Edge cloud collaborative software defined industrial control system

Abstract

This invention relates to the field of industrial network control technology, specifically to a software-defined industrial control system with edge-cloud collaboration. The system includes an edge control action configuration module, a multi-source signal aggregation module, a state change identification module, a cloud scheduling module, and a collaborative control generation module. In this invention, by decoupling and standardizing the control action configuration, multi-channel signal acquisition, data tag construction, change state identification, and remote scheduling execution processes, a basic edge control instruction structure can be quickly constructed based on equipment status, and a unified time-series data tagging mechanism can be established. Combined with the dynamic perception capability of change states and the feedback point identification method, efficient identification and rapid classification of key control points can be achieved. This allows task chain execution to be quickly mapped to the critical path and real-time control to be implemented, enabling rapid response to multi-source changes in the industrial field and accurate issuance of control commands, significantly improving the real-time performance and collaboration of the control process.

Description

Technical Field

[0001] This invention relates to the field of industrial network control technology, and in particular to a software-defined industrial control system with edge-cloud collaboration. Background Technology

[0002] The field of industrial network control technology mainly involves the construction and management of networked control systems for industrial environments. It encompasses data communication, command transmission, and collaborative scheduling mechanisms between various control devices in industrial automation systems. Core aspects include industrial Ethernet, real-time bus protocols, network topology, latency-tolerant mechanisms, network security control, and remote device access management. By building stable, reliable, and efficient communication networks, it achieves precise control and remote management of industrial processes, serving as a fundamental system pillar supporting the development of industrial automation, informatization, and intelligence. Software-defined industrial control systems (SDS) refer to industrial automation systems that configure control logic and function allocation through pre-set software. They primarily address the difficulties in system upgrades, insufficient flexibility, and long deployment cycles caused by the high coupling between hardware control units and control logic in industrial control systems. Typically, fixed program instructions are pre-set in the industrial controller, and functions are deployed based on specific protocols. The control flow relies on embedded programming and a centralized architecture, with functions defined and logic arranged using standard PLC programming languages. However, updates to all control nodes require adjustments to software instructions and execution logic one by one based on on-site maintenance, lacking flexible response capabilities to control requirements in dynamic industrial scenarios.

[0003] Existing industrial control systems generally rely on a centralized architecture for configuring and deploying control logic. The high coupling between control units and functional instructions in these systems makes it difficult to achieve unified scheduling and rapid adjustment of control nodes. The deployment process heavily depends on embedded programs and on-site maintenance methods, making it impossible to achieve flexible arrangement and remote response of nodes. In the face of changing operating conditions, they lack the ability to proactively identify and control them. Updates to control instructions require intervention in field devices one by one, which is cumbersome and susceptible to human intervention. In addition, there is a lack of a unified data timing structure to support multi-source data acquisition and monitoring of sudden states, making it impossible to build cross-node linkage response logic. This limits the ability to perceive and efficiently handle sudden changes, ultimately affecting the response efficiency and control flexibility of the entire industrial process. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a software-defined industrial control system that integrates edge and cloud technologies.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a software-defined industrial control system with edge-cloud collaboration includes:

[0006] The edge control action configuration module obtains the control requirement type during the self-test phase of industrial equipment startup, extracts the corresponding action configuration item, matches the basic instruction set template with the control action number, writes the matched template instruction into the edge node and starts the cache, and generates the edge control basic instruction structure.

[0007] The multi-source signal aggregation module collects multi-channel analog signals based on the signal acquisition configuration in the border control basic instruction structure, constructs a timestamp tag sequence by combining channel identifiers and channel values, classifies and writes it into the synchronization buffer according to node numbers, and generates a feedback tag time matrix.

[0008] The state change identification module calculates the difference between two adjacent acquisition cycles based on the current, temperature, and frequency tag values ​​of each node in the feedback tag time matrix, calculates the current fluctuation difference, temperature slope difference, and frequency jump amplitude, and determines whether they exceed the current change judgment coefficient, temperature slope limit, and frequency jump critical value, respectively. Feedback points that meet the conditions are marked and recorded, and a joint list of change triggers is generated.

[0009] The cloud scheduling module filters priority control template structures with corresponding numbers based on the mutation trigger joint list. Templates with scheduling cycles less than the set benchmark value are assigned to the priority distribution channel, while the rest are allocated to the standard scheduling pool and remotely transmitted according to their numbers to generate a module scheduling loading sequence.

[0010] The collaborative control generation module, based on the list of to-be-activated items in the module scheduling loading sequence, combined with the current mutation type marker of the node, parses the execution path graph in the template, maps the critical path to the edge control task execution channel, executes the task chain in real time, and generates the edge-cloud collaborative control scheduling result.

[0011] As a further embodiment of the present invention, the action configuration items include, but are not limited to, cooling fan start-up, motor no-load maintenance, and sensor calibration;

[0012] The multi-channel analog signals include, but are not limited to, servo system output frequency, spindle temperature, and power supply circuit current.

[0013] The task chain includes, but is not limited to, turning off the heat source, reducing the voltage, and lowering the speed.

[0014] As a further embodiment of the present invention, the border control basic instruction structure includes an action execution identifier, a cache trigger flag, and node mapping information; the feedback tag time matrix includes a tag timestamp, a channel data structure, and a node classification index; the mutation trigger joint list includes a trigger event record, a mutation feature parameter set, and node behavior tags; the module scheduling loading sequence includes a scheduling priority classification, a remote scheduling identifier, and an edge activation flag; and the edge-cloud collaborative control scheduling result includes a critical path mapping result, a task execution status set, and a border control response log.

[0015] As a further aspect of the present invention, the edge control action configuration module includes:

[0016] The control demand extraction submodule acquires the real-time control signal data stream during the self-test phase of industrial equipment startup. Based on the signal identifier and corresponding startup status carried by each type of signal, it extracts the control demand signals, classifies them by type, and assigns them numbers. Combining the numbering information, it extracts the corresponding control action numbers and generates a set of control action numbers.

[0017] The basic template matching submodule compares and matches the control number data in the control action number set with the basic instruction set template. For each number information, it performs template instruction indexing and instruction field mapping, arranges the successfully matched basic template instruction sequences in order, and establishes the correspondence between the number and the template field according to the logical order of the control numbers to generate a list of matched basic template instructions.

[0018] The instruction cache generation submodule constructs the instruction structure field format and cache initialization field set based on all template instruction content in the matching basic template instruction list, writes all structure instruction data into the edge node cache module, starts the structure data loading process and monitors the cache writing status, and obtains the edge control basic instruction structure set.

[0019] As a further aspect of the present invention, the multi-source signal aggregation module includes:

[0020] The signal acquisition execution submodule obtains the signal acquisition configuration in the border control basic instruction structure, parses the signal type number and channel acquisition parameters, continuously receives analog signals from each channel within a preset sampling period, parses the sampled values ​​and performs normalization conversion, converts the input analog signal values ​​to the target data range, and adds a sampling time point index at each sampling point to generate a multi-channel sampling data sequence.

[0021] The channel label construction submodule clusters and sorts the data sequences based on the time information corresponding to the multi-channel sampled data sequences and the channel number. It uses the sampling time information as the main index of the sequence, establishes a time sequence label for each group of sampling points, calculates the time interval to remove data points outside the time error range, and retains the valid sequence points after filtering to construct a labeled data set and generate a channel timestamp label sequence set.

[0022] The synchronous cache writing submodule writes the corresponding tag data into the synchronous cache space according to the data structure of the channel timestamp tag sequence set, using the node number as the index. It calculates the storage location offset through the mapping relationship between the node and the cache address, encapsulates the tag sequence data, performs the cache writing operation, and obtains the feedback tag time matrix.

[0023] As a further aspect of the present invention, the state change identification module includes:

[0024] The signal difference calculation submodule obtains the current, temperature and frequency signal values ​​of each node in the feedback tag time matrix, performs item-by-item difference calculation on the same type of signal in adjacent sampling periods, calculates the composite sudden change fluctuation intensity index, and generates a list of signal sudden change fluctuation values.

[0025] The mutation condition determination submodule compares the normalized fluctuation value of the signal corresponding to each node in the signal mutation fluctuation value list with the preset current mutation determination coefficient, temperature slope limit value and frequency jump critical value one by one. It adopts the greater than threshold determination method, marks the fluctuation value that meets any critical condition as an abnormal mutation item, records the node where the mutation signal is located, the signal type and the judgment time, and obtains the mutation signal judgment result set.

[0026] The anomaly labeling output submodule determines the abnormal mutation items in the mutation signal judgment result set, corresponds to the node number, signal type and judgment time in advance, and uniformly constructs an anomaly item identification structure for all records that meet the mutation conditions. It uses a structure list to summarize them in chronological order and uniformly writes them into the node anomaly log cache area to establish a joint list of mutation triggers.

[0027] As a further aspect of the present invention, the calculation formula for the composite mutation fluctuation intensity index is as follows:

[0028] ;

[0029] in, Indicates the first Signal type at time The normalized abrupt change fluctuation intensity, and They represent the first The label values ​​of the signal type at the current and previous time steps. The standard deviation of this type of signal. Used to measure the combined energy level of disturbances in the three signal channels.

[0030] As a further aspect of the present invention, the cloud scheduling module includes:

[0031] The template filtering and allocation submodule obtains all node numbers and corresponding mutation types in the variable trigger joint list, indexes the corresponding control template structure in the cloud template mapping table according to the node number, extracts the scheduling cycle value set in the template and judges it with the benchmark cycle value, extracts template items with scheduling cycles less than the benchmark value and classifies them into the priority distribution channel, and the remaining template items are classified into the standard scheduling pool. The template distribution priority list is generated according to the number order of the filtering operation.

[0032] The instruction remote transmission submodule constructs a remote scheduling instruction frame structure and binds the target node number according to the control template number in the template distribution priority list. It then transmits the instruction structure to the corresponding edge node in batches through the cloud-edge communication channel. After the transmission is completed, it records the corresponding number and the issuance time index and establishes a remote template issuance record table.

[0033] The edge registration and loading submodule locates the scheduling area corresponding to the current edge node based on the template number and node number information in the remote template distribution record table, registers the template number, the channel to which it belongs and the activation identifier in the cache area to be activated, and generates a control index sequence structure in chronological order after completing the loading and registration action, and establishes the module scheduling and loading sequence.

[0034] As a further aspect of the present invention, the collaborative control generation module includes:

[0035] The execution path parsing submodule extracts the corresponding template number and associated node number based on the content of the list to be activated in the module scheduling loading sequence. Combined with the mutation type marker information corresponding to the node in the mutation trigger joint list, it extracts the path graph index that matches the current mutation type, sequentially traverses all path nodes, filters key paths, and obtains the key path node sequence set.

[0036] The task channel mapping submodule performs task-resource matching judgment on the task type corresponding to each node based on the node number in the critical path node sequence set, establishes the mapping relationship between critical tasks and specific border control channels, completes the construction of the three-dimensional assignment structure of node-channel-task, and outputs the execution priority index and physical mapping location index of each task to generate a task channel mapping matrix.

[0037] The scheduling action execution submodule issues execution control commands to each channel according to the channel address and task sequence index in the task channel mapping matrix, and schedules them to the edge control interface in a sequential loading manner. For each scheduling, it records the current timestamp, execution channel number and return status code information, integrates all successful scheduling records and organizes them into a structure based on nodes, and establishes the edge-cloud collaborative control scheduling result.

[0038] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0039] In this invention, by decoupling and standardizing the control action configuration, multi-channel signal acquisition, data tag construction, abrupt state identification, and remote scheduling execution process, a basic edge control instruction structure can be quickly constructed based on equipment status, and a unified time-series data tagging mechanism can be established. Combined with the dynamic perception capability of abrupt states and the feedback point identification method, efficient identification and rapid classification of key control points can be achieved. With the help of remote scheduling strategies, differentiated template distribution is completed and a list of control points to be activated is formed, enabling task chain execution to be quickly mapped to the critical path and implemented in real time. Ultimately, this achieves rapid response to multi-source changes in the industrial field and accurate issuance of control commands, significantly improving the adaptability of the control system to the dynamic environment and the real-time performance and coordination of the control process. Attached Figure Description

[0040] Figure 1 This is a system flowchart of the present invention;

[0041] Figure 2 This is a flowchart of the control action configuration module of the present invention;

[0042] Figure 3 This is a flowchart of the multi-source signal aggregation module of the present invention;

[0043] Figure 4 This is a flowchart of the state change recognition module of the present invention;

[0044] Figure 5 This is a flowchart of the cloud scheduling module of the present invention;

[0045] Figure 6 This is a flowchart of the collaborative control generation module of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0047] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0048] Please see Figure 1A software-defined industrial control system with edge-cloud collaboration includes:

[0049] The edge control action configuration module obtains the control requirement type during the self-test phase of industrial equipment startup, extracts action configuration items such as cooling fan start, motor no-load maintenance, and sensor verification, matches the basic instruction set template with the control action number, writes the matched template instruction into the edge node and starts the cache, and generates the edge control basic instruction structure.

[0050] The multi-source signal aggregation module is based on the signal acquisition configuration in the edge control basic instruction structure. It collects multi-channel analog signals such as servo system output frequency, spindle temperature, and power circuit current. It combines channel identifiers and channel values ​​to construct a timestamp tag sequence and categorizes all sequences by node number and writes them into the synchronization cache to generate a feedback tag time matrix.

[0051] The state change identification module calculates the difference between two adjacent acquisition cycles based on the label values ​​of current, temperature and frequency at each node in the feedback label time matrix. By calculating the difference in current fluctuation, temperature slope and frequency jump amplitude, it determines whether the current change judgment coefficient, temperature slope limit and frequency jump threshold are exceeded respectively. The module then marks the feedback points that meet the conditions and records the node number, change type and trigger time, and generates a joint list of change triggers.

[0052] The cloud scheduling module selects priority control template structures based on the node number and mutation type in the mutation trigger joint list. Templates with a scheduling cycle of less than 5ms are assigned to the priority distribution channel, while the rest are assigned to the standard scheduling pool. The modules are remotely transmitted according to their numbers and registered in the list to be activated at the edge, generating a module scheduling loading sequence.

[0053] The collaborative control generation module, based on the list of to-be-activated items in the module scheduling loading sequence and combined with the current mutation type marker of the node, parses the execution path graph in the template, maps the key paths in the path graph to the edge control task execution channel, and executes task chains such as shutting down the heat source, reducing the voltage, and adjusting the speed in real time, generating edge-cloud collaborative control scheduling results.

[0054] The basic border control instruction structure includes action execution identifiers, cache trigger flags, and node mapping information. The feedback tag time matrix includes tag timestamps, channel data structures, and node classification indexes. The mutation trigger joint list includes trigger event records, mutation feature parameter sets, and node behavior tags. The module scheduling loading sequence includes scheduling priority classification, remote scheduling identifiers, and edge activation flags. The edge-cloud collaborative control scheduling results include critical path mapping results, task execution status sets, and border control response logs.

[0055] Please see Figure 2 The border control action configuration module includes:

[0056] The control demand extraction submodule acquires the real-time control signal data stream during the self-test phase of industrial equipment startup. Based on the signal identifier and corresponding startup status carried by each type of signal, it extracts the control demand signals, classifies them by type, and assigns them numbers. Combining the numbering information, it extracts the corresponding control action numbers and generates a set of control action numbers.

[0057] Real-time control signal data streams during the self-test phase of industrial equipment startup are acquired. This includes data collection on the power supply status, initial motor current, sensor analog signal levels, and cooling fan speed feedback. These data are parsed using analog signal input interfaces and communication protocol frame formats, converting each signal into a standardized data structure. Signal identifiers are used, such as value 1 for cooling fan start signal, value 2 for motor no-load maintenance signal, and value 3 for sensor calibration feedback signal. Signal numbers are identified through command header fields. The raw data (5Hz frequency, 60s sampling time) is smoothed to remove spikes. Signals are then categorized according to the equipment startup sequence and determined based on equipment type and action logic table. Whether the control action triggering conditions are met, such as the motor starting current fluctuation being within ±10% and the duration being greater than 5 seconds, is determined to be a valid no-load maintenance signal. Then, the signal number is converted into a control action number through a mapping table. For example, number 1 corresponds to the cooling fan pre-start action, number 2 corresponds to the motor no-load steady-state maintenance, and number 3 corresponds to the sensor zero-point calibration. The signal action type is converted into a set of identifiable control action numbers through a number lookup table method. This mapping table is stored using a static hash table, and the key-value pair structure is <signal identifier, control action number>. For example, signal identifier 2 corresponds to the number code segment 0201. By extracting and comparing each item, all action numbers that meet the logical conditions are filtered out to form a continuous control action number sequence, as shown in Table 1.

[0058] Table 1 Mapping Table of Control Signals and Action Numbers

[0059] ;

[0060] As shown in Table 1, after the signal is acquired and the corresponding action number is obtained, the subsequent instruction template matching stage is entered to obtain the set of control action numbers.

[0061] The basic template matching submodule compares and matches the control number data in the control action number set with the basic instruction set template. For each number information, it performs template instruction indexing and instruction field mapping, arranges the successfully matched basic template instruction sequences in order, and establishes the correspondence between the number and the template field according to the logical order of the control numbers to generate a list of matched basic template instructions.

[0062] Based on all the numbered data in the control action number set, taking the current number sequence 0101, 0201, and 0301 as an example, these correspond to three types of action requirements: cooling fan start-up, motor no-load maintenance, and sensor verification feedback, respectively. The control action number set is used as the key, and each entry is compared and searched in the preset basic instruction set template table. A hash index matching is performed using the number field in the template library as the primary key. By comparing the number consistency and verifying the field integrity, the corresponding template instruction fields are selected. Instruction fields include instruction type, execution duration, execution parameters, target port number, etc. For example, the template instruction field corresponding to number 0101 is <Start-up type: PWM, Duration: 30s, Parameters: Duty cycle 70%, Port: 0x01>. After matching, the data is sorted according to the execution order. The control numbers are arranged in order as follows: cooling fan start, motor no-load maintenance, and sensor verification. Field mapping adopts field comparison + parameter type classification. The field comparison operation is based on the number, and each control field is checked item by item and a mapping dictionary is built. For example, the mapping result of number 0201 is <execution type: constant speed maintenance, target value: 500rpm, duration: 60s, control terminal: 0x03>. All field data is stored as a structure array type. The data structure has a fixed field format and is arranged in the order of control number, execution type, parameter value, duration, and target port. After matching and sorting, all fields are summarized into a matching basic template instruction list, which is then generated.

[0063] The instruction cache generation submodule constructs the instruction structure field format and cache initialization field set based on the content of all template instructions in the matching basic template instruction list, writes all structure instruction data into the edge node cache module, starts the structure data loading process and monitors the cache writing status, and obtains the edge control basic instruction structure set.

[0064] Based on the content of all template instruction fields in the matching basic template instruction list, a cache format for the instruction structure of the edge control node is constructed. The cache format adopts a fixed-length structure layout, and the field order is consistent with the template instructions. The field values ​​in the matching template are called and filled into the cache fields in sequence. The execution duration, control parameters, and target ports involved in the fields are converted into binary codes and written to the local RAM cache module mounted on the edge node. The cache structure is organized in a queue manner. After the cache controller is started, control signal frames are written to the control interface of the edge node. The cache status is checked by the start and end bits of the control frame. If the write is successful... The current cache page status is recorded as READY. For example, the instruction structure field <Execution type: calibration operation, calibration deviation: ±0.05V, calibration duration: 20s, port: 0x05> corresponding to number 0301 is encoded, written into the structure, and appended to the cache queue. The cache queue length is set to 64 frames, and the size of each frame is fixed at 128 bytes. The loading status of the current cache frame and the page number mapping table are read through the controller status register. Finally, the physical address correspondence between the structure instruction data and the edge node control buffer is established, the control data loading process is completed, and the edge control basic instruction structure set is obtained.

[0065] Please see Figure 3 The multi-source signal aggregation module includes:

[0066] The signal acquisition execution submodule obtains the signal acquisition configuration in the border control basic instruction structure, parses the signal type number and channel acquisition parameters, continuously receives analog signals from each channel within a preset sampling period, parses the sampled values ​​and performs normalization conversion, converts the input analog signal values ​​to the target data range, and adds a sampling time point index at each sampling point to generate a multi-channel sampling data sequence.

[0067] To obtain the signal acquisition configuration from the basic command structure of the edge control system, first extract the channel number, signal type number, and sampling frequency settings for the acquired signal. Bind the servo system's output frequency signal to the specified analog channel, and initiate the sampling logic for that channel via the control register. The system triggers sampling at set intervals according to the established sampling period, recording the analog input voltage value at each sampling point. This value is then standardized and converted using linear interpolation based on the channel's preset mapping parameters. For example, the input voltage signal range is mapped to a frequency range; if the acquired voltage value is 2.5, the corresponding frequency is calculated to be 50 using a conversion factor. The conversion process is expressed in the form of a linear function. The conversion coefficients are obtained by looking up a table. The processing of spindle temperature and power circuit current signals also uses the same sampling logic for synchronous operation. After each sampling, the system adds sampling time information as a sampling index and temporarily stores the sampling value and time pair through a node caching mechanism. Each set of sampling points is sorted and combined in the system cache by index number. The system completes data reception, conversion and index construction for the three channels in a single sampling period. The sampling duration is set to 10 seconds. During this period, the three types of signals are received at equal intervals. The total amount of data received by each channel is 100, 50 and 10 sets respectively, as shown in Table 2.

[0068] Table 2 Signal Sampling Settings

[0069] ;

[0070] As shown in Table 2, the sampling configuration of different signals directly affects their data accuracy and response density. After completing the conversion and time matching, a multi-channel sampling data sequence is obtained.

[0071] The channel label construction submodule clusters and sorts the data sequences based on the time information corresponding to the multi-channel sampled data sequences and the channel number. It uses the sampling time information as the main index of the sequence, establishes a time sequence label for each group of sampling points, calculates the time interval to remove data points outside the time error range, and retains the valid sequence points after filtering to construct a labeled data set and generate a channel timestamp label sequence set.

[0072] Based on the sampling time and sampled value content in the multi-channel sampled data sequence, all sampling points are initially grouped according to the channel number information attached to each sampling point, grouping sampling data belonging to the same channel into the same category. Then, each group of data is sorted in ascending order according to the sampling time, constructing a data string with a time index structure. Subsequently, the time interval of each group of data is filtered and calculated, using the average sampling interval as a benchmark, setting an error threshold range of ±2%, and calculating the offset between two consecutive time intervals. If the time interval of a certain sampling point exceeds the deviation threshold from the previous data point, then that data is... Data points are marked as invalid and removed. For example, if the baseline sampling interval is set to 0.1 seconds and the actual interval is 0.103 seconds, the deviation rate is 3%, which exceeds the threshold. The data point is then removed. The remaining data points are labeled. The label content consists of three items: channel number, sampling time, and sampling value. Three types of label sequences are constructed for each channel. Each sequence uses channel + time to form a unique primary key index, which is used for subsequent cache mapping and structure matching operations. This index structure avoids duplicate data redundancy and address conflicts. After the labels are constructed, all label sequences are renumbered and archived to obtain the channel timestamp label sequence set.

[0073] The synchronous cache writing submodule writes the corresponding tag data into the synchronous cache space according to the data structure of the channel timestamp tag sequence set, using the node number as the index. It calculates the storage location offset through the mapping relationship between the node and the cache address, encapsulates the tag sequence data, performs the cache writing operation, and obtains the feedback tag time matrix.

[0074] Based on the data entry structure stored in the channel timestamp tag sequence, the storage location is determined according to the border control node number information corresponding to each group of data. Each node number determines its allocated data block in the synchronization cache through the node mapping table. The system queries the target cache address based on the node number, and then serializes the tag sequence to be written. The data of each field of the tag is combined into a continuous byte block in a fixed order using a structure organization method. The total byte length and check bit of each structure are calculated, and the integrity of the start identifier and the end bit is checked. After the check passes, the cache write instruction is executed. The system records the current start address, data length and cache offset, and updates the cache index table to maintain the data structure mapping relationship. If the cache address is out of bounds, the write failure status is returned and the previous record status is retained. The entire cache write process is based on the rule of one-to-one correspondence between data structure and address index to complete the encapsulation, verification and write process. After the write is completed, the system updates the data validity identifier of the current node through the synchronization interface and writes back the cache status byte to the status register, and finally obtains the feedback tag time matrix.

[0075] Please see Figure 4 The state change identification module includes:

[0076] The signal difference calculation submodule acquires the current, temperature, and frequency signal values ​​of each node in the feedback tag time matrix, and performs item-by-item difference calculation for similar signals within adjacent sampling periods, using the formula:

[0077] ;

[0078] The composite abrupt change fluctuation intensity index is obtained through calculation, and a list of signal abrupt change fluctuation values ​​is generated; among which... Indicates the first Signal type at time The normalized abrupt change fluctuation intensity, and They represent the first The label values ​​of the signal type at the current and previous time steps. The standard deviation of this type of signal. Used to measure the combined energy level of disturbances in the three signal channels;

[0079] To obtain the current, temperature, and frequency signal values ​​of each node in the feedback tag time matrix, the signal tag values ​​corresponding to the current and previous sampling periods are first extracted according to the node number. The tag values ​​are bound to the signal channels via a time-series index, and each signal belongs to an independent acquisition channel. The signal values ​​of two periods are paired according to the channel order. After signal pairing, the difference amplitude is calculated for each type of signal. The calculation method is the time-series difference formed by subtracting the previous period value from the current period value. This difference has direct quantitative significance for single-point mutations. However, to eliminate the influence of different signal dimensions, all differences are divided by the standard deviation calculated during the stable operation of each channel to normalize to a dimensionless mutation ratio. This standard deviation is estimated using the standard deviation of 100 consecutive sampled values ​​during the initial operation phase of the signal. For example, the standard deviation of the temperature signal is 1.8, the frequency is 3.2, and the current is 0.5. The sampled values ​​are current 4.5 / 3.8A, temperature 60.0 / 58.2°C, and frequency 45.0 / 40.6Hz. Then, the normalized single-signal difference is calculated as current: ,temperature: ,frequency: Then, substituting the values ​​into the formula, the composite fluctuation intensity of the current channel is calculated as follows:

[0080] ;

[0081] The same computational logic is used to process the channel data of all nodes. Finally, the channel fluctuation intensity of all nodes is organized into a standardized composite mutation index sequence according to time sequence for subsequent judgment steps, generating a list of signal mutation fluctuation values.

[0082] The calculation logic of this formula is mainly based on two parts, the first of which is... This represents the normalized abrupt change amplitude of a single signal channel within adjacent sampling periods. The difference between the current value and the previous period's value reflects the instantaneous jump of the signal in the local time domain, while the denominator contains... This represents the standard deviation of the signal channel during its stable operation period. It is used to uniformly process the magnitudes of different signals, thereby converting the fluctuations of all signals into dimensionless ratios; the second term... To jointly represent the disturbances of multiple signal channels, the normalized sum of squared differences of various signal channels is used, followed by square root operation. This is logically equivalent to calculating the Euclidean norm of the three-dimensional disturbance vector. This structure can accurately measure the overall disturbance intensity of multiple signals at the current time point. The squaring operation emphasizes the amplitude while ignoring the direction, and the square root is used to restore the scale and enhance the expression of linkage. The overall formula adds and combines the drastic changes of a single channel with the overall disturbance of the system, so that the local abnormal changes and global disturbances are superimposed, thereby improving the comprehensive response capability to multi-source sudden change signals. It has a dual design logic of linkage identification and precision control.

[0083] The composite mutation fluctuation intensity index is used to quantify the combined performance of the instantaneous mutation degree of a signal channel at a specific moment and the overall disturbance state of the system. This index integrates the normalized amplitude of the channel in two consecutive sampling periods with the composite value of the normalized disturbance of the three types of signal channels in the current period. The former reflects the mutation sensitivity of the channel itself, while the latter reflects the coordinated jump trend of the system-level signal. The index formed by the sum of the two can simultaneously reflect local drastic changes and global linkages, thereby effectively characterizing the abnormal fluctuation degree of the node at the current time point. The larger the value, the more the signal behavior of the node deviates from the normal operating range in the current period, and it has a high degree of anomaly identification indicativeness.

[0084] The mutation condition determination submodule compares the normalized fluctuation value of the signal corresponding to each node in the signal mutation fluctuation value list with the preset current mutation determination coefficient, temperature slope limit value and frequency jump critical value one by one. It adopts the greater than threshold determination method, marks the fluctuation value that meets any critical condition as an abnormal mutation item, records the node where the mutation signal is located, the signal type and the judgment time, and obtains the mutation signal judgment result set.

[0085] Based on the signal fluctuation intensity recorded in the signal mutation fluctuation value list, a mutation judgment operation is performed. First, fixed thresholds are set for each channel: the current mutation judgment coefficient is set to 2.8, the temperature slope limit is set to 2.0, and the frequency jump threshold is set to 2.5. These three thresholds are stable operating boundaries set through multiple system operation sampling and statistical analysis. When judging the fluctuation value, the fluctuation value of the current node in this period is read channel by channel and compared with the corresponding channel threshold. If the fluctuation value of a certain channel exceeds its set threshold, the node at that moment is considered to have triggered a signal mutation type event, and its channel number and signal category are recorded. If multiple channels of the same node trigger the condition, all types that meet the mutation condition are recorded separately and marked as joint mutation events. For example, if the current channel mutation value is 3.6, which is greater than the threshold 2.8, and the frequency mutation value is 2.1, which is less than the threshold 2.5, then only the current channel mutation at that moment is recorded, and the judgment result is a current mutation event. Finally, the index relationship between the node number and the signal channel number of all nodes that meet the trigger condition is calculated and written into the cache in chronological order to obtain the mutation signal judgment result set.

[0086] The anomaly labeling output submodule determines the abnormal mutation items in the result set based on the mutation signal, pre-assigns the corresponding node number, signal type, and judgment time, uniformly constructs an anomaly item identification structure for all records that meet the mutation conditions, summarizes them in chronological order using a structure list, uniformly writes them into the node anomaly log cache area, and establishes a joint list of mutation triggers.

[0087] Based on the node number, mutation type, and time index information corresponding to each mutation record in the mutation signal judgment result set, a unique identifier is first constructed for each record that meets the mutation judgment. The current mutation event that was judged as valid in the previous example is extracted, with a mutation value of 3.602, exceeding the current mutation judgment threshold of 2.8. Therefore, the node number is recorded as N001, the mutation type as current mutation, and the trigger time index as T57. This record is then assembled into an abnormal event entry, uniformly organized using a structure format. Each record contains three key fields: event node number, mutation type number, and trigger time index. A unique event identifier ID is generated for this structure entry, named in the form of "E + node number + time index + type number", such as E_N001_T57_I. The record order is arranged in ascending order of timestamps. Each structure entry has a fixed length of 64 bytes. The system constructs a list according to the timeline and writes it sequentially to the edges. Before writing, the system calculates the required cache size based on the number of events in the exception log cache area allocated by the control node. For example, if there are a total of 5 mutation events, the cache requirement is 64 bytes × 5 events = 320 bytes. The system retrieves the start and end addresses of the cache to ensure that the write space is complete and conflict-free. If the current free address of the cache area is 0x3000, the batch of exception records will be written from address 0x3000 to address 0x3120. After writing, the index block number Block_ID=20251105_01 of this record block is generated. At the same time, the storage location and write status code of the index block are registered in the global status identifier table. The system marks the batch of data as "Active", indicating that it has taken effect and can be called by subsequent modules. Finally, an array of exception event structures is constructed with the node number as the primary key, the mutation type as the index, and the trigger time as the retrieval identifier, and they are uniformly archived into the cache log structure to establish a mutation trigger joint list.

[0088] Please see Figure 5 The cloud-based scheduling module includes:

[0089] The template filtering and allocation submodule obtains all node numbers and corresponding mutation types in the mutation triggering joint list, indexes the corresponding control template structure in the cloud template mapping table according to the node number, extracts the scheduling cycle value set in the template and compares it with the baseline cycle value, extracts template items with scheduling cycles less than the baseline value and classifies them into the priority distribution channel, and the remaining template items are classified into the standard scheduling pool. The template distribution priority list is generated according to the number order of the filtering operation.

[0090] To obtain all node numbers and corresponding mutation types in the mutation triggering joint list, first parse the node number and mutation type fields of the structure in the list and store them in a mapping list as key-value pairs. This mapping result is used as the template index input. Based on each node number, locate the preset control template index table in the cloud, and extract the scheduling cycle parameter set in the corresponding template one by one. The scheduling cycle unit is milliseconds. Perform a value comparison operation between the extracted template scheduling cycles and the baseline value of 5ms. Set the judgment condition as follows: scheduling cycles less than 5ms are high-priority items, otherwise they are standard-priority items. The two types of template numbers are assigned to different distribution channels. For example, the period corresponding to number T8001 is 3.0ms and the period corresponding to number T8012 is 6.2ms. Therefore, T8001 is assigned to the priority distribution channel and T8012 is assigned to the standard scheduling pool. The template numbers within each channel are arranged in ascending order of node number and archived. To facilitate subsequent optimization of the system's scheduling strategy, the priority category, scheduling period value and channel allocation information of all archived templates are organized into a structured array. Each element in the list contains triple information, namely template number, priority label and scheduling period value. See Table 3 for an example.

[0091] Table 3 Template Distribution Priority Classification Table

[0092] ;

[0093] As shown in Table 3, after all templates have completed priority marking and channel classification, a template distribution priority list is generated.

[0094] The instruction remote transmission submodule constructs a remote scheduling instruction frame structure and binds the target node number according to the control template number in the template distribution priority list. It then transmits the instruction structure to the corresponding edge node in batches through the cloud-edge communication channel. After the transmission is completed, it records the corresponding number and the issuance time index and establishes a remote template issuance record table.

[0095] Based on the control template number and target node number in the template distribution priority list, the system first constructs a scheduling instruction frame structure. The instruction frame consists of a template number field, a target node number field, a scheduling period field, and a control check field. Each instruction frame is encoded in a fixed byte format; for example, an instruction frame with template number T8001, node number N09, and a period of 3.0ms is encoded as Frame_8001_N09_03. After construction, remote transmission scheduling is performed according to the distribution priority order. The cloud-edge communication link interface is called, and each instruction frame is written to the edge node's receive buffer via the TCP packet distribution module. The system records the transmission timestamp and target node number of each frame, storing them as transmission record index information in the remote distribution log structure. Simultaneously, network stability and frame return response codes are checked. Successful response frames are marked, while failed frames are recorded as "RETRY" and added to the retransmission queue. Each distribution operation registers the corresponding frame ID, node ID, time index, and status code in the log table. For example, instruction frame T8001 is distributed to node N09 on 2024-11-05. At 10:12:00, with the status code ACK, the system generates the following record in the log: LogID=L20241105_01. All distributed log records are indexed, sorted, and output, and a remote template distribution record table is established.

[0096] The edge registration and loading submodule locates the scheduling area corresponding to the current edge node based on the template number and node number information in the remote template distribution record table. It registers the template number, the channel it belongs to, and the activation identifier in the cache area to be activated. After completing the loading and registration action, it generates a control index sequence structure in chronological order and establishes the module scheduling and loading sequence.

[0097] Based on the template number and node number information in the remote template distribution record table, the corresponding template structure data is extracted from the buffer of the edge node receiver. The template number is mapped to the scheduling area of ​​the edge control channel. The storage location is determined by the preset correspondence table between the control channel and the template number. The system generates a list of items to be loaded in the scheduling activation buffer. The list item contains three items: template number, scheduling type label, and activation status identifier. The system automatically assigns an activation sequence number according to the template number and inserts the corresponding item into the buffer. If there are multiple template numbers to be activated on the same node, the order is adjusted according to the scheduling priority label. The priority channel template is inserted into the previous position, and the standard channel template is inserted into the next position. For example, if node N09 distributes two templates, T8001 (priority) and T8012 (standard), the loading order in the list is T8001→T8012. At the same time, a unique index value is assigned to each loading record and registered in the node scheduling control table in the form of an index structure table. After all items to be activated are registered and sorted, an edge-side control scheduling activation list is formed, and a module scheduling loading sequence is established.

[0098] Please see Figure 6 The collaborative control generation module includes:

[0099] The execution path parsing submodule extracts the corresponding template number and associated node number based on the list of to-be-activated items in the module scheduling loading sequence. Combined with the mutation type marking information of the nodes in the mutation trigger joint list, it extracts the path graph index that matches the current mutation type, sequentially traverses all path nodes, selects the path branch with the largest scheduling weight and continuous dependency coefficient greater than 0.8 in each path, marks it as a critical path, and obtains the critical path node sequence set.

[0100] The system retrieves the list of nodes to be activated from the module scheduling loading sequence. First, it performs index matching between each template number and its corresponding node number. Then, it sequentially extracts the node structure information from the loading cache and reads the mutation type marker for that node. For example, node N004 is marked as a current mutation, and node N006 is marked as a temperature mutation. Subsequently, based on the type index in the mutation trigger union list, it locates the matching execution path graph index field in the template structure. This path graph is stored as a node connection matrix, where the connection dependencies between nodes are defined by weight values. The weight values ​​are set between 0 and 1, representing the node's connection dependency. To determine inter-path dependency strength, the system iterates through and calculates the node weights for each path during parsing, recording the dependency coefficient of each node. Continuity is determined by comparing the dependency coefficients of adjacent nodes; nodes with a dependency coefficient greater than 0.8 are included in the continuous dependency path segment. Within the same path segment, the system selects the node with the highest weight value as the critical node. For example, if path segment P1 contains nodes A, B, and C with dependency coefficients of 0.83, 0.87, and 0.65 respectively, the system extracts nodes A and B to form the critical path segment. All segments are combined to form a complete critical path set, and the node number order and weight sequence are recorded in the parsing table. The table below shows the extracted data for an example path.

[0101] Table 4. Statistics on Path Node Dependency and Weight

[0102] ;

[0103] As shown in Table 4, after parsing, the system generates a sequence of path node numbers in the order of node numbers, thus obtaining a set of key path node sequences.

[0104] The task channel mapping submodule performs task-resource matching judgment on the task type corresponding to each node based on the node number in the critical path node sequence set, establishes the mapping relationship between critical tasks and specific border control channels, completes the construction of the three-dimensional assignment structure of node-channel-task, and outputs the execution priority index and physical mapping location index of each task, generating a task channel mapping matrix.

[0105] Based on the node numbers in the critical path node sequence set, the system reads the task type field information corresponding to each node and matches the task type with the channel identifier in the edge control resource table. Each channel in the resource table is marked with a channel number, channel capacity, and current task occupancy status. The system performs a matching judgment on the task type of each critical node. If the task is of the "heat source control" type, the temperature control execution channel is matched first; if the task is of the "voltage adjustment" type, the voltage control channel is matched; if the task is of the "speed regulation" type, the spindle control channel is matched. For cases where there are multiple tasks under the same node, the system calculates the task weight and sorts them according to the scheduling priority parameter. The priority parameter is divided into a range of 1 to 5 according to the urgency of the task, with the smaller the value, the higher the priority. The scheduling order is arranged in ascending order of priority. For example, if the critical path tasks of node N004 include current downgrading (priority 1) and temperature stabilization (priority 3), then the current downgrading task is first assigned to execution channel C001, and the temperature stabilization task is assigned to execution channel C005. After mapping all tasks, the system generates a three-dimensional structure matrix. Each element in the matrix consists of a node number, a task type, and a channel number, and records the execution priority index and the channel position index to obtain the task channel mapping matrix.

[0106] The scheduling action execution submodule issues execution control commands one by one according to the channel address and task sequence index in the task channel mapping matrix, and schedules them to the edge control interface in a sequential loading manner. It records the current timestamp, execution channel number and return status code information for each scheduling, integrates various successful scheduling records and organizes them in terms of nodes to establish the edge-cloud collaborative control scheduling result.

[0107] Based on the channel address and task sequence index in the task channel mapping matrix, the system traverses all task items by node number, constructing control instruction frames for each item. Each instruction frame contains four fields: task number, channel number, control action value, and time index. For example, for the current downgrading task under node N004, the system constructs instruction frame F_N004_C001_D, writes it to the target channel C001, sets the action value to downgrading level 2, and the time index to T214. All instructions are arranged from 1 to n according to the task sequence number, loaded into the edge control interface cache, and then issued in sequence. The system monitors the execution feedback of each scheduling in real time, recording the return status code. Status code 200 indicates successful execution, 400 indicates delayed execution, and 500 indicates channel blockage. After all tasks are issued, the system calculates the execution results, the task success rate, and the latency ratio. All successfully executed task records are summarized by node number to generate a table structure, which is then input into the cloud synchronization cache as the collaborative execution result, forming a structure array indexed by node number. This array is sorted in ascending order by timestamp and archived, ultimately establishing the edge-cloud collaborative control scheduling result.

[0108] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A software-defined industrial control system with edge-cloud collaboration, characterized in that, include: The edge control action configuration module obtains the control requirement type during the self-test phase of industrial equipment startup, extracts the corresponding action configuration item, matches the basic instruction set template with the control action number, writes the matched template instruction into the edge node and starts the cache, and generates the edge control basic instruction structure. The multi-source signal aggregation module collects multi-channel analog signals based on the signal acquisition configuration in the border control basic instruction structure, constructs a timestamp tag sequence by combining channel identifiers and channel values, classifies and writes it into the synchronization buffer according to node numbers, and generates a feedback tag time matrix. The state change identification module calculates the difference between two adjacent acquisition cycles based on the current, temperature, and frequency tag values ​​of each node in the feedback tag time matrix, calculates the current fluctuation difference, temperature slope difference, and frequency jump amplitude, and determines whether they exceed the current change judgment coefficient, temperature slope limit, and frequency jump critical value, respectively. Feedback points that meet the conditions are marked and recorded, and a joint list of change triggers is generated. The cloud scheduling module filters priority control template structures with corresponding numbers based on the mutation trigger joint list. Templates with scheduling cycles less than the set benchmark value are assigned to the priority distribution channel, while the rest are allocated to the standard scheduling pool and remotely transmitted according to their numbers to generate a module scheduling loading sequence. The collaborative control generation module, based on the list of to-be-activated items in the module scheduling loading sequence, combined with the current mutation type marker of the node, parses the execution path graph in the template, maps the key path to the edge control task execution channel, executes the task chain in real time, and generates the edge-cloud collaborative control scheduling result. The collaborative control generation module includes: The execution path parsing submodule extracts the corresponding template number and associated node number based on the content of the list to be activated in the module scheduling loading sequence. Combined with the mutation type marker information corresponding to the node in the mutation trigger joint list, it extracts the path graph index that matches the current mutation type, sequentially traverses all path nodes, filters key paths, and obtains the key path node sequence set. The task channel mapping submodule performs task-resource matching judgment on the task type corresponding to each node based on the node number in the critical path node sequence set, establishes the mapping relationship between critical tasks and specific border control channels, completes the construction of the three-dimensional assignment structure of node-channel-task, and outputs the execution priority index and physical mapping location index of each task to generate a task channel mapping matrix. The scheduling action execution submodule issues execution control commands to each channel according to the channel address and task sequence index in the task channel mapping matrix, and schedules them to the edge control interface in a sequential loading manner. For each scheduling, it records the current timestamp, execution channel number and return status code information, integrates all successful scheduling records and organizes them into a structure based on nodes, and establishes the edge-cloud collaborative control scheduling result.

2. The edge-cloud collaborative software-defined industrial control system according to claim 1, characterized in that, The action configuration items include, but are not limited to, starting the cooling fan, maintaining the motor no-load, and calibrating the sensor. The multi-channel analog signals include, but are not limited to, servo system output frequency, spindle temperature, and power supply circuit current. The task chain includes, but is not limited to, turning off the heat source, reducing the voltage, and lowering the speed.

3. The edge-cloud collaborative software-defined industrial control system according to claim 1, characterized in that, The border control basic instruction structure includes action execution identifier, cache trigger flag, and node mapping information. The feedback tag time matrix includes tag timestamp, channel data structure, and node classification index. The mutation trigger joint list includes trigger event record, mutation feature parameter set, and node behavior tag. The module scheduling loading sequence includes scheduling priority classification, remote scheduling identifier, and edge activation flag. The edge-cloud collaborative control scheduling result includes critical path mapping result, task execution status set, and border control response log.

4. The edge-cloud collaborative software-defined industrial control system according to claim 1, characterized in that, The edge control action configuration module includes: The control demand extraction submodule acquires the real-time control signal data stream during the self-test phase of industrial equipment startup. Based on the signal identifier and corresponding startup status carried by each type of signal, it extracts the control demand signals, classifies them by type, and assigns them numbers. Combining the numbering information, it extracts the corresponding control action numbers and generates a set of control action numbers. The basic template matching submodule compares and matches the control number data in the control action number set with the basic instruction set template. For each number information, it performs template instruction indexing and instruction field mapping, arranges the successfully matched basic template instruction sequences in order, and establishes the correspondence between the number and the template field according to the logical order of the control numbers to generate a list of matched basic template instructions. The instruction cache generation submodule constructs the instruction structure field format and cache initialization field set based on all template instruction content in the matching basic template instruction list, writes all structure instruction data into the edge node cache module, starts the structure data loading process and monitors the cache writing status, and obtains the edge control basic instruction structure set.

5. The edge-cloud collaborative software-defined industrial control system according to claim 1, characterized in that, The multi-source signal aggregation module includes: The signal acquisition execution submodule obtains the signal acquisition configuration in the border control basic instruction structure, parses the signal type number and channel acquisition parameters, continuously receives analog signals from each channel within a preset sampling period, parses the sampled values ​​and performs normalization conversion, converts the input analog signal values ​​to the target data range, and adds a sampling time point index at each sampling point to generate a multi-channel sampling data sequence. The channel label construction submodule clusters and sorts the data sequences based on the time information corresponding to the multi-channel sampled data sequences and the channel number. It uses the sampling time information as the main index of the sequence, establishes a time sequence label for each group of sampling points, calculates the time interval to remove data points outside the time error range, and retains the valid sequence points after filtering to construct a labeled data set and generate a channel timestamp label sequence set. The synchronous cache writing submodule writes the corresponding tag data into the synchronous cache space according to the data structure of the channel timestamp tag sequence set, using the node number as the index. It calculates the storage location offset through the mapping relationship between the node and the cache address, encapsulates the tag sequence data, performs the cache writing operation, and obtains the feedback tag time matrix.

6. The edge-cloud collaborative software-defined industrial control system according to claim 1, characterized in that, The state change identification module includes: The signal difference calculation submodule obtains the current, temperature and frequency signal values ​​of each node in the feedback tag time matrix, performs item-by-item difference calculation on the same type of signal in adjacent sampling periods, calculates the composite sudden change fluctuation intensity index, and generates a list of signal sudden change fluctuation values. The mutation condition determination submodule compares the normalized fluctuation value of the signal corresponding to each node in the signal mutation fluctuation value list with the preset current mutation determination coefficient, temperature slope limit value and frequency jump critical value one by one. It adopts the greater than threshold determination method, marks the fluctuation value that meets any critical condition as an abnormal mutation item, records the node where the mutation signal is located, the signal type and the judgment time, and obtains the mutation signal judgment result set. The anomaly labeling output submodule determines the abnormal mutation items in the mutation signal judgment result set, corresponds to the node number, signal type and judgment time in advance, and uniformly constructs an anomaly item identification structure for all records that meet the mutation conditions. It uses a structure list to summarize them in chronological order and uniformly writes them into the node anomaly log cache area to establish a joint list of mutation triggers.

7. The edge-cloud collaborative software-defined industrial control system according to claim 6, characterized in that, The formula for calculating the composite mutation fluctuation intensity index is as follows: ； in, Indicates the first Signal type at time The normalized abrupt change fluctuation intensity, and They represent the first The label values ​​of the signal type at the current and previous time steps. The standard deviation of this type of signal. Used to measure the combined energy level of disturbances in the three signal channels.

8. The edge-cloud collaborative software-defined industrial control system according to claim 1, characterized in that, The cloud-based scheduling module includes: The template filtering and allocation submodule obtains all node numbers and corresponding mutation types in the variable trigger joint list, indexes the corresponding control template structure in the cloud template mapping table according to the node number, extracts the scheduling cycle value set in the template and judges it with the benchmark cycle value, extracts template items with scheduling cycles less than the benchmark value and classifies them into the priority distribution channel, and the remaining template items are classified into the standard scheduling pool. The template distribution priority list is generated according to the number order of the filtering operation. The instruction remote transmission submodule constructs a remote scheduling instruction frame structure and binds the target node number according to the control template number in the template distribution priority list. It then transmits the instruction structure to the corresponding edge node in batches through the cloud-edge communication channel. After the transmission is completed, it records the corresponding number and the issuance time index and establishes a remote template issuance record table. The edge registration and loading submodule locates the scheduling area corresponding to the current edge node based on the template number and node number information in the remote template distribution record table, registers the template number, the channel to which it belongs and the activation identifier in the cache area to be activated, and generates a control index sequence structure in chronological order after completing the loading and registration action, and establishes the module scheduling and loading sequence.

Images