An edge intelligent computing method and system supporting multi-protocol awareness and rule-driven processing

Abstract

The application discloses an edge intelligent computing method and system supporting multi-protocol awareness and rule-driven processing, relates to the technical field of edge intelligent computing, and comprises the following steps: identifying protocol characteristics by listening to communication behaviors, and completing semantic registration in parallel with an animal model structure; collecting data streams based on the field structure of semantic registration, and performing rule judgment on the fields of semantic registration; when the rule judgment meets a trigger condition, executing a function chain bound with an event, outputting a response result, and feeding back a running state. The method disclosed by the application constructs a complete edge computing process of protocol characteristic identification, field semantic registration, rule-driven calculation and function chain response processing, realizes device plug-and-recognize, data structured processing and event intelligent response, has high adaptability, high real-time performance and cloud-edge collaborative evolution capability, and significantly improves the autonomy and response efficiency of the edge intelligent system.

Description

Technical Field

[0001] This invention relates to the field of edge intelligent computing technology, specifically to an edge intelligent computing method and system that supports multi-protocol perception and rule-driven processing. Background Technology

[0002] With the rapid development of the Industrial Internet of Things (IIoT) and edge computing, higher demands are being placed on the protocol compatibility, data processing, and event response of multi-source heterogeneous devices at the edge. Traditional centralized processing architectures are no longer sufficient to meet the application requirements of low latency, high throughput, and local autonomy. In recent years, researchers and engineering practitioners have gradually promoted the integration of key technologies such as "automated protocol parsing," "object model-driven management," and "edge rule execution engines," driving edge intelligence to evolve from static deployment to dynamic perception, rule-driven processes, and real-time linkage. In practical applications, how to achieve automatic device identification, structured semantic parsing, and complex condition judgment at the edge has become the core foundation for ensuring real-time control and intelligent response capabilities in industry.

[0003] Although some edge computing systems currently support multi-protocol data access and rule judgment, they still generally suffer from poor adaptability, lagging triggering mechanisms, and fragmented rule execution processes. Existing systems mostly rely on manually predefined communication protocol configuration files, lacking the ability to identify protocol features, resulting in long access cycles for new devices and making it difficult to achieve automatic modeling and semantic recognition. Although some platforms have introduced rule judgment mechanisms, the structure of rule expressions is generally simplified, lacking the ability to express based on time windows, multi-field interactions, and context awareness, making it difficult to adapt to complex real-world business scenarios. Existing event response strategies often rely on mapping a single rule to a single instruction, lacking function chain execution and status feedback capabilities, resulting in coarse-grained responses and a lack of linkage. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that existing edge data processing and event judgment methods lack an automatic protocol identification mechanism, have weak rule expression capabilities that cannot adapt to complex temporal logic, and have a disconnect between event response and function processing. The problem is how to achieve a closed-loop self-driven intelligent computing process from device identification, rule judgment to response execution at the edge.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: an edge intelligent computing method supporting multi-protocol perception and rule-driven processing, comprising: identifying protocol features by monitoring communication behavior; completing semantic registration by parallel animal model structure; collecting data streams based on the semantically registered field structure; performing rule judgment on the semantically registered fields; executing a function chain bound to the event when the rule judgment meets the triggering condition, outputting the response result, and providing feedback on the running status; the animal model structure includes the naming rules, numerical types, logical levels, and operation method sets of attribute fields; the rule judgment includes: traversing the archived cached data in chronological order, with each time window serving as a rule judgment cycle; executing computational tasks within each cycle; performing real-time judgment on rule conditions and generating judgment results; substituting field values ​​from the data stream into the corresponding rule expressions; and performing Boolean calculation operations.

[0007] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol awareness and rule-driven processing described in this invention, the monitoring communication behavior includes: extracting frame start characters, field offset structures, encoding formats and verification rules from device communication data, and constructing a protocol feature vector; matching the protocol feature vector with a preset protocol family to identify the communication protocol type.

[0008] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol perception and rule-driven processing described in this invention, the method of identifying communication protocol type includes calling the object model template corresponding to the communication protocol. The object model template includes the attribute name, data type, logical hierarchy structure, and supported operation method definition of the accessed device fields, and is used to generate the field parsing structure.

[0009] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol awareness and rule-driven processing described in this invention, the data stream includes: a field structure based on semantic registration, collecting device communication data according to a preset sampling period, parsing it into a structured format, and encapsulating it into a standardized data record containing field names, values, timestamps, and device identifiers.

[0010] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol perception and rule-driven processing described in this invention, the rule judgment includes: structured data records are organized into time window buffers, each window serves as a rule judgment period, and the rule judgment performs traversal analysis on the recorded data within each period.

[0011] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol perception and rule-driven processing described in this invention, each window, as a rule judgment cycle, includes substituting each field value into the rule expression associated with the field, combining historical averages and rate change time-series information, and performing Boolean calculations to determine whether the event triggering conditions are met.

[0012] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol perception and rule-driven processing described in this invention, the following is included: satisfying the event triggering conditions includes binding events that satisfy the triggering conditions to a set of functions, selecting the function execution order according to the event type, using the output of the previous function as the input of the next function, and generating a response result after the chained functions are executed.

[0013] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol perception and rule-driven processing described in this invention, the response result includes event level classification, function call trajectory, processing time, field labels and response conclusion, and is output in a structured data format for recording and scheduling reference.

[0014] As a preferred embodiment of the edge intelligent computing method supporting multi-protocol perception and rule-driven processing described in this invention, the operating state includes response accuracy, event triggering frequency, and rule hit rate.

[0015] Another objective of this invention is to provide an edge intelligent computing system that supports multi-protocol awareness and rule-driven processing. This system can collect data streams through a field structure based on semantic registration and perform rule judgment on the semantically registered fields, thus solving the problem that current edge data processing and event judgment methods have weak rule expression capabilities and cannot adapt to complex temporal logic.

[0016] As a preferred embodiment of the edge intelligent computing system supporting multi-protocol perception and rule-driven processing described in this invention, it includes: a communication protocol feature recognition and object model linkage registration module, a structured field-based data acquisition and rule judgment module, and an event response function execution and running status feedback module; the communication protocol feature recognition and object model linkage registration module is used to extract communication protocol feature information by monitoring device communication behavior and matching preset object model templates based on the features; the structured field-based data acquisition and rule judgment module is used to continuously collect device data streams based on the registered field structure, organize them into cached data blocks according to time windows, and substitute field values ​​into rule expressions for real-time Boolean logic calculation; the event response function execution and running status feedback module is used to execute the function chain task bound to the event if the rule conditions are met, output the response conclusion, and construct a running summary for rule and model version feedback and cloud-edge collaborative updates.

[0017] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement an edge intelligent computing method supporting multi-protocol perception and rule-driven processing.

[0018] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of an edge intelligent computing method that supports multi-protocol perception and rule-driven processing.

[0019] The beneficial effects of this invention are as follows: The edge intelligent computing method supporting multi-protocol awareness and rule-driven processing provided by this invention identifies protocol features by monitoring communication behavior and automatically matches the object model structure, breaking through the limitations of traditional methods that rely on manual configuration of protocols and field rules. This enables plug-and-play device recognition and automatic binding of field semantics, effectively improving the access efficiency and management consistency of heterogeneous devices. Secondly, through a data acquisition mechanism based on registered field structures, the collected values ​​are injected into rule expressions. Boolean judgments are performed by combining features such as time windows, field offsets, and historical trends to construct a rule-driven computing capability that supports context awareness, thereby meeting the needs of event recognition under complex conditions. Finally, response tasks associated with events are executed in a chain-like manner, and the running status is fed back in the form of a structured summary to drive cloud-based policy collaboration and version evolution. The overall solution not only improves the real-time decision-making capability and local intelligent response depth of edge computing, but also constructs a cloud-edge collaborative and adaptively evolving edge intelligent architecture, with significant advantages such as strong deployment flexibility, transparent response mechanism, and scalable rule management. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 The first embodiment of the present invention provides an overall flowchart of an edge intelligent computing method that supports multi-protocol perception and rule-driven processing. Detailed Implementation

[0022] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0023] Example 1, referring to Figure 1 As an embodiment of the present invention, an edge intelligent computing method supporting multi-protocol awareness and rule-driven processing is provided, comprising:

[0024] S1: Identify protocol features by monitoring communication behavior and complete semantic registration using a parallel animal model structure.

[0025] Furthermore, by monitoring the data exchange behavior during the initial communication of the monitoring device, communication features including frame start characters, field offset patterns, data encoding formats, and checksum methods are extracted to form a protocol feature vector. This vector is used to characterize the type of communication protocol used by the device.

[0026] It should be noted that the extracted protocol feature vectors are matched with predefined protocol feature clusters, and the protocol category to which the communication behavior belongs is identified through similarity measurement, classifying it into a specified protocol family (e.g., Modbus, MQTT, OPC, etc.). The behavior identification is based on communication patterns rather than protocol identifier fields, and is adaptive.

[0027] It should also be noted that, based on the identified protocol category, a data structure template matching the protocol family is selected from the locally defined set of device object models. This template includes naming rules, numeric types, logical levels, and a set of operable methods for multiple attribute fields, used to standardize the format of subsequent data acquisition and processing.

[0028] Furthermore, based on the selected template, a unique semantic structure description is generated for the device, including field encoding and semantic tag mapping rules. A unique identifier is constructed by combining the device's network attributes (such as IP address, port, and communication path), and this identifier is written into the list of devices to be processed as a registration entry point. It should be noted that after completing the above identification and modeling, the device identifier, field structure, data format, and the associated object model structure are packaged into a unified data record and stored in the logical processing list for direct reference by the data acquisition and rule calculation steps, without manual intervention.

[0029] S2: Collect data streams based on the semantically registered field structure and perform rule-based judgments on the semantically registered fields.

[0030] Furthermore, after completing device protocol identification and object model binding, continuous real-time acquisition, protocol parsing, and format normalization of communication data are performed to support edge data standard access in high-frequency scenarios. Based on the generated device identifier and field structure definition, uplink data frames from the device are continuously received from the corresponding communication channel according to the set sampling period. The acquisition frequency can reach the second or millisecond level to meet the needs of high-frequency signal processing.

[0031] It should be noted that a protocol-level parsing operation is performed on the received data frame to identify the address offset, data type, starting position, and parsing method corresponding to the field content. The parsing behavior is based on the field templates defined in the binding model, ensuring semantic consistency during the extraction process.

[0032] It should also be noted that the field values ​​parsed from each data frame are associated and mapped with their logical tags, and the results are organized into a unified structured format (such as key-value pairs or standard JSON structure), while also attaching the collection timestamp, device identifier and sampling batch number.

[0033] Furthermore, the standardized data is archived and temporarily buffered according to a set time window, which can be a sliding window, a skip window, or an event-based dynamic window. The buffer content is used as input for subsequent rule triggering and streaming computation steps.

[0034] It should be noted that the constructed window cache data is labeled with usable tags and declared as an input data stream that can be called by the rule calculation engine. This allows functional modules such as event triggering and expression condition judgment to directly reference it, thus establishing a data-driven processing chain.

[0035] It should also be noted that a predefined set of rule expressions is read, and each rule consists of a Boolean condition, a threshold setting, comparison logic, and a trigger action identifier. The field names and data structures referenced are derived from the cached data format; each rule is bound to a unique device identifier to ensure that the calculation objective is clear and the context is consistent.

[0036] Furthermore, according to the defined time window division method, the archived cached data is traversed in chronological order. Each time window is regarded as a rule judgment period; within each period, parallel rule calculation tasks for a single device and multiple fields are executed.

[0037] It should be noted that substituting field values ​​from the data stream into the corresponding rule expression and performing Boolean calculations is represented as follows:

[0038] ;

[0039] in, This represents a regular expression function based on the linkage of multiple fields in a data stream. The sampled value of the main field in the current data stream. This is the second field value (e.g., temperature or current) within the same cycle. For fields The sliding history sampling sequence, For sequence The moving average, For sequence The sliding standard deviation, This is the sampled value of the second field in the previous time period. For fields The rate of change of is the first difference value. To prevent small constants with zero denominators (such as...) ), The lower threshold constant for rule triggering is usually set to 0.85 to control the sensitivity of events.

[0040] The range of values ​​is ,when When the rule expression is satisfied, an event is triggered; when When the value is close to 1, only logs are recorded and no events are triggered; a value close to 1 indicates that the current data state is highly abnormal or highly consistent with the rules, while a value close to 0 indicates that the data deviates from the rules.

[0041] If the judgment result is "true", that is, the condition is met, a structured event judgment result is generated immediately; if the condition is not met, no event is output in the current cycle, only the log status is retained.

[0042] The event determination result includes: event type, trigger time, trigger field, trigger value and corresponding device ID, which are used for subsequent processing.

[0043] It should also be noted that, based on the event type specified in the event determination result, a preset set of function identifiers is searched and associated. Functions can be diagnostic functions, exception label generators, predictive inference models, early warning strategy functions, etc.; the function is not actually executed in this step, but only its call preparation and parameter encapsulation are completed here; the encapsulated function call task is packaged into an instruction stream and output along with the event for the next step. The event result and the bound function instruction stream are jointly encapsulated into an "event-function pair," which is pushed to subsequent processing stages in the form of a data structure, realizing the logical transition between rule judgment and intelligent processing.

[0044] S3: If the rule determines that the triggering condition is met, execute the function chain bound to the event, output the response result, and provide feedback on the running status.

[0045] Furthermore, for events that meet the specified conditions, the corresponding function instruction structure is determined by the event type and parameter fields. Based on information such as field labels, numerical intensity, and trigger time contained in the event structure, a set of matching response functions is selected. Each function is defined as follows:

[0046] ;

[0047] in, This represents the set of matched response functions. This represents the absolute value of the main field in the event determination. This indicates the offset of the field relative to its historical average. Indicates the rate of change when an event occurs.

[0048] It should be noted that if the event belongs to the category of high-priority chain responses (such as prediction) control (Record), then the current function The output is used as the next function. The input is used to continue executing subsequent function chains. Each function level can change its dimensions or perform reasoning and classification, ultimately generating a comprehensive response result.

[0049] It should also be noted that the output value of the final first-level function chain is compared with a preset response threshold: if it is greater than the judgment threshold... If the event is within the critical range, a control command will be generated; if it is in the middle range, a medium-level record label will be generated; if it is below the lower threshold, it will be marked as minor or ignored.

[0050] After local event response processing is completed, in order to improve the long-term adaptability of rule configuration and model strategy, a collaborative approach based on version tracking and strategy synchronization is adopted to achieve continuous consistency and secure evolution between edge processing logic and remote control strategy.

[0051] The generated event response information, function call traces, output labels, and response levels are summarized and used to construct periodic runtime summary data. This summary includes field statistical characteristics, event count, response success rate, and data coverage. This summary is used to evaluate the current rule effectiveness and the execution status of the model behavior.

[0052] The currently deployed rule expression structure and function chain logic are hashed to generate a globally unique version fingerprint number; at the same time, its loading time, the most recent modification record and the version number of the source file in the cloud are recorded to form a complete version triple: local version number, local hash code and the corresponding original version number in the cloud.

[0053] Based on the performance data reported in the runtime summary, determine whether the current rules have issues such as response latency, increased false positives, or event redundancy. If any preset threshold is triggered, request a remote update suggestion via the cloud-edge communication channel and receive the new version of the rule expression, function structure, or policy parameter set.

[0054] Update operations are controlled by version identifiers and are only executed when the cloud version number is higher than the local version number, or when the local version number is marked as "pending update".

[0055] The received new version content undergoes structural consistency verification and instruction specification conversion, and then replaces the current local version. Simultaneously, an update record is generated, including a summary of the update content, the previous and current version numbers, the replacement time, the execution result, and whether rollback protection was triggered. If the update fails or is incompatible with the local data structure, a rollback operation is triggered, restoring the previous valid version.

[0056] After the update, the new rule configuration and function chain will be registered in the edge configuration manifest, along with version identifiers and loading information. This information will serve as a historical baseline in subsequent runtime summaries for performance evaluation, supporting future multi-version comparisons and dynamic adjustments.

[0057] Example 2, an embodiment of the present invention, provides an edge intelligent computing system that supports multi-protocol perception and rule-driven processing, including a communication protocol feature recognition and object model linkage registration module, a data acquisition and rule judgment module based on structured fields, and an event response function execution and running status feedback module.

[0058] The communication protocol feature recognition and object model linkage registration module is used to extract communication protocol feature information by monitoring device communication behavior, and construct a protocol feature vector based on the extraction results; compare the protocol feature vector with a preset protocol family to identify the communication protocol type; then match the corresponding object model template according to the identification results to complete the registration of device field structure, field naming rules, numerical type, logical level and operation method; generate a unique identifier by combining the device network attributes and write it into the semantic structure list for downstream data collection and rule calculation processes to call.

[0059] It should also be noted that after completing the extraction of device communication features and the identification of protocol types, the communication protocol feature recognition and object model linkage registration module loads the corresponding object model template based on the identification results and generates a complete semantic description structure containing field structure, data type, named labels, and operation methods. This structure, together with the device's network attributes, generates a unique identifier and is uniformly encapsulated as standard structured registration information for direct reference by subsequent modules. This achieves automatic connection between field templates, protocol parsing methods, and rule parameter configurations, avoiding manual configuration operations and improving the system's automatic deployment capabilities and device adaptability.

[0060] The data acquisition and rule judgment module based on structured fields is used to continuously collect data frame content uploaded by the device based on communication protocol feature recognition and semantic registration results; it parses field values ​​according to the set sampling period and time window mode, performs protocol-level decoding, unit conversion and type normalization processing, and outputs structured data records; further, it substitutes the field values ​​in the data stream into preset rule expressions, performs Boolean logic calculations in combination with the field change trend within the time window, determines whether the event triggering conditions are met, and constructs an event-function binding structure for subsequent response processing.

[0061] It should also be noted that the data acquisition and rule judgment module based on structured fields utilizes the semantic structure and field configuration provided by the registration module to perform high-frequency data acquisition, and organizes the acquired data stream according to a time window caching method to form a standardized field data structure. During the rule judgment phase, the system dynamically injects the values ​​of each field into the rule expression, performs Boolean logic calculations, and determines whether the event triggering conditions are met. When the rule meets the conditions, the system constructs an event-function binding relationship, binding the event type, field label, and trigger strength to the function identifier, encapsulating it into an event-function pair structure, and pushing it to the response processing module.

[0062] The event response function execution and runtime status feedback module receives the event-function binding structure generated based on rule judgment, selects the matching function set according to the event type, executes the response function tasks sequentially in a chained order, and outputs the response result and processing label. At the same time, it collects the function trajectory, processing time, event level and trigger field, etc. during the response execution process, constructs a runtime status summary, and feeds it back to the remote management terminal. When the system receives a new version policy from the cloud, it performs policy replacement, function chain upgrade or automatic rollback operation according to the version fingerprint and verification mechanism to ensure that local rules and models are continuously optimized and consistent with cloud-edge collaboration.

[0063] It should also be noted that after receiving an event-function pair, the event response function execution and runtime status feedback module selects and executes the corresponding function chain task according to the event type, and outputs the processing result, response tag, and call path. Simultaneously, the system records the response latency, event level, and field participation information during function execution, constructing a runtime status summary for the cloud system to perform rule performance evaluation and policy adjustment. When the system receives a new version of rule configuration or function set from the cloud, the module performs difference verification based on the local version fingerprint information. After ensuring structural compatibility, it performs version replacement or rollback, completing the closed loop of dynamic policy evolution and cloud-edge collaborative update for edge nodes.

Claims

1. An edge intelligence computing method supporting multi-protocol awareness and rule-driven processing, characterized in that, Comprise: Identify protocol features by monitoring communication behavior, and complete semantic registration with animal model structure; Collect data stream based on the field structure of semantic registration, and perform rule judgment on the fields of semantic registration; Rule judgment meets the trigger condition, executes the function chain bound to the event, outputs the response result, and feeds back the running state; The physical model structure includes the naming rules, numerical types, logical levels, and operation method set of attribute fields; Rule judgment includes traversing the archived cache data in chronological order, taking each time window as a rule judgment period, performing calculation tasks in each period, and generating judgment results by real-time judging rule conditions and substituting field values in the data stream into corresponding rule expressions to perform Boolean calculation operations; Each window as a rule judgment period includes substituting each field value into the rule expression associated with the field, combining historical mean value and rate change time series information, and performing Boolean calculation to determine whether the event trigger condition is met; Satisfying the event trigger condition includes that the event meeting the trigger condition is bound to a function set, the function execution order is selected according to the event type, the output of the previous function is taken as the input of the next function, and a response result is generated after the chain function execution is completed.

2. The edge intelligence computing method of claim 1, wherein: The monitoring of communication behavior includes, Extracting the frame start symbol, field offset structure, encoding format, and check rule in device communication data, and constructing a protocol feature vector; The protocol feature vector is matched with the preset protocol family to identify the communication protocol type.

3. The edge intelligent computing method of claim 2, wherein: The identification of the communication protocol type includes, Calling the physical model template corresponding to the communication protocol, which includes the attribute name, data type, logical level structure, and supported operation method definition of the accessed device field, and is used to generate a field parsing structure.

4. The edge intelligent computing method of claim 3, wherein: The data stream collection includes, Based on the field structure of semantic registration, collect device communication data according to the preset sampling period, and parse it into a structured format, and encapsulate it into a standardized data record containing field name, value, timestamp, and device identifier.

5. The edge intelligent computing method of claim 1, wherein: The rule judgment includes, The structured data record is organized into a time window cache area, and each window is taken as a rule judgment period. Rule judgment traverses and analyzes record data in each period.

6. The edge intelligent computing method of claim 1, wherein: The response result includes, Event level classification, function call track, processing time, field label, and response conclusion, and is output in a structured data format for record and dispatch reference.

7. The edge intelligent computing method of claim 1, wherein: The running state includes, Response accuracy, event trigger frequency, and rule hit rate.

8. An edge intelligent computing system supporting multi-protocol awareness and rule-driven processing, adopting the edge intelligent computing method supporting multi-protocol awareness and rule-driven processing according to any one of claims 1-7. Comprise communication protocol feature identification and physical model registration module, structured field-based data collection and rule judgment module, event response function execution and running state feedback module; The communication protocol feature identification and physical model registration module is used to extract communication protocol feature information by monitoring device communication behavior, and match the feature with the preset physical model template; The structured field-based data collection and rule judgment module is used to continuously collect device data stream based on the registered field structure, organize it into cache data blocks according to time windows, and substitute field values into rule expressions for real-time Boolean logic calculation; The event response function executes the function chain task bound with the event, outputs the response conclusion, and constructs the operation summary for version feedback and cloud-edge collaborative update of the rule and the model if the rule condition is met. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor implements the steps of the edge intelligent computing method supporting multi-protocol awareness and rule-driven processing of any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The processor implements the steps of the edge intelligent computing method supporting multi-protocol awareness and rule-driven processing of any one of claims 1 to 7 when executing the computer program.

Images