensors simultaneously vying for the wireless channel, and repeated transmissions after data packet corruption caused by industrial electromagnetic interference. By superimposing these three protocol parameters, a timing benchmark conforming to the operational patterns of industrial networks is constructed. The beneficial effect of this step is that, unlike conventional propagation models that only consider physical distance, the evaluation benchmark can reflect the constraints of the current electromagnetic environment and network load on communication, improving the rationality of the collaborative efficiency quantification system.

[0075] For S33, this step focuses on solving the technical problem of comparing the actually observed asynchronous time difference with the theoretically allowed delay benchmark, and then outputting a standardized network cooperative state index. The typical time difference is divided by the standard time difference to obtain the processing ratio, and the time difference amount is obtained by subtracting the processing ratio from the value. If this value is greater than zero, it is directly used as the initial cooperative value; if it is not greater than zero, the value is zero as the initial cooperative value. Based on the data, the typical time difference in the first subspace is 16.0 milliseconds, the standard time difference is 20.0 milliseconds, and the processing ratio is calculated to be 0.800. Subtracting 0.800 from 1 gives the time difference amount of 0.200. Since it is greater than zero, the initial cooperative value is 0.200.

[0076] This calculation logic, which involves proportional subtraction with a conditional check, converts the physical quantity of time units into a value between 0 and 1. A higher score indicates that the collaborative reporting between nodes is closer to synchronization. When severe congestion occurs in a local network, causing the actual asynchronous time to exceed the protocol's allowed baseline time, the difference is negative. In this case, a conditional check assigns a value of zero to prevent negative scores from affecting subsequent calculations. The beneficial effect of this step is that it achieves normalization of network layer communication quality, providing a quantitative parameter that reflects the data perception balance and responsiveness within this subspace, thus providing input support for subsequent data fusion.

[0077] like Figure 5 As shown, in one specific embodiment, S4 includes: S41, calculating the angle difference between the actual attitude angle and the theoretical attitude angle of the sensing node to obtain the attitude offset angle. Determining whether the attitude offset angle exceeds a preset matching threshold. If the attitude offset angle exceeds the matching threshold, performing a linear mapping operation on the attitude offset angle based on the cross-polarization discrimination rate parameter of the radio frequency signal to obtain an attitude penalty value.

[0078] S42. Based on the anchoring drift rate, attitude penalty value, and initial cooperative value of each subspace, obtain the target cooperative value for each subspace. The formula for calculating the target cooperative value is:

[0079] in, Indicates the first The target collaborative value of each subspace; This represents the minimum value function, used to output the smallest item among multiple values ​​within parentheses; This indicates the pre-set compensation coefficient; Indicates the anchorage drift rate; Indicates the attitude penalty value; Indicates the first The initial cooperative values ​​of each subspace. This operation includes... This is the correction coefficient value. Multiplying it by the initial synergy value yields the evaluation result, and the maximum upper limit of the target synergy value is ensured to be one through the minimum value function.

[0080] S43. For the multiple subspaces, allocate spatial weights according to the process criticality of their corresponding regions, and ensure that the sum of the weight values ​​of all subspaces is equal to one. The specific method for determining the spatial weights assigned to each subspace is as follows: Based on the historical archives of Failure Mode, Effects, and Criticality Analysis (FMECA) of industrial equipment in multiple past production cycles, the weights are determined through a data processing process that combines entropy weighting with expert scoring. The system comprehensively extracts historical panel data on the maintenance records of each component of the target machinery over the past 5 years, the duration of unplanned downtime caused by corresponding area failures, and direct economic losses. The severity and frequency of historical failures occurring in each physical subspace are used as the input vector for information entropy calculation. The information entropy weights of objective risks in each area are calculated, and iterative corrections are made using the initial assessment coefficients of mechanical process criticality provided by the equipment manufacturer. For example, historical data from a workshop reveals that downtime caused by failures in the main bearing area (first subspace) accounts for 62% of the total unplanned downtime, while failures in the peripheral auxiliary equipment areas (second and third subspaces) account for only 23% and 15%, respectively. After inputting this set of historical hazard percentage data into the entropy weighting model for calculation and standardization, strictly meeting the requirement that the sum of the values ​​equals one, the system ultimately pre-assigns weights of 0.60, 0.25, and 0.15 to the first, second, and third subspaces, respectively.

[0081] S44. Multiply the target coordination value of each subspace by its corresponding spatial weight and sum them to obtain the total coordination value. Compare the total coordination value with a preset safety threshold. If the total coordination value is lower than the safety threshold, generate and issue a control command. Specifically, trigger a control command to wake up redundant nodes pre-deployed in the physical area to take over the monitoring task, or trigger a control command to issue to the device through the programmable logic controller to adjust operating parameters (such as reducing the operating speed). The preset safety threshold is determined as follows: based on the extreme tolerance edge data of secondary failures caused by long-term operation of industrial equipment with defects, and combined with the decision boundary extracted after training with the Support Vector Machine (SVM) classification algorithm; collect 200 equipment overall coordination status score data recorded in the target workshop in the past 3 years, as well as binary classification labels for whether subsequent shutdown failures occurred, and input these labeled historical coordination efficiency data into the SVM model for hyperplane training. Extract the one-dimensional numerical coordinates corresponding to the best hyperplane used to distinguish between the categories of "safe operation" and "high-risk warning" after training convergence, and use it as the safety threshold; for example, after inputting 200 historical overall coordination values ​​into the model, the model identifies that the historical coordination values ​​that caused shutdown accidents are all below 0.63, while the values ​​of stable operation are all above 0.68. After extracting the classification hyperplane boundary and superimposing safety redundancy, the final output preset safety threshold is 0.65.

[0082] In this embodiment, it should be noted that step S41 aims to address the technical problem that the antenna RF polarization attenuation caused by the physical device undergoing a flipping posture cannot be quantified and fused into the communication evaluation. The attitude offset angle is obtained by calculating the difference between the actual and theoretical attitude angles of the sensing node. When the offset angle exceeds the matching threshold, a linear mapping is performed based on the cross-polarization discrimination rate to obtain the attitude penalty value. Scenario data shows that Node_1's actual yaw angle of 98.5 degrees and the theoretical 90.0 degrees result in an 8.5-degree attitude offset angle. The preset matching threshold is 5.0 degrees, and the excess is 3.5 degrees. Following a mapping rule that increases the weight by 0.05 for every 1.0-degree exceedance, the attitude penalty value is calculated to be 0.175. This calculation logic is adopted because industrial antennas have polarization characteristics, and the physical tilt of the sensor under mechanical vibration can lead to polarization mismatch between the transmitting and receiving antennas, resulting in a decrease in signal amplitude and multipath effects. By setting a threshold, slight shaking within the normal installation tolerance range is eliminated; and the linear mapping converts the physical angle parameter into a deduction coefficient for the network layer. The beneficial effect of this step is that it establishes a mathematical relationship between physical displacement and electromagnetic transmission, allowing the antenna performance degradation caused by mechanical loosening to participate in the evaluation of the overall network cooperative status in numerical form, thus enriching the data dimensions of the evaluation.

[0083] The specific process for determining the matching threshold and attitude penalty weight mapping coefficient is as follows: The threshold is obtained through a limited number of RF polarization attenuation simulation tests conducted in a laboratory environment, combined with polynomial fitting. In an anechoic chamber, the sensor antenna is controlled to undergo gradient deflection from 1 degree to 15 degrees, and the percentage decrease in signal received strength and the corresponding increase in data packet retransmission rate are recorded simultaneously for each deflection. The deflection angle corresponding to the retransmission rate mutation point is selected as the matching threshold, and a first-order linear regression is performed on the retransmission rate curve after exceeding the limit. The slope of this regression line is extracted as the penalty weight mapping coefficient. For example, tests show that the retransmission rate begins to surge at a deflection of 5 degrees, so the matching threshold is set to 5.0 degrees. Linear fitting is performed on historical attenuation data from 5 to 15 degrees, revealing that for every 1 degree increase in angle, the signal communication success rate is equivalently reduced by approximately 5%. Therefore, the penalty weight mapping relationship is set to increase by 0.05 for every 1.0 degree exceeding the limit.

[0084] Regarding S42, this step aims to address the technical problem in existing technologies where the mechanical anchoring state of the physical layer and the communication coordination performance of the network layer are independent, making it impossible to obtain a comprehensive evaluation. The compensation coefficient is multiplied by the anchoring drift rate, one is added, and the attitude penalty value is subtracted to obtain the correction coefficient value. This correction coefficient is then multiplied by the initial coordination value to obtain the evaluation result. The minimum value between this result and the initial value is extracted as the target coordination value.

[0085] Specifically, regarding the calculation expression in S42 used to obtain the target collaborative value... The core design of this computational process aims to address the technical problem in existing IoT evaluation systems where the physical layer's mechanical state and network layer's communication quality are disconnected, failing to objectively reflect the multi-physics coupling mechanism of antenna deflection caused by equipment vibration, leading to communication congestion. The preset compensation coefficient is determined based on the Pearson correlation coefficient between mechanical physical stability and network communication recovery rate in the historical operation and maintenance database of the Industrial IoT. The system extracts historical event archives of 100 instances where communication quality deteriorated due to local electromagnetic interference, statistically analyzes the time taken for the network layer to successfully restore communication through protocol retransmission under different mechanical anchoring drift rates, and uses data mining tools to calculate the statistical contribution of the physical anchoring state to the final communication success probability. The empirical expected value of this positive contribution is set as the compensation coefficient. For example, through correlation analysis of 100 sets of historical event data, it was found that when the underlying mechanical fasteners are not loose (i.e., the anchoring drift rate is high), there is a positive correlation of approximately 50% improvement in network anti-interference and self-healing capabilities. Therefore, the system directly quantifies and sets the compensation coefficient α to 0.5.

[0086] In complex industrial monitoring environments, the data packet reporting delay (reflected in the initial coordination value) is a significant factor. The issue is often not simply due to bandwidth limitations of the network channel itself, but largely due to physical displacement of the sensor hardware caused by mechanical vibration (reflected in the anchoring drift rate). And the accompanying spatial attitude flip causes antenna pointing deviation (reflected in attitude penalty value). This is caused by the deflection of the antenna attitude, which changes the polarization direction of the radio frequency signal, causing cross-polarization attenuation and exacerbating multipath effects. This forces the medium access control layer to increase the number of data retransmissions, ultimately resulting in delay tailing at the network layer.

[0087] To mathematically accurately represent this cross-domain coupling causal relationship, the expression constructs a dynamic correction coefficient value dominated by physical domain parameters. This is then used as a multiplier to apply to the initial cooperative value of the network domain. Within the internal structure of the correction coefficients, constants... This represents the baseline neutral state when no physical domain influence is applied. This constitutes a positive compensation mechanism, in which As a compensation coefficient that can be preset manually or adaptively adjusted according to the process environment, it is used to control the positive gain of mechanical stability on the overall score. When the anchoring drift rate is high, it is considered that a stable hardware foundation provides a reliable premise for data acquisition, and the communication score should be appropriately boosted.

[0088] However, the fact that the hardware base has not shifted does not equate to a intact communication link. If the sensor rotates in its original position, it will also cause signal attenuation. Therefore, the expression directly introduces a subtraction term after the positive compensation. As a negative penalty mechanism, the RF performance loss caused by antenna polarization mismatch is forcibly deducted from the correction coefficient. This internal combination of addition and subtraction allows the correction coefficient to comprehensively cover the combined effects of translation and rotation on communication. After multiplying the correction coefficient by the initial coordination value to obtain the fusion evaluation result, due to the existence of the positive compensation mechanism, if the initial coordination value in the region is high and the mechanical state is excellent, the product result may exceed expectations. This refers to the mathematical risk of exceeding the maximum score, which would undermine the evaluation system. The principle of normalized comparison within an interval. Therefore, a minimum value function is nested within the outer layer of the expression. Its function is to forcibly smooth out the peak when the comprehensive score overflows. This ensures that the mathematical dimensions of the collaborative values ​​of objectives in each subspace are rigorous and consistent.

[0089] The compensation coefficient was set based on the application scenario data of the first subspace. for The anchoring drift rate calculated beforehand for Attitude penalty value due to yaw angle exceeding the limit for The initial coordination value calculated based on the underlying protocol latency Only Substituting into the expression for calculation, the internal correction coefficients are first solved: the positive compensation part is... Then the sum of the correction factors was calculated as follows: This value is greater than This objectively demonstrates that the current mechanically stable base and controllable attitude deflection of this subspace generally provide for communication. The physical domain trust gain. Then, this gain coefficient... Multiply by the initial collaborative value A comprehensive evaluation result was obtained. Finally, the minimum value function is used. The comparison and judgment result in the output. As the target collaborative value for the first subspace, this calculation process clearly demonstrates that although the hardware maintains a certain stable gain at the physical level, the final collaborative efficiency after cross-domain fusion remains at a low level due to the severe asynchronous congestion caused by the underlying network protocol in this region. This expression, through the multiplication and adjustment of multiple parameters, successfully quantifies the cascading impact of the "mechanical-electromagnetic-communication" chain in the industrial sensing network, providing comprehensive and reliable fusion data support for management to accurately identify the root causes of faults and trigger targeted closed-loop control commands.

[0090] For S43, this step addresses the technical problem of insufficient response to core faults during overall assessment due to varying levels of process criticality in different areas of the industrial production physical environment. For the divided subspaces, spatial weights are assigned based on the process criticality of their corresponding areas, ensuring that the sum of the weights of all subspaces equals one. In the scenario of a large rotating electric motor, the target coordination values ​​for the second and third subspaces are 0.8500 and 0.9200, respectively. Considering that the operating state of the first subspace (main bearing) directly affects the motor's operation, it is assigned a higher spatial weight of 0.60; while the second and third subspaces, used for auxiliary monitoring or external environment monitoring, are assigned weights of 0.25 and 0.15, respectively.

[0091] This non-uniform weighting mechanism is adopted because, in industrial applications, sensor anomalies of equal severity occurring in peripheral areas may only result in missing auxiliary data, but if they occur in core power areas, they could trigger equipment shutdowns. If all areas are assigned equal weight, the overall assessment result will be averaged by a large amount of peripheral safety data, thus weakening the reflection of potential hazards in the core area. The beneficial effect of this step is that spatial weighting reflects the management needs of the industrial site, enabling subsequent aggregate assessments to prioritize the network coordination status of core process areas, thereby improving the effectiveness of safety monitoring.

[0092] Regarding S44, this step ultimately solves the technical problem of IoT management's inability to effectively intervene after obtaining quantitative evaluation indicators, and the lack of automated closed-loop capability to transform data flow into control commands. The target coordination value of each subspace is multiplied by its corresponding spatial weight and summed to obtain the total coordination value. When this value is lower than a preset safety threshold, a control command is triggered. Combining complete data, 0.2258 multiplied by 0.60, 0.8500 multiplied by 0.25, and 0.9200 multiplied by 0.15 are added to obtain a total coordination value of 0.48598. Given that this value is lower than the set industrial safety operating threshold of 0.65000, an automatic control command is triggered: the motor's operating speed is immediately reduced from 3000.0 RPM to 1500.0 RPM via the programmable logic controller, while simultaneously waking up the three redundant sensing nodes in this high-frequency vibration area to take over the monitoring task.

[0093] This computational logic, based on aggregated threshold triggering commands, is adopted because it's necessary to assess the overall sensing and assurance capabilities of the industrial equipment. When the core area's efficiency decreases, impacting the overall coordination value, the machine's operating parameters are proactively adjusted at the physical level to improve the communication environment, and node backups are initiated at the network level. The beneficial effect of this step is that it establishes a seamless control link from state perception and quantitative assessment to physical execution, realizing the transformation of the Industrial Internet of Things (IIoT) from data display to equipment state intervention, thus ensuring the operational order of machinery under complex working conditions.

[0094] like Figure 6 As shown, a sensor cluster management system based on the Industrial Internet of Things (IIoT) is also provided. The system includes a management platform, a sensor network platform, and an object platform that are connected in sequence. The management platform includes: The acquisition module is used to acquire the physical space, divide the physical space into multiple subspaces, acquire multiple sensing nodes in each subspace, acquire the theoretical coordinates and theoretical attitude angles of each sensing node, and acquire the actual coordinates and actual attitude angles of each sensing node. The first processing module is used to obtain the sensitivity matrix, obtain the unit offset value based on the theoretical coordinates, actual coordinates and the sensitivity matrix; obtain the group offset value based on the unit offset value of all sensing nodes; and obtain the anchoring drift rate based on the group offset value and the standard offset value. The second processing module is used to obtain the reporting time point of each sensing node, obtain the typical time difference of each subspace based on the reporting time point; obtain the backoff delay and retransmission delay, obtain the standard time difference based on the backoff delay and retransmission delay; and obtain the initial coordination value of each subspace based on the typical time difference and the standard time difference. The third processing module is used to obtain the attitude penalty value based on the theoretical attitude angle and the actual attitude angle; to obtain the target coordination value of each subspace based on the anchoring drift rate, the attitude penalty value and the initial coordination value of each subspace; to obtain the total coordination value based on the target coordination value and the spatial weight of each subspace; and to trigger control commands based on the total coordination value.

[0095] An electronic device is also provided, comprising a memory and a processor. The memory stores a computer program containing instructions corresponding to the various processing logics and model calculations described in the above embodiments. The processor is connected to the memory and is used to read and execute the computer program stored in the memory to implement the aforementioned sensor cluster management method based on the Industrial Internet of Things.

[0096] To enable those skilled in the art to fully understand and implement the technical solutions described in this specification, the following section, in conjunction with a specific application scenario, provides a detailed deduction and data analysis of the entire implementation principle of a sensor cluster management system based on the Industrial Internet of Things.

[0097] In the rotating motor operation workshop of a large heavy machinery manufacturing plant, a sensor cluster management system based on the Industrial Internet of Things (IIoT) begins the initialization and data acquisition of S1. The system uses the southwest corner of the fixed base of main bearing No. 2 as its origin. A three-dimensional spatial coordinate system was established, and the entire motor monitoring area was evenly divided into three subspaces of equal volume. Taking the first subspace, located in the core high-frequency vibration zone, as an example, a total of five sensing nodes for monitoring equipment status were deployed in this area. The system retrieved the theoretical coordinates of sensing node numbered Node_1 from the equipment ledger database. millimeters and theoretical attitude angles (where the yaw angle is 1 / 2) When the equipment is operating at full load, continuous high-frequency mechanical vibration causes physical slippage of the fasteners. The positioning and attitude calculation module captures the actual coordinates of Node_1 in real time and changes them to... Millimeters, while the yaw angle in its actual attitude angle shifts to .

[0098] The system then proceeds to S2 to quantify the mechanical displacement characteristics. First, a sensitivity matrix customized for this type of rotary motor is obtained. Since radial vibration has a significant impact on equipment lifespan, the system assigns anisotropic axis weight values: X-axis weight... Y-axis weight Z-axis weight Based on the actual and theoretical coordinates of Node_1, the coordinate difference is calculated. millimeters millimeters Millimeters. Substitute into the formula for calculating the anisotropic unit offset value. Get the unit offset value of Node_1 Millimeters. The system synchronously collects and calculates the unit offset values ​​of the remaining four nodes within the first subspace as follows: millimeters millimeters millimeters The population offset value is obtained by summing the values ​​in millimeters and then dividing by the total number of nodes (5). Millimeters. Sets the standard offset value for safe operation of the equipment. millimeters, substituting into the anchoring drift rate formula Calculation This value accurately represents whether the physical structure of the current hardware device has become loose, but the system still determines that it retains its functionality. The mechanical anchoring stability coefficient.

[0099] After completing the quantization of the spatial physical domain, the system executes S3 to evaluate the communication timing coordination of the network layer. For the first subspace, the system extracts the first data reporting timestamps of the five sensing nodes within the current monitoring period. The reporting timestamps of these five nodes are as follows: millisecond, millisecond, millisecond, millisecond, Milliseconds. The system iterates through all node combinations and calculates the absolute value of the difference between any two time points, where the 3rd node ( (milliseconds) and the first node ( The largest time difference occurred between milliseconds. Therefore, the system extracts the absolute value of this largest difference. As a typical time difference in the first subspace, milliseconds directly reveal that there are obvious timing trailing and asynchronous phenomena in the wireless communication links within this local area.

[0100] To establish a baseline for the aforementioned asynchronous phenomenon, the system continues to calculate the standard time difference in S3. The system detects that the number of currently active nodes in the first subspace is 5, and the background interference index of the current operating frequency band is high. Based on the contention window mechanism of the Medium Access Control (MAC) layer, the number of active nodes (5) is substituted into the backoff delay calculation formula. Combined with the preset contention window coefficient and the basic time slot length, the backoff delay of these 5 nodes in the current channel state is calculated to be 6.5 milliseconds. At the same time, due to the increased retransmission probability caused by electromagnetic interference, the current interference index is obtained and substituted into the retransmission delay calculation formula, resulting in a retransmission delay of 11.0 milliseconds. The system also extracts the buffer delay preset at the bottom layer of the communication protocol. Milliseconds. Adding these three values ​​arithmetically, we obtain the standard time difference of this subspace. Milliseconds. Subsequently, the system calculates the initial coordination value: dividing the typical time difference by the standard time difference yields the proportion to be processed. Next, the value Subtracting this ratio yields the time difference value. Because the time difference is greater than The system directly uses it as the initial cooperative value of the first subspace. This low value indicates that current communication synchronization has deteriorated and is approaching the limit of the protocol's tolerance.

[0101] Subsequently, the system enters S4 for bidirectional coupling correction of physical attitude and network communication. First, the attitude penalty caused by antenna polarization attenuation is calculated, given that the actual yaw angle of Node_1 is... The theoretical yaw angle is The angle difference between the two produces the attitude offset angle. The system's preset matching threshold is... Because the offset angle exceeds the limit, the system performs a linear mapping based on the radio frequency cross-polarization discrimination rate, setting a limit for each deviation exceeding the limit. Increase The penalty weights are used to calculate the attitude penalty value. Next, the system extracts the preset compensation coefficients. Substitute into the target collaborative value calculation formula Substitute the actual calculation data: correction coefficient value Multiplying the correction factor value by the initial synergy value yields the evaluation result. By using a minimum value function, the target cooperative value of the first subspace is finally determined. This calculation result shows that although the mechanical anchoring state provides Despite a gain of several times, the overall sensing performance in this local area remains extremely low due to the offsetting of antenna flipping losses and severe trailing at the underlying communication layer.

[0102] After completing the calculations for all local regions, the system continues with the global aggregation and closed-loop intervention in S4. The system simultaneously acquires the target cooperative values ​​for the second and third subspaces, which have a better external environment and less vibration. and Based on the criticality of the mechanical process, the system assigns the highest weight to the first subspace located in the core area of ​​the main bearing. Weights were assigned to the second and third subspaces respectively. and The system multiplies the collaborative values ​​of each objective by their spatial weights and sums the results to calculate the total collaborative value. The system's preset industrial safety operation threshold is: Due to the current total collaborative value Far below the safety threshold, the system immediately triggers a closed-loop control command: within milliseconds, the programmable logic controller receives a deceleration intervention signal, forcibly reducing the current speed of the rotating motor. The RPM operating speed was reduced to RPM is used to curb the further deterioration of mechanical vibration, while the system automatically wakes up the three backup redundant nodes deployed in the area to take over the sensing tasks, thus completing a complete industrial-grade intelligent management closed loop.

[0103] It should also be noted that the entire sensor cluster management system based on the Industrial Internet of Things can be applied to the optimized Industrial Internet of Things. Figure 6 This is a schematic diagram of the composition of the sensor cluster management system based on the Industrial Internet of Things of the present invention. Figure 7 This is a schematic diagram illustrating the optimized industrial Internet of Things (IIoT) involved in this invention. (See diagram below.) Figure 6 and Figure 7As shown, the optimized Industrial Internet of Things (IIoT) includes a user platform, a service platform, a management platform, a sensor network platform, and an object platform that establish communication in sequence. The user platform is configured to provide front-end services to users; users obtain the necessary perception service information through the user platform, process the perception service information, and transform it into user perception information; users analyze the user perception information and make corresponding decisions based on their own wishes, and transform the user perception information into user control information through the corresponding information system and send it to the service platform, thereby demonstrating the user's corresponding service needs and wishes.

[0104] The physical entities of the user platform include various user terminals, such as mobile phones, computers, and dedicated terminals, which provide user services through integration with user information system software.

[0105] The service platform is configured as an API server or other server used to establish communication between the management platform and the user platform to achieve corresponding functions; the physical entity of the service platform includes various servers.

[0106] The management platform is configured to perform at least one of the following: device operation status monitoring and management, data monitoring and management, device parameter management, and lifecycle management; the management platform is the overall operation platform for the Internet of Things, which may include various management sub-platforms, with different management sub-platforms performing different management tasks; the physical entities of the management platform include various servers.

[0107] The sensor network platform is configured to perform at least one of the following functions: network management, command management, device status management, data protocol management, data parsing, data classification, data transmission monitoring, and data transmission security management. The sensor network platform provides functions such as data communication, transmission, parsing, identification, and classification, avoiding the direct aggregation of data from various object platforms on the management platform, which would result in data redundancy and low data processing efficiency. The physical entities of the sensor network platform include various gateways and edge computing devices.

[0108] The object platform is configured to perform specific production control, detection, measurement and other production tasks; the physical entities of the production objects include various production equipment, sensors and so on.

[0109] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0110] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0111] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

[0112] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A sensor cluster management method based on the Industrial Internet of Things, characterized in that, The methods include: Obtain the physical space and divide the physical space into multiple subspaces; Acquire multiple sensing nodes within each subspace; Obtain the theoretical coordinates and theoretical attitude angles of each sensing node, and obtain the actual coordinates and actual attitude angles of each sensing node; Obtaining the sensitivity matrix, and obtaining the unit offset value based on the theoretical coordinates, the actual coordinates, and the sensitivity matrix, includes: obtaining the axis weight values ​​in the sensitivity matrix; obtaining the coordinate difference based on the actual coordinates and the theoretical coordinates; multiplying the square of each coordinate difference with its corresponding axis weight value and summing the results, and then taking the square root of the summation result to obtain the unit offset value. The process involves obtaining a group offset value based on the unit offset value of all sensing nodes, and obtaining an anchoring drift rate based on the group offset value and a standard offset value, including: dividing the group offset value by the standard offset value to obtain an offset ratio; subtracting the offset ratio from the value to obtain a difference result; and extracting the maximum value between the difference result and the value of zero as the anchoring drift rate. Obtain the reporting time point of each sensing node, and obtain the typical time difference of each subspace based on the reporting time point, including: calculating the absolute value of the difference between the reporting time points of any two sensing nodes in the same subspace; selecting the absolute value of the largest difference as the typical time difference of the subspace. The backoff delay and retransmission delay are obtained, and a standard time difference is obtained based on the backoff delay and retransmission delay, including: obtaining the number of active nodes and the interference index; determining the backoff delay based on the number of active nodes and the contention window mechanism; determining the retransmission delay based on the interference index; obtaining the buffer delay, and adding the backoff delay, retransmission delay and buffer delay to obtain the standard time difference; Based on the typical time difference and the standard time difference, the initial coordination value of each subspace is obtained, including: dividing the typical time difference by the standard time difference to obtain the proportion to be processed; subtracting the proportion to be processed from the value to obtain the time difference amount; if the time difference amount is greater than zero, the time difference amount is used as the initial coordination value; if the time difference amount is not greater than zero, the value of zero is used as the initial coordination value. The attitude penalty value is obtained based on the theoretical attitude angle and the actual attitude angle; Based on the anchoring drift rate, the attitude penalty value, and the initial cooperative value of each subspace, the target cooperative value of each subspace is obtained; The total collaboration value is obtained based on the target collaboration value and spatial weight of each subspace, and control commands are triggered based on the total collaboration value.

2. The sensor cluster management method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of obtaining the attitude penalty value based on the theoretical attitude angle and the actual attitude angle includes: The attitude deviation angle is obtained by calculating the angle difference between the actual attitude angle and the theoretical attitude angle. When the attitude offset angle exceeds the matching threshold, the attitude offset angle is linearly mapped to obtain the attitude penalty value.

3. The sensor cluster management method based on the Industrial Internet of Things according to claim 1, characterized in that, The step of obtaining the target cooperative value of each subspace based on the anchoring drift rate, the attitude penalty value, and the initial cooperative value of each subspace includes: Multiply the compensation coefficient by the anchoring drift rate, add a value of one, and subtract the attitude penalty value to obtain the correction coefficient value; Multiply the correction coefficient value by the initial collaborative value to obtain the evaluation result; Extract the minimum value between the evaluation result and value one, and use it as the target synergy value.

4. A sensor cluster management system based on the Industrial Internet of Things, characterized in that, The system is used to implement the sensor cluster management method based on the Industrial Internet of Things as described in any one of claims 1 to 3. The system includes a management platform, a sensor network platform, and an object platform that are sequentially connected in communication. The management platform includes: The acquisition module is used to acquire the physical space, divide the physical space into multiple subspaces, acquire multiple sensing nodes in each subspace, acquire the theoretical coordinates and theoretical attitude angles of each sensing node, and acquire the actual coordinates and actual attitude angles of each sensing node. The first processing module is used to obtain a sensitivity matrix, obtain a unit offset value based on the theoretical coordinates, the actual coordinates and the sensitivity matrix; obtain a group offset value based on the unit offset values ​​of all sensing nodes; and obtain an anchoring drift rate based on the group offset value and the standard offset value. The second processing module is used to obtain the reporting time point of each sensing node, obtain the typical time difference of each subspace based on the reporting time point; obtain the backoff delay and retransmission delay, obtain the standard time difference based on the backoff delay and retransmission delay; and obtain the initial coordination value of each subspace based on the typical time difference and the standard time difference. The third processing module is used to obtain an attitude penalty value based on the theoretical attitude angle and the actual attitude angle; obtain a target coordination value for each subspace based on the anchoring drift rate, the attitude penalty value and the initial coordination value of each subspace; obtain a total coordination value based on the target coordination value and spatial weight of each subspace; and trigger a control command based on the total coordination value.

5. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program in the memory to implement the sensor cluster management method based on the Industrial Internet of Things as described in any one of claims 1 to 3.

Images