An internet of things intelligent multimedia integrated discipline table control method and system

Abstract

The present application relates to program control system technical field, specifically for a kind of thing internet of things intelligent multimedia integrated education table control method and system, including the following steps: receiving from thing internet of things intelligent multimedia integrated education table original event, the original event is remote forced intervention instruction, online examination task and user autonomous learning request, each event identifier is matched with the preset event priority definition table.Processing logic in the present application, through the structured priority identification of original event, remote intervention, examination task, autonomous learning and other multi-source events are classified by relying on static priority mechanism, realize the structural ordering control between tasks, improve the distinguishability of multi-task scheduling.Introduce preemption judgment mechanism under the priority driving, in combination with task context saving and resource exclusive judgment, form the dynamic switching strategy of condition trigger, so that terminal running logic has the traceability and preemption flexibility.

Description

Technical Field

[0001] This invention relates to the field of program control system technology, and in particular to a control method and system for an IoT-enabled intelligent multimedia integrated discipline desk. Background Technology

[0002] The field of program control system technology mainly involves system behavior control and scheduling mechanisms based on preset logic, operation models or rule-driven approaches, and is widely used in industrial automation, intelligent terminal control, human-machine interface, network communication management, embedded platforms and other scenarios.

[0003] Existing technologies for task scheduling often rely on static flowcharts or event-triggered mechanisms, lacking the ability to prioritize and dynamically compare task attributes. This results in high-value tasks being difficult to process promptly in resource-constrained scenarios, impacting system efficiency. In behavior recognition, relying solely on behavior logs for single-point event responses leads to delayed and inaccurate behavior management responses. Existing control strategies often map execution instructions to single rules, lacking dynamic correlation with behavior states or task weights. This can easily lead to generalized control operations, weakening the guidance effect on individual behaviors. For example, in educational management scenarios, when students repeatedly switch between non-learning windows without constituting a violation, existing solutions cannot quantify the impact of the behavior or generate appropriate intervention strategies, easily leading to false alarms or control failures, reducing the sensitivity and practicality of management. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an IoT-based intelligent multimedia integrated discipline desk control method and system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for controlling an IoT-enabled smart multimedia integrated discipline desk, comprising the following steps:

[0006] Receive raw events from the IoT smart multimedia integrated discipline desk, the raw events being remote mandatory intervention commands, online examination tasks, and user self-learning requests. Match each event identifier with a preset event priority definition table, attach a static priority value to each task request event, and generate a priority task sequence.

[0007] Based on the priority task sequence, a priority set is obtained and generated. If the priority value of the newly arrived task is higher, the current task is interrupted and the context is saved to establish a preemptive scheduling execution scheme. If the priority value of the newly arrived task is not higher, it is further determined whether the requested resource is exclusively locked. If so, the task request is placed in the waiting queue.

[0008] Based on the application start-up and window switching events recorded by the preemptive scheduling execution scheme, a continuous behavioral event stream is extracted. The behavioral event stream is then compared with a preset positive behavior pattern library and a negative behavior pattern library to establish a matching behavior pattern set. The matching behavior pattern set is then traversed to calculate the user's digital health status code.

[0009] Based on the user's digital health status code, a corresponding strategy is selected, a selected guidance strategy identifier is generated, and a corresponding operation instruction is retrieved according to the selected guidance strategy identifier. The operation instruction is to lock non-learning applications, display prompt information on the discipline desk screen, or adjust network interface access permissions, thereby constructing terminal device control instructions.

[0010] Preferably, the step of obtaining the priority task sequence is as follows:

[0011] Receive raw events from the IoT smart multimedia integrated discipline desk, parse the event identifier in each raw event, read the event identifier character sequence one by one through the traversal method, and extract the feature fields representing the event type from the event identifier character sequence to form a set of event identifier feature fields;

[0012] Based on the set of event identifier feature fields, each event identifier feature field is matched against a preset event priority definition table one by one, and the priority value corresponding to each event feature field in the event priority definition table is called to obtain a set of static priority values ​​that match each event identifier feature field.

[0013] Based on the set of static priority values, the original events are bound one by one to their corresponding static priority values, and the original events are rearranged in descending order of priority values ​​to form a priority task sequence.

[0014] Preferably, the steps for obtaining the priority set are as follows:

[0015] Based on the priority task sequence, the newly arrived task entries and currently running task entries in the task assignment record table are parsed, and the bound priority field values ​​are read to obtain the priority set.

[0016] Preferably, the step of obtaining the preemptive scheduling execution scheme is as follows:

[0017] Calculate the dynamic preemption decision score based on the priority set;

[0018] Based on the dynamic preemption decision score, when the dynamic preemption decision score is greater than the system preemption trigger threshold, the currently running task is interrupted and a complete running context snapshot is written. If the resource requested by a newly arrived task is in a locked state, the task request is pushed into the waiting queue. A resource allocation mapping table is constructed by combining the current task status, the status of the newly arrived task, and the priorities of all tasks in the waiting queue. A preemptive scheduling execution scheme is generated based on the resource allocation mapping table and the context snapshot.

[0019] Preferably, the step of obtaining the matching behavior pattern set is as follows:

[0020] Based on the application start / stop and window switching events recorded by the preemptive scheduling execution scheme, the behavior tag, timestamp and priority value of the task to which the event belongs are extracted in chronological order, and a time-series triplet sequence containing the event behavior tag, the event occurrence time and the priority of the associated task is constructed to obtain a continuous behavior event stream.

[0021] Based on the continuous stream of behavioral events, the behavioral label sequence is compared with the behavioral structure of the positive behavioral pattern library and the negative behavioral pattern library one by one. The sliding matching method is used to record the behavioral pattern matching position. At the same time, the corresponding priority value and behavior occurrence time are extracted from the triples to obtain the matching behavioral pattern set.

[0022] Preferably, the steps for obtaining the user's digital health status code are as follows:

[0023] The user's digital health status code is calculated based on the set of matching behavior patterns.

[0024] Preferably, the step of obtaining the selected guidance strategy identifier is as follows:

[0025] Based on the user's digital health status code, the threshold of each level in the multi-level score threshold table is called sequentially, the numerical difference between the user's digital health status code and the current threshold is calculated level by level, and it is determined whether the numerical difference reaches the policy triggering condition defined in the threshold table to obtain the policy triggering judgment result.

[0026] Based on the policy trigger judgment result, the policy entries corresponding to the trigger judgment result are retrieved item by item from the policy configuration table, the unique identifier of the guidance policy marked in the policy entry is extracted, and the selected guidance policy identifier is generated based on the unique identifier of the guidance policy.

[0027] Preferably, the step of obtaining the terminal device control command is as follows:

[0028] Based on the selected boot strategy identifier, the terminal device control instruction set is retrieved, the control instruction entry corresponding to the selected boot strategy identifier is called, the operation instructions in the control instruction entry are parsed, and the terminal device control instructions are generated.

[0029] This invention also provides a discipline desk control system, comprising:

[0030] The event processing module receives raw events from the IoT smart multimedia integrated discipline desk. The raw events are remote mandatory intervention commands, online examination tasks, and user self-learning requests. It matches each event identifier with a preset event priority definition table, adds a static priority value to each task request event, and generates a priority task sequence.

[0031] The task scheduling module obtains a priority set based on the priority task sequence and generates a priority set. If the priority value of the newly arrived task is higher, the current task is interrupted and the context is saved to establish a preemptive scheduling execution scheme. If the priority value of the newly arrived task is not higher, it continues to determine whether the requested resource is exclusively locked. If so, the task request is placed in the waiting queue.

[0032] The behavior analysis module extracts a continuous stream of behavior events based on the application start-up and shutdown and window switching events recorded by the preemptive scheduling execution scheme. It then compares the behavior event stream with a pre-set positive behavior pattern library and a negative behavior pattern library to establish a matching behavior pattern set. Finally, it traverses the matching behavior pattern set to calculate the user's digital health status code.

[0033] The strategy execution module selects the corresponding strategy based on the user's digital health status code, generates a selected guidance strategy identifier, and retrieves the corresponding operation instructions based on the selected guidance strategy identifier. The operation instructions are to lock non-learning applications, display prompt information on the disciplinary desk screen, or adjust network interface access permissions, thereby constructing terminal device control instructions.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0035] In this invention, the processing logic involves structured priority identification of original events, categorizing multi-source events such as remote intervention, examination tasks, and self-learning into hierarchical classes based on a static priority mechanism. This achieves structured sorting control between tasks, improving the identifiability of multi-task scheduling. A preemption judgment mechanism is introduced under priority-driven conditions, combining task context saving and resource exclusivity judgment to form a condition-triggered dynamic switching strategy, enabling the terminal's operational logic to have traceability and preemption flexibility. A continuous stream of behavioral events is formed by recording task start / stop and window switching behaviors. Positive and negative behavior comparisons and behavior pattern recognition are performed based on the behavioral sequence, enhancing the accuracy of behavioral perception of user status. A user digital health status code is constructed through sequence accumulation and deduction, and a response strategy is established based on the correspondence between this status code and graded thresholds, giving terminal feedback behavior a status-driven attribute. Device control commands are generated based on the strategy, completing responses to non-learning operation restrictions, network permission adjustments, and information prompts, ensuring the closed-loop consistency of behavior management and resource scheduling. This achieves integrated processing from task control to behavior intervention, improving the dynamic adaptability to complex application scenarios. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the steps of the present invention. Detailed Implementation

[0037] 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.

[0038] Please see Figure 1 This invention provides a technical solution: a method for controlling an IoT-enabled smart multimedia integrated discipline desk, comprising the following steps:

[0039] Receive raw events from the IoT smart multimedia integrated discipline desk. The raw events are remote mandatory intervention commands, online examination tasks, and user self-learning requests. Match each event identifier with a preset event priority definition table, attach a static priority value to each task request event, and generate a priority task sequence.

[0040] Based on the priority task sequence, obtain the priority set, generate the priority set, if the priority value of the newly arrived task is higher, interrupt the current task and save the context, and establish a preemptive scheduling execution scheme; if the priority value of the newly arrived task is not higher, continue to determine whether the requested resource is exclusively locked, if so, put the task request into the waiting queue.

[0041] Based on the application start-up, shutdown and window switching events recorded by the preemptive scheduling execution scheme, a continuous behavioral event stream is extracted. The behavioral event stream is then compared with a pre-set positive behavioral pattern library and a negative behavioral pattern library to establish a matching behavioral pattern set. The matching behavioral pattern set is then traversed to calculate the user's digital health status code.

[0042] Based on the user's digital health status code, select the corresponding strategy, generate the selected guidance strategy identifier, and retrieve the corresponding operation instructions according to the selected guidance strategy identifier. The operation instructions are to lock non-learning applications, display prompt information on the discipline desk screen, or adjust network interface access permissions, thus constructing terminal device control instructions.

[0043] The steps to obtain a priority task sequence are as follows:

[0044] Receive raw events from the IoT smart multimedia integrated discipline desk, parse the event identifier in each raw event, read the event identifier character sequence one by one through the traversal method, and extract the feature fields representing the event type from the event identifier character sequence to form a set of event identifier feature fields;

[0045] Based on the set of event identifier feature fields, each event identifier feature field is matched against a preset event priority definition table one by one, and the priority value corresponding to each event feature field in the event priority definition table is called to obtain a set of static priority values ​​that match each event identifier feature field.

[0046] Based on the set of static priority values, the original events are bound one by one to their corresponding static priority values, and the original events are rearranged in descending order of priority values ​​to form a priority task sequence.

[0047] Specifically, through the built-in message queue interface, such as the subscription endpoint of the MQTT or AMQP protocol, it receives raw event data packets in JSON format reported by various functional components of the IoT smart multimedia integrated teaching desk in real time. Each data packet contains a unique event ID, an event timestamp, and a structured event identifier string. The event identifier follows the specification of "source_category_action_additional information", with fields separated by underscores. For example, the event identifier for a remote mandatory intervention command from the teacher's end is represented as "TCON_CMD_FORCE_INTERVENE_001". The online exam task is identified as "SYS_TASK_EXAM_MATH_102", while a user-initiated self-study request is identified as "STU_REQ_STUDY_VIDEO_PHY03". Upon receiving the original event, the system immediately initiates the parsing process. This process first extracts the event identifier string from the JSON data packet. Then, using a string splitting function with a preset underscore as the delimiter, it decomposes the event identifier string into a string array. For example, for the identifier "TCON_CMD_FORCE_INTERVENE_001", after splitting, it yields the array ["TCON", "CMD", ...]. To accurately identify the core nature of an event, the system extracts feature fields representing the event type from this array according to preset rules. These rules define fields at specific positions in the identifier as feature fields. Specifically, the system identifies the first three elements of the segmented array—the source field, the category field, and the action field—as the core feature fields, ignoring the final additional information field. Therefore, for the example above, the extracted feature fields are ["TCON", "CMD", "FORCE_INTERVENE"], ["SYS", "TASK", "EXAM"], and ["STU", "REQ", "STUDY_VIDEO"]. After performing this parsing and extraction operation on all the original events to be processed, the feature field arrays corresponding to each event are collected to form the event identifier feature field set.

[0048] Based on the set of event identifier feature fields generated in the previous step, the system calculates the static priority for each raw event. This calculation relies on a preset event priority definition table, which is a set of key-value pairs loaded into memory during system initialization. The table's content is based on the experience and knowledge of educational management experts and statistical analysis of historical task execution data. Each event feature field in the table maps to a base score. For example, "FORCE_INTERVENE" is assigned the highest base score of 100 due to its mandatory nature and importance; "EXAM" corresponds to 80 points; and "STUDY_VIDEO" corresponds to 50 points. Simultaneously, the source and category fields serve as adjustment scores; for example, "TCON" represents the teacher's side. The system assigns a +10-point adjustment value to each of the following: "CMD" as the instruction category, +5 points; "STU" representing the student side, with a lower initial weight, -5 points; and "TASK" and "REQ" each correspond to 0 points. The system initiates a matching calculation process, iterating through the feature field array of each event in the event identifier feature field set. An accumulator is initialized to 0 for each array. Then, the feature field strings are retrieved one by one and queried against the event priority definition table. If the query is successful, the score corresponding to that field is added to the accumulator for that event. If no match is found for a field in the table, its score is recorded as 0. For example, for the feature field ["TCON", ...] For events ["CMD", "FORCE_INTERVENE"], the static priority is calculated as 100 (from "FORCE_INTERVENE") + 10 (from "TCON") + 5 (from "CMD"), for a total of 115. For ["SYS", "TASK", "EXAM"], the score is 80 + 0 + 0, for a total of 80. For ["STU", "REQ", "STUDY_VIDEO"], the score is 50 - 5 + 0, for a total of 45. After calculating the priority for all events, all calculated priority scores are summed to obtain a set of static priority values ​​that match the feature fields of each event identifier.

[0049] Based on the set of static priority values ​​and the corresponding list of original events obtained in the previous step, the system performs task serialization. First, to firmly bind events to their priorities, the system creates a new composite data structure for each original event. This structure contains two main fields: one field stores the complete JSON data packet of the original event, and the other field stores its corresponding static priority value calculated in the previous step. The system iterates through the list of original events and the set of static priority values, matching them according to the unique event ID, and instantiates these composite data structures one by one. This combines each independent event information with its quantified importance into an indivisible unit, forming a temporary, unordered list of task objects. Next, to ensure the efficiency and fairness of task scheduling, the system needs to sort this list of task objects. The sorting process employs a dual standard. The primary sorting key is the static priority value, arranged in descending order of value. This means that tasks with the highest priority value will be placed at the front of the list and receive the highest priority processing rights. To handle cases with the same static priority value, the system introduces a secondary sorting key, namely the timestamp of the original event, arranged in ascending order of timestamp from earliest to latest. This supplementary "first-come, first-served" rule ensures that tasks of equal importance are processed according to the order in which they are initiated, avoiding the situation where later-initiated tasks starve earlier-initiated tasks. The sorting operation is implemented by calling a stable sorting algorithm (such as Timsort), which can efficiently process partially ordered data and ensure the stability of the sorting. After sorting, the originally unordered list of task objects is transformed into a strictly ordered queue, forming a priority task sequence.

[0050] The steps to obtain the priority set are as follows:

[0051] Based on the priority task sequence, the task entries marked as newly arrived and the task entries in the current running status in the task assignment record table are parsed, and the bound priority field values ​​are read to obtain the priority set.

[0052] Specifically, based on the priority task sequence generated in the previous step, the system starts a scheduler monitoring loop, continuously scanning a task assignment record table in memory. This table, as a dynamic data structure, contains fields such as task ID, task status (e.g., newly arrived, running, waiting, completed), bound priority value, task initiation timestamp, and estimated resource requirements. When a new task appears in the priority task sequence, it is immediately added to the task assignment record table, and its status is marked as "newly arrived." This operation triggers a scheduling decision evaluation. The system then performs two query operations in the task assignment record table. The first query locates the task entry with the status "newly arrived," which is the one that was just added. The second query located the task entry with the status "running," which is the active task currently consuming CPU and other resources. In a normal single-task scenario, there is at most one "running" task entry in the table. After locating these two entries, the system directly reads the value of the "bound priority value" field from their respective records. For example, the currently running task is "user self-learning request," which has a priority value of 45 in the record table, while the newly arrived task is "remote forced intervention command," which has a priority value of 115. The system extracts these two values, 45 and 115, and temporarily stores them as a set containing two elements, thus obtaining the priority set.

[0053] The steps to obtain the preemptive scheduling execution plan are as follows:

[0054] Based on the priority set, the dynamic preemption decision score is calculated using the following formula:

[0055] ;

[0056] in, This indicates the dynamic acquisition of decision points. This indicates the priority value of the newly arrived task. This indicates the priority value of the currently running task. This indicates the expected resource consumption of the newly arrived task. This represents the smallest positive number to prevent division by zero anomalies. This indicates the real-time resource usage of the currently running task. Indicates the total system resource capacity. This represents the adjustment coefficient for the product of priority density and resource pressure.

[0057] Based on the dynamic preemption decision score, when the dynamic preemption decision score is greater than the system preemption trigger threshold, the currently running task is interrupted and a complete running context snapshot is written. If the resource requested by a newly arrived task is in a locked state, the task request is pushed into the waiting queue. A resource allocation mapping table is constructed by combining the current task status, the status of the newly arrived task, and the priorities of all tasks in the waiting queue. A preemptive scheduling execution scheme is generated based on the resource allocation mapping table and the context snapshot.

[0058] Specifically, the formula: The advantage of the formula lies in the introduction of a dynamic system state factor, which enables fine-grained adjustment of scheduling decisions. (Part 1) This directly reflects the static difference in importance between the old and new tasks, forming the basis for decision-making. The second part is a dynamic adjustment item, in which... This item defines the "priority density" of a new task, which is the priority benefit that a unit of expected resource consumption can bring. This makes it easier for tasks with the same priority and lower resource requirements to preempt successfully. This term quantifies the current system's "resource pressure." The higher the system's resource utilization rate, the closer this term's value is to 1, thus increasing the preemption decision score and giving high-priority tasks a stronger preemption capability when the system is busy. This is achieved by adjusting the coefficient... By weighting these two dynamic factors, the scheduler can make a more intelligent and balanced judgment between the intrinsic importance of a task and the real-time load of current system resources, thus avoiding the problem that high-priority critical tasks cannot respond in a timely manner due to low-priority but resource-intensive tasks occupying resources for a long time.

[0059] This parameter represents the priority value of the newly arrived task. It is directly obtained from the priority set generated in the previous step and represents the static importance of the new task. This value is determined by the source, category, and specific action of the event. For example, a teacher's mandatory intervention instruction has a higher priority than a system-issued exam task, and the exam task has a higher priority than a student's self-initiated learning request. This value has already been calculated and bound in the previous steps and can be directly called here. In this example, the newly arrived task is a "remote mandatory intervention instruction," and its priority has been calculated in the previous steps. .

[0060] This parameter represents the priority value of the currently running task. It is also obtained from the priority set generated in the previous step and represents the static importance of the currently executing task. It reflects the importance of the current task if it is not interrupted. This is achieved by comparing it with... By comparing these parameters, a baseline score for preemptive decision-making is formed. The method for obtaining this parameter is similar to... The methods are the same; both are read directly from the task assignment record table. In this example, the currently running task is "User Self-Learning Request," and its priority has been calculated in the previous steps. .

[0061] This parameter represents the expected resource consumption of a newly arrived task. It is not a real-time measurement but is retrieved from a pre-defined "task resource profile library." This library was established during the initial system deployment phase through multiple benchmark tests on various typical tasks (such as video playback, online quiz, and application locking). It records the average CPU utilization, memory consumption (MB), and network bandwidth (Mbps) for each task under standard hardware conditions. To obtain a single quantitative indicator, the system uses a weighted summation method to calculate the overall resource consumption. The calculation formula is as follows: The weighting coefficients are set according to the importance of the resource, for example... , , For the newly arrived "remote forced intervention command" task, a database query reveals its resource profile: CPU 10%, memory 32MB, network 1Mbps. Therefore, its expected resource consumption is... .

[0062] This represents the smallest positive number to prevent division by zero anomalies, and is set to... .

[0063] This parameter represents the real-time resource usage of the currently running task. It is collected in real-time by calling the operating system's underlying performance monitoring interface. The system acquires the CPU utilization, physical memory usage, and real-time network interface traffic of the currently running task process at 1-second intervals, and uses this data for calculation. The same weighted summation formula is used to calculate a dynamic, comprehensive resource utilization value that reflects the current real-time load. For example, the currently running "User Self-Learning Request" task (playing an online video) has the following real-time resource utilization: CPU 40%, memory 256MB, network 4Mbps. Therefore, its real-time resource utilization is... .

[0064] This represents the total system resource capacity. This value is determined during system startup by detecting hardware configuration information and is calculated using a weighted summation method to obtain a comprehensive baseline resource capacity. For example, if the discipline desk is configured with a 4-core CPU (total 400% utilization), 4096MB of memory, and 100Mbps network bandwidth, then the total system resource capacity is... .

[0065] This represents an adjustment coefficient for the product of priority density and resource pressure. This coefficient is used to balance the weight of static priority differences and dynamic resource conditions in decision-making. Its value is set based on offline analysis and simulation tests of a large number of historical scheduling scenarios. By running a series of benchmark test cases covering different task combinations and system loads, optimization algorithms such as grid search are used to find the method that minimizes the average waiting time of high-priority tasks and maximizes the total system throughput. The value, after calibration, is determined to be a balanced value that performs well in most scenarios. Here, it is set. .

[0066] Calculation process:

[0067] Based on the results of obtaining the aforementioned parameters, the specific values ​​are substituted into the dynamic preemption decision score calculation formula for derivation:

[0068] , , , , , , .

[0069] ;

[0070] ;

[0071] ;

[0072] ;

[0073] ;

[0074] ;

[0075] The results show that the calculated dynamic preemption decision score is 78.525. This value integrates the static priority difference between the old and new tasks, the resource efficiency of the new task, and the current resource tension of the system. It will be used as the direct basis for judging whether to perform the preemption operation in the next step. A positive value indicates that the new task has the potential to preempt. The larger the value, the more sufficient the reason for preemption. The score of 78.525 here is a relatively high score. The main contribution comes from the basic priority difference of 70 points. At the same time, since the current system resource pressure is not high (the tanh value is small), the contribution of the dynamic adjustment term (8.525) is relatively mild.

[0076] Based on the dynamic preemption decision score calculated in the previous step, the system compares it with a preset system preemption trigger threshold. This threshold is not fixed but is determined through statistical analysis of the dynamic preemption decision score data generated by the equipment running in a standard teaching environment for one week. Specifically, all decision scores are collected, and the top 5% and bottom 5% extreme values ​​are removed to avoid noise interference. Then, the arithmetic mean and standard deviation of the remaining data are calculated. Finally, the threshold is set as the mean plus 1.5 times the standard deviation. For example, if the mean is 30 and the standard deviation is 10, then the system preemption trigger threshold is... In this example, the calculated dynamic preemption decision score of 78.525 is clearly greater than 45, therefore the preemption condition is met. The system immediately sends an interrupt signal to the operating system kernel, targeting the currently running task process. Simultaneously, a context saver is started to take a complete snapshot of the process's runtime context, including the current values ​​of all CPU registers, program counter, stack pointer, memory page table, list of open file descriptors, and key state data registered by the application itself through a specific interface (such as the current timestamp of video playback). This snapshot is serialized and stored in a reserved memory buffer. After the interrupt is triggered, the system performs an additional resource lock status check. It queries a global resource lock table, which records whether exclusive resources such as cameras, microphones, and full-screen display permissions are currently locked by a task. If a newly arrived task requests a resource (such as a forced intervention requiring full-screen display permissions), and this resource is already locked by the current task, the system will perform a check. Even if the preemption decision score is met, newly arriving task requests will be temporarily suspended, and their task IDs and requested resource information will be pushed into a first-in-first-out waiting queue. Otherwise, the system will combine the priorities and resource requirements of the current task (which has been interrupted), the newly arrived task, and all tasks in the waiting queue to construct a temporary resource allocation mapping table. This table is organized by task ID as the row and system resources (such as CPU core 0, memory segment A) as the column. The table entries are the allocation status (such as "allocated", "reserved", "released"). The table generation logic follows a strict priority order, starting with the highest priority task and satisfying its resource requirements one by one until all tasks are allocated or resources are exhausted. Finally, based on this resource allocation mapping table and the saved context snapshot, the system generates a preemptive scheduling execution scheme containing a series of atomic operation instructions. For example, the instruction sequence is: "[Save process 456 context to address 0x] -> [Release process 456 display resources] -> [Allocate display resources to process 789] -> [Load and resume the execution of process 789]".

[0077] The steps to obtain the set of matching behavior patterns are as follows:

[0078] Based on the application start-up, shutdown, and window switching events recorded by the preemptive scheduling execution scheme, the behavior label, timestamp, and priority value of the task to which the event belongs are extracted in chronological order. A time-series triplet sequence containing the event behavior label, the event occurrence time, and the priority of the associated task is constructed to obtain a continuous behavior event stream.

[0079] Based on the continuous flow of behavioral events, the behavioral label sequence is compared with the behavioral structure of the positive behavioral pattern library and the negative behavioral pattern library one by one. The sliding matching method is used to record the behavioral pattern matching position. At the same time, the corresponding priority value and behavior occurrence time are extracted from the triples to obtain the matching behavioral pattern set.

[0080] Specifically, based on the preemptive scheduling execution scheme generated in the previous step, the system starts a log parser. This parser is specifically responsible for extracting events related to user behavior analysis from the structured log records of the scheme. The preemptive scheduling execution scheme is essentially a sequence of operation instructions ordered by timestamps, which includes records such as "Start application [ID: A, Task ID: T1]", "Switch window focus to [Window handle: H, Task ID: T2]", and "Terminate application [ID: B, Task ID: T1]". The parser reads these records one by one and applies a preset event filter. This filter only allows events of type "application start / stop" and "window switching". For each event record that passes the filter, the system performs an information extraction operation. First, a standardized behavior label is parsed from the record. For example, for "Start application [ID: com.learning.math, Task ID: T_EXAM_01]", its behavior label is converted to "START_APP_LEARN", while "Switch window focus to [Window handle: GameWindow, Task ID: T1]" is converted to "START_APP_LEARN". "T_STUDY_02" is then labeled as "SWITCH_TO_NON_LEARN". The mapping rules for these labels are stored in a configurable lookup table. Next, the UNIX timestamp, accurate to milliseconds, is directly extracted from the record as the event occurrence time. Finally, based on the task ID in the record, the system will look up the task assignment record table in reverse to read the static priority value bound to that task ID. For example, the priority of task T_EXAM_01 is 80, and the priority of task T_STUDY_02 is 45. By repeating this process for all filtered events, the system converts each original operation record into a time-series triple containing the behavior label, event occurrence time, and associated task priority, for example, ("START_APP_LEARN", 1678886400000, 80). All these triples are arranged in ascending order of their timestamps to form a complete and ordered continuous flow of behavior events.

[0081] Based on the continuous stream of behavioral events generated in the previous step, the system activates the pattern matching engine to compare the user's actual behavioral sequence with pre-set positive and negative behavior pattern libraries. These libraries were jointly constructed by educational psychology experts and experienced teachers during the system deployment phase. They define various typical learning behavior sequences. The positive behavior pattern library includes patterns such as "Focused Learning" (defined as: [START_APP_LEARN, FOCUS_WINDOW_LEARN], requiring the duration of the FOCUS_WINDOW_LEARN event to exceed 600 seconds) and "Smooth Task Switching" (defined as: [STOP_APP_LEARN_A, START_APP_LEARN_B], with the time interval between the two events less than 10 seconds). The negative behavior pattern library includes "Distracted Learning" (defined as: [START_APP_LEARN, ...). Patterns such as "SWITCH_TO_NON_LEARN" and "frequent switching" (defined as more than 5 window switching events within 300 seconds) are matched using a sliding window method. The system maintains a configurable window (e.g., 5 events) on the continuous stream of behavioral events. This window slides forward along the time axis, moving one event unit at a time. At each window position, the system extracts the behavioral label subsequence within the window and performs exact or fuzzy matching with all patterns in two behavioral pattern libraries. When the behavioral label sequence within the window completely matches the pattern structure in one of the libraries, a successful match is recorded, and the system immediately records the match result. The results are stored in a temporary list, which records the matched pattern ID (e.g., positive pattern P01 or negative pattern N03), the basic behavior score of the pattern (which is also predefined in the library), the time when the behavior pattern occurred (usually recorded as the timestamp of the last event in the matching sequence), and the associated task priority (the average or maximum value of the associated priorities of all events in the matching sequence is taken; the maximum value strategy is used here). For example, if the "distracted learning" pattern is matched, its basic score is -10, the time of occurrence is 1678887000000, and the behavior occurs during the self-learning task with a priority of 45, then it is recorded as (N03, -10, 1678887000000, 45). After traversing the entire continuous stream of behavior events, this temporary list is the final set of matched behavior patterns.

[0082] The steps to obtain a user's digital health status code are as follows:

[0083] Based on a set of matching behavioral patterns, the user's digital health status code is calculated using the following formula:

[0084] ;

[0085] in, Indicates the current time User digital health status code Indicates basic health score. Indicates the first The basic behavioral score for each behavior match (positive is positive, negative is negative).

[0086] Indicates the first The priority value of the associated task when an action occurs. Indicates the task priority modulation factor. Indicates the first The time when the behavior occurs, Indicates the current calculation time. The time decay coefficient representing the effect of behavior. The total number of matching behavior patterns.

[0087] Specifically, the formula: The advantage of the formula lies in the introduction of basic health scores. This provides a stable baseline for evaluation, giving score fluctuations a clear reference point. Secondly, the task priority modulation factor... This allows the same behavior to have differentiated effects in task contexts of varying importance. For example, distraction during important online exams has a significantly amplified negative impact, while distraction during free-exploration learning has a relatively smaller impact. This makes the assessment results closer to real educational situations. Finally, the exponential decay term... The introduction of this feature gives the assessment the characteristics of "memory" but not "permanent memory". The impact weight of recent behavior is much greater than that of long-term behavior. The model can dynamically reflect the latest trend of user status changes without being overly dragged down by outdated historical behavior data. The combination of these three factors makes the calculation of user digital health status code take into account the nature of behavior, the importance of context, and the timeliness of impact.

[0088] The baseline health score is a system-level initial parameter representing the generally expected health level for a standard user. Its value is set by the system administrator based on educational goals and management strategies. This score is typically set at a relatively positive level to encourage users to maintain a good state. It serves as a floating benchmark for the user's digital health status code, and all subsequent behavioral scores are adjusted based on this. During the initial system deployment, data from 100 students of different grades was tracked for one month, and combined with assessments by educational experts, their average digital health assessment score was used as the initial baseline. The reference value was ultimately determined to be a standard value. For example, in a system with a maximum score of 100, this value could be set to 80 points as the starting score for all new users. In this example, .

[0089] Indicates the first Each behavior pattern is matched with a base behavior score, which quantifies the inherent positive or negative degree of each identified behavior pattern. The scores are derived from the aforementioned positive and negative behavior pattern databases. In constructing these databases, five educational psychology experts independently scored each behavior pattern using the Delphi method (range -20 to +20). Then, through multiple rounds of anonymous feedback and statistical analysis (calculating the mean and removing extreme outliers), a universally accepted base score was determined for each pattern. For example, the pattern "completing a full learning unit" might be assigned +10 points, while the pattern "opening an entertainment app during an exam" might be assigned -20 points. Positive scores represent positive behaviors, and negative scores represent negative behaviors; the absolute value reflects the intensity of the behavior's impact. In this example, two behaviors are selected from the matched behavior pattern set: a positive behavior "focusing on learning for 600 seconds," and... One negative behavior is "distraction from studying." .

[0090] Indicates the first The priority value of the associated task when a behavior occurs is extracted directly from the set of matching behavior patterns generated in the previous step. It reflects the importance of the specific task context when the user's behavior occurs. This priority value is a static value calculated by the system based on factors such as the source and type of the event when the task is created. It has already been bound to the task and behavior in the preceding steps. For example, a mandatory task initiated by a teacher has a priority of 115, an exam task issued by the system has a priority of 80, and student self-study has a priority of 45. In this example, the positive behavior "focus on studying for 600 seconds" occurs in an exam preparation task, and its associated task priority... The negative behavior of "distraction from learning" occurred during a task of independently browsing learning materials, and its associated task priority... .

[0091] This represents the task priority modulation factor, a key adjustment function used to amplify or reduce the influence of the base behavior score based on the importance of the task. It is a factor related to priority. An increasing function is defined as: in, and These are the system's highest and lowest preset priorities, as set above. , , and It is a moderating coefficient, which is used to find the parameter value that best fits the expert's judgment of the severity of the "behavior-context" combination through regression analysis of historical data. Controlling the intensity of modulation, The degree of nonlinearity of the modulation is controlled here. , This design aims to achieve a quadratic amplification effect on high-priority task scenarios. This setup makes the behavioral impact in low-priority tasks close to its base score, while the impact in high-priority tasks is significantly amplified.

[0092] Indicates the first The timestamp of each behavior is also directly obtained from the set of matched behavior patterns. It records the point in time when the behavior pattern was identified. In this example, the positive behavior occurred 2 hours ago, and the negative behavior occurred 30 minutes ago. The current time is set as follows. If it is 1678893600, then , .

[0093] This represents the current calculation time, i.e., the current system time for calculating the user's digital health status code, expressed as a second-level UNIX timestamp. It serves as the benchmark for time decay calculation, ensuring the real-time nature of the model evaluation. The latest timestamp is retrieved each time the calculation is called. In this example, .

[0094] This represents the time decay coefficient of a behavior's impact. This coefficient determines how quickly the influence of historical behavior on the current status code diminishes over time. Its value is based on the half-life of behavioral habits in educational theory, which generally considers the influence of a behavior to decrease significantly after 24 hours. The half-life is then set accordingly. If the time is 24 hours (i.e., 86400 seconds), then the attenuation coefficient can be calculated using the formula... The calculation shows that, This setting ensures that scores can sensitively reflect recent behavioral changes.

[0095] The total number of matched behavior patterns represents the total number of behaviors included in the calculation from the set of matched behavior patterns within this calculation period. In this example, the two most recent behaviors are considered. .

[0096] Calculation process:

[0097] First, calculate the modulation factors for each of the two behaviors. :

[0098] For positive behaviors (k=1): ;

[0099] ;

[0100] For negative behaviors (k=2): ;

[0101] ;

[0102] Then, calculate the total score:

[0103] ;

[0104] Substitute the values:

[0105] ;

[0106] ;

[0107] ;

[0108] ;

[0109] ;

[0110] ;

[0111] The results indicate that at the current moment, the user's digital health status code is 76.042, which is lower than the base health score of 80. This means that the impact of the user's recent negative behavior outweighs that of positive behavior, leading to a decrease in the health score. Although the base score of positive behavior is not low and it occurred in a high-priority task, the negative behavior occurred more recently (with less time decay) and its base negative score is larger, resulting in a decrease in the total score. This value of 76.042 will be passed to the next processing step for comparison with a multi-level score threshold table. If the score falls into the "warning" range (e.g., 70-80 points), a corresponding mild guidance strategy will be triggered.

[0112] The steps for obtaining the selected boot strategy identifier are as follows:

[0113] Based on the user's digital health status code, the threshold of each level in the multi-level score threshold table is called in turn, and the numerical difference between the user's digital health status code and the current threshold is calculated level by level. It is then determined whether the numerical difference reaches the policy triggering condition defined in the threshold table, and the policy triggering judgment result is obtained.

[0114] Based on the policy trigger judgment result, retrieve the policy entries corresponding to the trigger judgment result from the policy configuration table one by one, extract the unique identifier of the boot strategy marked in the policy entry, and generate the selected boot strategy identifier based on the unique identifier of the boot strategy.

[0115] Specifically, based on the user's digital health status code calculated in the previous step, the system activates the strategy decision engine. The core of this engine is to interact with a pre-set multi-level score threshold table. This table is loaded during system initialization, and its content is jointly formulated by education management experts based on the severity of the user's status corresponding to different score ranges. The table defines multiple health levels, such as "Excellent" (90-100 points), "Good" (80-89 points), "Warning" (70-79 points), "Dangerous" (60-69 points), and "Intervention" (below 60 points). Each level corresponds to a threshold and a strategy triggering condition. For example, the threshold for the "Warning" level is 70, and the triggering condition is "the score is less than or equal to the previous level threshold (80) and greater than the current level threshold (70)". The system receives the user's digital health status code. For example, after 76.042, the system compares the values ​​sequentially according to the threshold order from high to low, starting from 90. First, the system calculates the difference between 76.042 and 90, which is -13.958, failing to meet the trigger condition for the "Excellent" level (greater than or equal to 90). Next, the system calls the next threshold, 80, and calculates the difference, which is -3.958, failing to meet the trigger condition for the "Good" level (greater than or equal to 80). Then, the system calls the threshold 70. Since 76.042 meets the condition of "less than 80 and greater than 70", the policy trigger condition is activated, and the system determines that the current user's status belongs to the "Warning" level. This judgment result, "Warning", is output. Once a matching level is found, the loop terminates and no longer compares with lower thresholds, thus obtaining the policy trigger judgment result.

[0116] Based on the strategy trigger judgment result obtained in the previous step, such as "warning," the system then queries a pre-set strategy configuration table. This table is a detailed set of rules that maps the strategy trigger judgment result to specific guidance strategies. Its content is designed in conjunction with educational goals, user acceptability, and technical feasibility, and is completed jointly by system designers and educational consultants. Each row in the table contains three core fields: trigger condition (corresponding to the level in the multi-level score threshold table), a unique identifier for the guidance strategy, and a detailed description of the strategy. For example, the strategy configuration table might contain entries like: trigger condition "good," corresponding to the identifier "GUIDE_POS_01," described as "pop up positive feedback information to encourage continued engagement"; trigger condition "warning," corresponding to the identifier "GUIDE_WARN_01," described as "display a gentle prompt message in the corner of the screen, suggesting focus." The trigger condition is "dangerous," with the corresponding identifier "GUIDE_INTV_01," described as "briefly locking non-learning applications for 5 minutes." The trigger condition is "intervention," with the corresponding identifier "GUIDE_INTV_02," described as "adjusting network interface access permissions to only allow access to whitelisted learning websites." The system takes the "warning" result and performs a line-by-line string matching search in the policy configuration table. When a row's "trigger condition" field is found to be completely consistent with "warning," the search stops, and the value of the "unique identifier for the guidance policy," i.e., "GUIDE_WARN_01," is extracted from that row. This identifier is a globally unique string, serving as a unique index for subsequent operations. The system encapsulates this extracted string into a standard format data object and generates a selected guidance policy identifier based on the unique identifier for the guidance policy.

[0117] The steps for obtaining control commands from the terminal device are as follows:

[0118] Based on the selected boot strategy identifier, retrieve the terminal device control instruction set, call the control instruction entry corresponding to the selected boot strategy identifier, parse the operation instructions in the control instruction entry, and generate the terminal device control instructions.

[0119] Specifically, based on the selected boot strategy identifier generated in the previous step, such as "GUIDE_WARN_01", the system finally enters the instruction construction stage. The core of this stage is to query a terminal device control instruction set. This instruction set is a low-level command library tightly integrated with the hardware and operating system interfaces. It stores detailed instructions that translate abstract strategies into specific device executable operations. This instruction set is written and pre-configured by system development engineers according to the API documentation of the IoT smart multimedia integrated teaching desk. Each entry corresponds to a unique boot strategy identifier and defines in detail a series of operation instructions required to execute the strategy. For example, in the entry corresponding to the identifier "GUIDE_WARN_01", its operation instruction is defined as a JSON format string: {"action": "display_toast", "params": {"text": "Please stay focused, the learning effect will be better!", "position": "bottom_right", "duration": 5000}}, while the instruction corresponding to "GUIDE_INTV_01" is {"action": "lock_apps", "params": {"category": After receiving "GUIDE_WARN_01", the system uses it as a key to quickly search the instruction set and locate the corresponding instruction entry. Then, the system parses the JSON string in the entry, extracts the value of the "action" field "display_toast" and the "params" object. This parsed information constitutes the core content of the control command sent to the terminal device. The system formats this information according to the device communication protocol (e.g., encapsulates it into a TCP packet with a specific command header) to generate the final terminal device control command that can be directly executed by the discipline desk device.

[0120] 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 control method for an IoT-enabled smart multimedia integrated discipline desk, characterized in that, Includes the following steps: Receive raw events from the IoT smart multimedia integrated discipline desk, the raw events being remote mandatory intervention commands, online examination tasks, and user self-learning requests. Match each event identifier with a preset event priority definition table, attach a static priority value to each task request event, and generate a priority task sequence. Based on the priority task sequence, a priority set is obtained and generated. If the priority value of the newly arrived task is higher, the current task is interrupted and the context is saved to establish a preemptive scheduling execution scheme. If the priority value of the newly arrived task is not higher, it is further determined whether the requested resource is exclusively locked. If so, the task request is placed in the waiting queue. Based on the application start-up and window switching events recorded by the preemptive scheduling execution scheme, a continuous behavioral event stream is extracted. The behavioral event stream is then compared with a preset positive behavior pattern library and a negative behavior pattern library to establish a matching behavior pattern set. The matching behavior pattern set is then traversed to calculate the user's digital health status code. Based on the user's digital health status code, a corresponding strategy is selected, a selected guidance strategy identifier is generated, and a corresponding operation instruction is retrieved according to the selected guidance strategy identifier. The operation instruction is to lock non-learning applications, display prompt information on the discipline desk screen, or adjust network interface access permissions, thereby constructing terminal device control instructions.

2. The IoT-based intelligent multimedia integrated discipline desk control method according to claim 1, characterized in that, The steps for obtaining the priority task sequence are as follows: Receive raw events from the IoT smart multimedia integrated discipline desk, parse the event identifier in each raw event, read the event identifier character sequence one by one through the traversal method, and extract the feature fields representing the event type from the event identifier character sequence to form a set of event identifier feature fields; Based on the set of event identifier feature fields, each event identifier feature field is matched against a preset event priority definition table one by one, and the priority value corresponding to each event feature field in the event priority definition table is called to obtain a set of static priority values ​​that match each event identifier feature field. Based on the set of static priority values, the original events are bound one by one to their corresponding static priority values, and the original events are rearranged in descending order of priority values ​​to form a priority task sequence.

3. The IoT-based intelligent multimedia integrated discipline desk control method according to claim 1, characterized in that, The steps for obtaining the priority set are as follows: Based on the priority task sequence, the newly arrived task entries and currently running task entries in the task assignment record table are parsed, and the bound priority field values ​​are read to obtain the priority set.

4. The IoT-based intelligent multimedia integrated discipline desk control method according to claim 1, characterized in that, The steps for obtaining the preemptive scheduling execution scheme are as follows: Calculate the dynamic preemption decision score based on the priority set; Based on the dynamic preemption decision score, when the dynamic preemption decision score is greater than the system preemption trigger threshold, the currently running task is interrupted and a complete running context snapshot is written. If the resource requested by a newly arrived task is in a locked state, the task request is pushed into the waiting queue. A resource allocation mapping table is constructed by combining the current task status, the status of the newly arrived task, and the priorities of all tasks in the waiting queue. A preemptive scheduling execution scheme is generated based on the resource allocation mapping table and the context snapshot.

5. The IoT-based intelligent multimedia integrated discipline desk control method according to claim 1, characterized in that, The steps for obtaining the set of matching behavior patterns are as follows: Based on the application start / stop and window switching events recorded by the preemptive scheduling execution scheme, the behavior tag, timestamp and priority value of the task to which the event belongs are extracted in chronological order, and a time-series triplet sequence containing the event behavior tag, the event occurrence time and the priority of the associated task is constructed to obtain a continuous behavior event stream. Based on the continuous stream of behavioral events, the behavioral label sequence is compared with the behavioral structure of the positive behavioral pattern library and the negative behavioral pattern library one by one. The sliding matching method is used to record the behavioral pattern matching position. At the same time, the corresponding priority value and behavior occurrence time are extracted from the triples to obtain the matching behavioral pattern set.

6. The IoT-based intelligent multimedia integrated discipline desk control method according to claim 1, characterized in that, The steps for obtaining the user's digital health status code are as follows: The user's digital health status code is calculated based on the set of matching behavior patterns.

7. The IoT-based intelligent multimedia integrated teaching desk control method according to claim 1, characterized in that, The steps for obtaining the selected boot strategy identifier are as follows: Based on the user's digital health status code, the threshold of each level in the multi-level score threshold table is called sequentially, the numerical difference between the user's digital health status code and the current threshold is calculated level by level, and it is determined whether the numerical difference reaches the policy triggering condition defined in the threshold table to obtain the policy triggering judgment result. Based on the policy trigger judgment result, the policy entries corresponding to the trigger judgment result are retrieved item by item from the policy configuration table, the unique identifier of the guidance policy marked in the policy entry is extracted, and the selected guidance policy identifier is generated based on the unique identifier of the guidance policy.

8. The IoT-based intelligent multimedia integrated discipline desk control method according to claim 1, characterized in that, The steps for obtaining the terminal device control command are as follows: Based on the selected boot strategy identifier, the terminal device control instruction set is retrieved, the control instruction entry corresponding to the selected boot strategy identifier is called, the operation instructions in the control instruction entry are parsed, and the terminal device control instructions are generated.

9. A discipline desk control system, used to implement the Internet of Things intelligent multimedia integrated discipline desk control method according to any one of claims 1-8, characterized in that, The system includes: The event processing module receives raw events from the IoT smart multimedia integrated discipline desk. The raw events are remote mandatory intervention commands, online examination tasks, and user self-learning requests. It matches each event identifier with a preset event priority definition table, adds a static priority value to each task request event, and generates a priority task sequence. The task scheduling module obtains a priority set based on the priority task sequence and generates a priority set. If the priority value of the newly arrived task is higher, the current task is interrupted and the context is saved to establish a preemptive scheduling execution scheme. If the priority value of the newly arrived task is not higher, it continues to determine whether the requested resource is exclusively locked. If so, the task request is placed in the waiting queue. The behavior analysis module extracts a continuous stream of behavior events based on the application start-up and shutdown and window switching events recorded by the preemptive scheduling execution scheme. It then compares the behavior event stream with a pre-set positive behavior pattern library and a negative behavior pattern library to establish a matching behavior pattern set. Finally, it traverses the matching behavior pattern set to calculate the user's digital health status code. The strategy execution module selects the corresponding strategy based on the user's digital health status code, generates a selected guidance strategy identifier, and retrieves the corresponding operation instructions based on the selected guidance strategy identifier. The operation instructions are to lock non-learning applications, display prompt information on the disciplinary desk screen, or adjust network interface access permissions, thereby constructing terminal device control instructions.

Images