Medical manikin real-time feedback teaching method and system

Abstract

The application discloses a medical simulation person real-time feedback teaching method and system, relates to the technical field of medical education, and discloses the medical simulation person real-time feedback teaching method and system, multi-source behavior data generated by a sensor group collecting the operation of a medical simulation person of a student is collected, feature extraction is carried out, and decision deviation data is generated by real-time comparison; when the deviation exceeds a threshold value, an intervention instruction is triggered and teaching intervention is executed, operation deviation can be captured in time and targeted guidance can be dynamically generated, multi-source behavior data is collected and analyzed in real time, and comparison with a standard clinical decision path can be carried out, operation deviation of a student can be detected in time, teaching intervention can be automatically triggered when the deviation is too large, the timeliness and accuracy of medical teaching feedback are improved, error operation is avoided to be solidified, and training efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of medical education technology, and in particular to a real-time feedback teaching method and system using medical mannequins. Background Technology

[0002] In medical clinical skills training, the timeliness and accuracy of trainee feedback directly affect skill mastery. Traditional teaching models rely on manual observation and post-training feedback from teachers, leading to prolonged feedback cycles and the potential for trainees' errors to become ingrained. For example, in cardiopulmonary resuscitation (CPR) simulation training, insufficient chest compression depth or frequency deviations, if not pointed out immediately, may lead to repeated reinforcement of incorrect movement patterns. In endotracheal intubation, incorrect sequence of steps or omission of key equipment often require multiple reminders from the teacher for correction, resulting in low training efficiency. A deeper problem lies in the lack of multi-dimensional quantitative monitoring capabilities for operational behaviors in existing teaching systems. They cannot accurately assess the standardization of voice commands, the compliance of the timing of operational actions, or the degree of visual attention focus. Although some institutions have introduced video monitoring systems for offline analysis, these systems can only process two-dimensional planar image information and are ill-suited to the multi-source data fusion requirements of complex scenarios. In high-fidelity training environments such as emergency room simulations, heterogeneous data such as semantic understanding of voice commands, time interval control of operational steps, and correlation analysis of eye movements and key equipment cannot be processed synchronously, resulting in significant delays in teaching intervention. Existing technical solutions cannot capture subtle deviations in trainees' decision-making process in real time, nor can they dynamically generate targeted guidance based on the degree of deviation, making it difficult to meet the core requirements of modern medical education for precise and timely teaching feedback.

[0003] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide a real-time feedback teaching method and system using medical mannequins, which aims to improve the timeliness and accuracy of medical teaching feedback and enhance training efficiency.

[0005] To achieve the above objectives, this application proposes a real-time feedback teaching method using a medical mannequin, the method comprising:

[0006] The sensor array collects multi-source behavioral data generated by the trainee operating the medical mannequin. The multi-source behavioral data includes voice data, operation timing data, and eye movement trajectory data.

[0007] The multi-source behavioral data is subjected to feature extraction processing to obtain a behavioral feature dataset;

[0008] The behavioral feature dataset is compared with the standard clinical decision-making path in real time to generate decision deviation data.

[0009] When the decision deviation data exceeds a preset threshold, the intervention instruction generation process is triggered, and real-time intervention instruction data is output.

[0010] The real-time intervention instruction data is sent to the feedback execution terminal to perform real-time teaching intervention operations.

[0011] In one embodiment, the step of performing real-time comparison processing between the behavioral feature dataset and a standard clinical decision-making path to generate decision deviation data includes:

[0012] Extract speech keyword data, operation step sequence data, and visual attention area data from the behavioral feature dataset;

[0013] The voice keyword data is matched with the medical terminology database of the current scene to generate instruction matching degree data.

[0014] The sequence data of the operation steps is compared with the operation sequence of the standard clinical decision-making path in terms of time sequence alignment to generate time sequence deviation data.

[0015] The visual attention area data is analyzed and processed to perform spatial overlap analysis with the preset key equipment area to generate attention missing data.

[0016] The instruction matching data, the timing deviation data, and the missing attention data are weighted and fused to generate decision deviation data.

[0017] In one embodiment, the step of performing time-series alignment and comparison processing on the sequence data of the operation steps and the operation sequence of the standard clinical decision-making pathway to generate time-series deviation data includes:

[0018] Identify the start and end timestamps of each operation action in the operation step sequence data;

[0019] Calculate the time interval data between adjacent operations based on the start timestamp and end timestamp;

[0020] The time interval data is compared with the allowed time window of the standard clinical decision-making pathway.

[0021] The number of operation steps exceeding the allowed time window is counted, and time series deviation data is generated.

[0022] In one embodiment, the step of performing spatial overlap analysis on the visual attention area data and the preset key equipment area to generate attention-missing data includes:

[0023] Mark the 3D region boundary data of key equipment in the spatial coordinate system of the medical simulator;

[0024] The coordinate position relationship between the gaze point coordinates in the visual attention area data and the three-dimensional region boundary data is determined.

[0025] Calculate the percentage of gaze points that are not covered by key equipment within a preset time period to generate attention gap data.

[0026] In one embodiment, the step of performing weighted fusion processing on the instruction matching degree data, the timing deviation data, and the attention missing data to generate decision deviation degree data includes:

[0027] Assign a first weight value to the instruction matching data, assign a second weight value to the timing deviation data, and assign a third weight value to the missing data of concern;

[0028] The weighted instruction matching data, weighted timing deviation data, and weighted attention missing data are input into the deviation calculation function for processing;

[0029] The weighted data is standardized and scored using the deviation calculation function to generate decision deviation data.

[0030] In one embodiment, the method further includes:

[0031] Parse the pre-set medical guideline document to obtain operational process node data;

[0032] Based on the operational process node data, an initial clinical decision path is generated;

[0033] The system receives innovative operational data from trainees that has been verified by experts, and dynamically updates the initial clinical decision-making path based on this data to obtain an updated standard clinical decision-making path.

[0034] In one embodiment, when the decision deviation data exceeds a preset threshold, the step of triggering the intervention instruction generation process and outputting real-time intervention instruction data includes:

[0035] Identify the error type data with the largest deviation value in the decision deviation data;

[0036] Extract correction scheme template data corresponding to the error type data from the knowledge graph engine;

[0037] The correction scheme template data is converted into an executable instruction format to generate real-time intervention instruction data.

[0038] In one embodiment, the step of converting the correction scheme template data into an executable instruction format to generate real-time intervention instruction data includes:

[0039] When the error type data is an operation sequence error, standard process animation data containing step sequence number markers is generated;

[0040] When the error type data is a device operation omission, visual highlight data pointing to a specific location on the medical mannequin is generated;

[0041] The standard process animation data or visual highlight data is encapsulated into augmented reality display instruction data to generate real-time intervention instruction data.

[0042] In one embodiment, the feedback execution terminal includes a head-mounted augmented reality device and / or a feedback control unit integrated into a medical mannequin; the execution of real-time teaching intervention operations includes:

[0043] Operation instructions are displayed via the aforementioned head-mounted augmented reality device;

[0044] And / or, through the feedback control unit integrated into the medical mannequin, an operation guidance layer is overlaid on the vital signs display of the medical mannequin;

[0045] And / or, through the feedback control unit integrated into the medical simulator, a directional voice prompt is initiated when the student's operation exceeds a preset time.

[0046] Furthermore, to achieve the above objectives, this application also proposes a real-time feedback teaching system for medical mannequins, which includes: a memory, a processor, and a real-time feedback teaching program for medical mannequins stored in the memory and executable on the processor, wherein the real-time feedback teaching program for medical mannequins is configured to implement the steps of the real-time feedback teaching method for medical mannequins.

[0047] The real-time feedback teaching method and system using medical mannequins proposed in this application collects multi-source behavioral data generated by trainees operating the medical mannequins through a sensor array, extracts features, and compares them in real time to generate decision deviation data. When the deviation exceeds a threshold, an intervention command is triggered and a teaching intervention is executed. This method can instantly capture operational deviations and dynamically generate targeted guidance. By collecting and analyzing multi-source behavioral data in real time and comparing it with standard clinical decision-making paths, it can instantly detect trainees' operational deviations and automatically trigger teaching interventions when the deviations are too large. This improves the timeliness and accuracy of medical teaching feedback, prevents erroneous operations from being solidified, and enhances training efficiency. Attached Figure Description

[0048] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating an embodiment of the real-time feedback teaching method for medical mannequins in this application.

[0051] Figure 2 This is a structural schematic diagram of an embodiment of the real-time feedback teaching system for medical simulators provided in this application.

[0052] Explanation of icon numbers:

[0053] 10. Memory; 20. Processor.

[0054] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0055] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of this application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0056] It should be understood that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0057] In existing medical teaching practices, student feedback is often delayed, error correction is inefficient, and the teaching is susceptible to subjective factors. Current teaching models lack the ability to quantitatively assess the timing of procedures, and existing offline assessment systems struggle to achieve real-time acquisition and dynamic feedback of multi-dimensional operational data. This is particularly problematic in complex clinical scenario simulations, where the inability to simultaneously process heterogeneous data from multiple sources leads to insufficient timeliness and accuracy in teaching interventions.

[0058] Based on this, the embodiments of this application provide a real-time feedback teaching method using a medical mannequin, referring to... Figure 1 The real-time feedback teaching method using a medical simulator includes steps S100 to S500, wherein:

[0059] Step S100: Collect multi-source behavioral data generated by the trainee operating the medical mannequin through the sensor group. The multi-source behavioral data includes voice data, operation timing data and eye movement trajectory data.

[0060] Step S200: Perform feature extraction processing on the multi-source behavioral data to obtain a behavioral feature dataset;

[0061] Step S300: The behavioral feature dataset is compared with the standard clinical decision-making path in real time to generate decision deviation data;

[0062] Step S400: When the decision deviation data exceeds a preset threshold, the intervention instruction generation process is triggered, and real-time intervention instruction data is output.

[0063] Step S500: The real-time intervention instruction data is sent to the feedback execution terminal to perform real-time teaching intervention operations.

[0064] In this embodiment, the sensor array is a collection of devices used to collect various physiological and behavioral signals generated by the trainee during the operation of the medical mannequin. It may include microphones, inertial measurement units, and eye trackers to achieve comprehensive perception of the trainee's operational behavior. Multi-source behavioral data refers to various types of data collected by the sensor array that reflect the trainee's operational process. Voice data can refer to the verbal commands and communication content issued by the trainee during operation; operation timing data can refer to the sequence of actions, duration, and intervals between actions performed by the trainee on the medical mannequin; eye-tracking data can refer to the movement path, fixation area, and fixation duration of the trainee's gaze focus point during operation. The behavioral feature dataset refers to the structured data set obtained after preprocessing and feature extraction of the original multi-source behavioral data. This dataset contains key information extracted from voice, operational, and eye-tracking data for subsequent analysis and comparison.

[0065] In this embodiment, the standard clinical decision-making path is a pre-set and validated operational procedure, decision logic, and expected outcome that conforms to regulations, tailored to a specific medical operation or clinical situation. It may include a series of chronologically arranged operational steps, key decision points, and allowed time windows. Decision deviation data is a quantitative indicator of the degree of difference between the trainee's actual operational behavior and the standard clinical decision-making path. This data reflects the extent to which the trainee deviates from the standard procedure during the operation. A preset threshold is used to determine whether the decision deviation data reaches a critical value that requires triggering intervention. When the decision deviation data exceeds this threshold, the system will consider the trainee's operation to have a significant deviation and require real-time intervention.

[0066] In this embodiment, the intervention instruction generation process automatically analyzes error types and generates corresponding correction plans based on decision deviation data, aiming to provide trainees with timely and accurate guidance. Real-time intervention instruction data is the specific instruction information output by the intervention instruction generation process, used to guide trainees in correcting erroneous operations. This data can include various forms such as text prompts, voice prompts, visual highlights, and animation demonstrations. The feedback execution terminal is the device that receives real-time intervention instruction data and presents it to the trainee. It can include head-mounted augmented reality devices, displays or speakers integrated into medical mannequins, tablet computers, etc., to achieve various forms of teaching intervention. Real-time teaching intervention operation refers to the behavior of the feedback execution terminal providing immediate feedback and guidance to trainees based on real-time intervention instruction data. This operation aims to help trainees promptly identify and correct errors during the operation process.

[0067] In this embodiment, the real-time feedback teaching method using a medical mannequin first collects multi-source behavioral data generated by the trainee's operation of the medical mannequin through a sensor array. This sensor array may include a microphone, an inertial measurement unit (IMU), and an eye tracker. For example, the microphone can be used to collect the trainee's voice data, the IMU can be used to record the trainee's movements and postures, and the eye tracker can acquire the trainee's eye movement trajectory data. As one implementation, the sensor array may contain only one or a few sensors; for example, using only a microphone and a camera to collect voice and overall motion data.

[0068] Secondly, feature extraction processing is performed on the multi-source behavioral data to obtain a behavioral feature dataset. This process aims to identify events or patterns with specific meanings from raw, continuous multi-source data. For example, spectral analysis can be performed on speech data to identify changes in sound intensity, action segmentation can be performed on temporal data to identify the start and end of actions, and cluster analysis can be performed on eye-tracking data to identify fixation points. As one implementation method, feature extraction processing can employ basic statistical methods, such as calculating the average volume of speech, the total duration of actions, or the average velocity of eye-tracking trajectories, to form an initial behavioral feature dataset.

[0069] Next, this behavioral feature dataset is compared in real time with a standard clinical decision-making pathway to generate decision deviation data. This comparison aims to quantify the differences between the trainee's actual actions and the standard procedure. For example, the trainee's action sequence can be matched one by one with the expected action sequence in the standard pathway, or the trainee's voice commands can be compared with preset correct commands using text similarity calculation. As one implementation method, the comparison can use a basic counting approach, such as counting the number of standard steps that the trainee did not perform or performed incorrectly, and using this as the decision deviation data.

[0070] Subsequently, when the decision deviation data exceeds a preset threshold, the intervention instruction generation process is triggered, outputting real-time intervention instruction data. This preset threshold can be set according to teaching objectives and error tolerance. For example, when the decision deviation data reaches a certain value, the system determines that the student's operation has a significant deviation and immediate intervention is required. As one implementation method, the intervention instruction generation process can select a general prompt from a general feedback list based on the magnitude of the deviation data, such as "Please check your operation steps" or "Please pay attention to operation details," and output it as real-time intervention instruction data.

[0071] Finally, the real-time intervention instruction data is sent to the feedback execution terminal to perform the real-time teaching intervention. The feedback execution terminal is a device that presents the intervention instruction to the learner. For example, the instruction data can be sent to a display screen where the prompt information is displayed in text form, or a pre-recorded voice prompt can be played through a speaker. As one implementation, the feedback execution terminal can be an indicator light that illuminates to alert the learner when the intervention instruction data is received.

[0072] In this embodiment, the real-time feedback teaching method using medical mannequins collects multi-source behavioral data generated by trainees operating the medical mannequins in real time and compares it with standard clinical decision-making paths, enabling timely detection of deviations in trainee operations. Therefore, this method overcomes the problems of delayed feedback and low error correction efficiency in traditional teaching, achieving quantitative evaluation of the trainee's operational process. Especially in complex clinical scenario simulations, this method can simultaneously process multi-source heterogeneous data such as speech, operation timing, and eye movement trajectories, thereby providing immediate and accurate teaching interventions and improving the timeliness and accuracy of medical skills training.

[0073] In one feasible implementation, the step of comparing the behavioral feature dataset with a standard clinical decision-making path in real time to generate decision deviation data includes: extracting voice keyword data, operation step sequence data, and visual attention area data from the behavioral feature dataset; performing terminology matching processing on the voice keyword data with a medical terminology database of the current scene to generate instruction matching data; performing temporal alignment comparison processing on the operation step sequence data and the operation sequence of the standard clinical decision-making path to generate temporal deviation data; performing spatial position overlap analysis processing on the visual attention area data and a preset key equipment area to generate attention missing data; and performing weighted fusion processing on the instruction matching data, the temporal deviation data, and the attention missing data to generate decision deviation data.

[0074] In this embodiment, speech keyword data, operation step sequence data, and visual attention area data are extracted from the behavioral feature dataset to perform fine-grained decomposition of multi-source behavioral data generated by trainees during their operation on the medical mannequin. Speech keyword data is extracted by performing speech recognition and semantic analysis on the trainees' speech data to identify key medical terms, operational instructions, or diagnostic descriptions in their spoken expressions. This helps assess the accuracy of the trainees' language expression and their mastery of professional terminology. Operation step sequence data is obtained by performing action recognition and sequence analysis on the operation time sequence data to identify the discrete operational actions performed by the trainees on the mannequin and their order of occurrence. This reflects the standardization of the trainees' operation process and the accuracy of their step execution. Visual attention area data is obtained by processing eye movement trajectory data to analyze the trainees' fixation points, fixation durations, and saccade paths, thereby determining their visual focus areas on the medical mannequin or in the surrounding environment. This reveals the trainees' attention allocation, information acquisition, and observation of key areas.

[0075] In this embodiment, after acquiring the aforementioned fine-grained data, the voice keyword data is matched with a medical terminology database for the current scenario to generate instruction matching data. This process compares the keywords spoken by the learner with preset standard medical terms, using techniques such as string matching, semantic similarity calculation, or ontology mapping to quantify the degree of conformity between the learner's voice instructions and the standard terms. Lower instruction matching data may indicate that the learner is unfamiliar with professional terminology or expresses it inaccurately.

[0076] Simultaneously, in this embodiment, the sequence data of operational steps is compared with the operational order of the standard clinical decision-making pathway through temporal alignment to generate temporal deviation data. This step uses sequence alignment algorithms (such as Dynamic Time Warping (DTW), Hidden Markov Models (HMM), or other sequence similarity algorithms) to evaluate the similarity between the trainee's actual operational sequence and the prescribed operational order in the standard clinical decision-making pathway. The temporal deviation data can quantify errors in the trainee's operational process, such as missing steps, reversed order, or redundant operations, thereby reflecting the trainee's standardization and efficiency.

[0077] Furthermore, in this embodiment, the visual attention area data is spatially overlapped with a preset key equipment area to generate attention gap data. This processing determines the student's level of attention to key information by analyzing whether the student's gaze points cover key equipment on the medical mannequin (such as monitors, infusion pumps, drug labels, etc.), as well as the duration and frequency of such coverage. Attention gap data can quantify the student's deficiencies in observation and information acquisition, such as neglecting important vital signs or omitting drug labels.

[0078] Finally, in this embodiment, the instruction matching data, temporal deviation data, and attention gap data are weighted and fused to generate decision deviation data. Considering that different types of deviations may have varying importance in clinical decision-making, a preset weight value can be assigned to each type of deviation data. For example, incorrect operation sequence may have a higher weight than inaccurate speech expression. Through weighted summation, multifactor analysis models, or other fusion algorithms, these weighted data are integrated into a comprehensive decision deviation dataset. This data can comprehensively and quantitatively reflect the degree of deviation between the learner's overall performance in speech, operation, and visual attention and the standard clinical decision-making path.

[0079] In this embodiment, the multi-source behavioral data generated by trainees operating the medical mannequin is refined into voice keyword data, operation step sequence data, and visual attention area data through the above technical solution. These are then compared with standards to generate instruction matching data, temporal deviation data, and attention gap data. This multi-dimensional and refined deviation analysis overcomes the limitations of only general comparisons, accurately identifying specific problems trainees encounter in verbal expression, operational procedures, and attention allocation. By weighted fusion of these deviation data, a comprehensive and quantitative decision deviation data can be generated. This data not only reflects the degree of deviation but also implicitly contains the type of deviation, thus providing more accurate and targeted input for the subsequent real-time intervention instruction generation process. This allows teaching interventions to directly address trainees' weaknesses, significantly improving the effectiveness of teaching feedback and trainees' learning efficiency.

[0080] In one feasible implementation, the step of performing time-series alignment comparison processing between the operation step sequence data and the operation sequence of the standard clinical decision-making path to generate time-series deviation data includes: identifying the start and end timestamps of each operation action in the operation step sequence data; calculating the time interval data between adjacent operation actions based on the start and end timestamps; comparing the time interval data with the allowed time window of the standard clinical decision-making path; and counting the number of operation steps that exceed the allowed time window to generate time-series deviation data.

[0081] In this embodiment, the step of identifying the start and end timestamps of each operation in the sequence data of operation steps aims to precisely define the start and end times of the student's execution of each specific operation. The system analyzes the operation timing data collected by the sensor array to identify event signals related to specific operations. For example, when a student picks up a syringe, presses a button, or moves a specific device, these physical interaction events are recorded with precise timestamps. Using a preset action recognition algorithm, the start and end timestamps of each operation can be extracted from these continuous event streams, thus providing basic data for subsequent time analysis.

[0082] In this embodiment, when calculating the time interval data between adjacent operations based on the start and end timestamps, after identifying the precise time boundaries of each operation, this step is used to quantify the pause or transition time between consecutive operations. Specifically, for any two adjacent operations in the operation step sequence data, such as operation A and operation B, the system calculates the time difference between the start timestamp of operation B and the end timestamp of operation A. This time difference is the time interval data between adjacent operations, which intuitively reflects the thinking, preparation, or hesitation time spent by the student after completing one operation and before starting the next operation. It is a key indicator for evaluating the smoothness and efficiency of operation.

[0083] In this embodiment, the step of comparing the time interval data with the allowed time window of the standard clinical decision-making path aims to assess whether the efficiency and fluency of the trainee's operation meet preset professional standards. The standard clinical decision-making path not only specifies the correct order of operations but also typically includes provisions for the maximum or minimum allowed time interval between key operations, i.e., the allowed time window. The system compares the time interval data calculated by the trainee during actual operation with the corresponding allowed time window in the standard path. For example, if the standard specifies that the time interval between two operations should be between 5 and 10 seconds, the system determines whether the trainee's actual interval falls within this range.

[0084] In this embodiment, when generating timing deviation data by counting the number of operation steps exceeding the allowed time window, this step quantifies operations that do not meet the time standard after comparing the time intervals with the allowed time window. The system iterates through all calculated time interval data and counts each interval that exceeds the standard allowed time window (whether too long or too short). The final count of operation steps exceeding the allowed time window is the timing deviation data. This data directly reflects the severity of problems students have in terms of operational fluency, efficiency, or reaction speed, providing a clear quantitative basis for subsequent teaching interventions.

[0085] In this embodiment, the present application overcomes the limitations of relying solely on the order of operations by employing the above technical solution. By accurately identifying the start and end timestamps of each operation and calculating the time intervals between adjacent operations, the system can deeply analyze the trainee's time efficiency and fluency during the operation. Furthermore, by comparing these time interval data with the preset allowable time windows in the standard clinical decision-making path and counting the number of operation steps exceeding the allowable time windows, more refined and comprehensive timing deviation data is generated. This method not only identifies errors in the trainee's operation sequence but also reveals problems in operation rhythm, reaction speed, or decision hesitation, such as pausing too long between key steps or operating too slowly in emergency situations. Therefore, the present application can provide more targeted real-time feedback to help trainees correct their operating habits, improve their operational proficiency and clinical efficiency, thereby significantly improving the effectiveness of medical simulation teaching.

[0086] In one feasible implementation, the step of performing spatial position overlap analysis on the visual attention area data and the preset key equipment area to generate attention-deficient data includes: marking the three-dimensional region boundary data of the key equipment in the spatial coordinate system of the medical simulator; determining the coordinate position relationship between the gaze point coordinates in the visual attention area data and the three-dimensional region boundary data; calculating the proportion of gaze points that are not covered by the key equipment within a preset time period to generate attention-deficient data.

[0087] In this embodiment, the three-dimensional boundary data of key equipment is first annotated in the spatial coordinate system of the medical mannequin. This three-dimensional boundary data is used to precisely define the specific location and geometry of all key medical equipment or important anatomical structures on the medical mannequin in three-dimensional space. For example, precise three-dimensional boundary information (such as bounding boxes, mesh models, or point cloud data) of key equipment or areas such as syringes, stethoscopes, monitor screens, and specific blood vessels or organs can be entered into the system using computer-aided design (CAD) models, 3D scanning technology, or manual annotation by experts, and then unified into the global coordinate system of the medical mannequin. This data provides an objective reference standard for subsequently determining whether the trainee's line of sight is on the key area.

[0088] Secondly, in this embodiment, the coordinate relationship between the gaze point coordinates in the visual attention area data and the aforementioned three-dimensional region boundary data is determined. While the learner is performing the operation, their gaze point coordinates are collected in real time using an eye-tracking device; these coordinates are typically points in three-dimensional space. The system performs a geometric comparison between each gaze point coordinate and the three-dimensional region boundary data of preset key devices to determine whether the gaze point falls within the boundary of any key device. For example, if the gaze point coordinates are within the bounding box of a key device, it is considered that the learner's line of sight covers that key device. This step aims to identify whether the learner is focusing their attention on the key device or area required for operation at a specific moment.

[0089] Finally, in this embodiment, the proportion of gaze points not covering key equipment within a preset time period is calculated, and this is used to generate attention gap data. After completing the matching judgment between gaze points and key equipment areas, the system will count the proportion of all gaze points that fail to fall within the 3D area boundary data of any key equipment within a preset time period (e.g., the duration of a certain operation step, or a fixed time window, such as 5 seconds or 10 seconds). This proportion directly reflects the degree of lack of attention the learner pays to key equipment within a specific time period. For example, if 30% of the learner's gaze points do not fall on any key equipment in an operation step, then this 30% is attention gap data, indicating that the learner may be at risk of distraction or missing key operations.

[0090] In this embodiment, through the above technical solution, this application can accurately identify whether trainees have focused on important medical equipment or areas during operation by precisely annotating the three-dimensional boundary data of key equipment in the spatial coordinate system of the medical simulator and combining the coordinates of gaze points in the visual attention area data to determine the coordinate position relationship. Furthermore, by calculating the proportion of gaze points not covering key equipment within a preset time period, the degree of trainees' lack of attention to key equipment can be quantified, thereby generating objective attention deficit data. This allows the system to more accurately assess trainees' visual attention allocation, effectively identify potential oversights or omissions in key operational steps, provide specific and quantitative evidence for subsequent real-time teaching intervention, and significantly improve the pertinence and effectiveness of teaching feedback.

[0091] In one feasible implementation, the step of weighted fusion processing of the instruction matching data, the time-series deviation data, and the missing attention data to generate decision deviation data includes: assigning a first weight value to the instruction matching data, assigning a second weight value to the time-series deviation data, and assigning a third weight value to the missing attention data; inputting the weighted instruction matching data, the weighted time-series deviation data, and the weighted missing attention data into a deviation calculation function for processing; and performing standardized scoring calculation on the weighted data through the deviation calculation function to generate decision deviation data.

[0092] In this embodiment, different types of operational deviations have varying degrees of impact on the final clinical outcome in medical simulation teaching. For example, in emergency rescue scenarios, the accuracy of the timing of operations may be far more critical than the wording of verbal instructions. Therefore, a first weight value, a second weight value, and a third weight value are assigned to instruction matching data, timing deviation data, and attention gap data, respectively, to assign different priorities based on their importance in specific clinical situations or their impact on teaching objectives. These weight values ​​can be pre-set by medical experts based on clinical guidelines, teaching experience, or risk assessments of specific simulation scenarios. For example, in cardiopulmonary resuscitation simulations, the weight of timing deviation can be significantly higher than that of instruction matching or attention gaps to emphasize the accuracy of the operational rhythm. Furthermore, these weight values ​​can also be dynamically adjusted according to the learner's learning stage or training focus to achieve personalized teaching.

[0093] In this embodiment, after weighting the instruction matching data, timing deviation data, and attention gap data separately, it is necessary to integrate these weighted data into a unified decision deviation index. To this end, the weighted instruction matching data, weighted timing deviation data, and weighted attention gap data are input into a preset deviation calculation function for processing. This function combines multiple weighted input variables into a single output value to quantify the overall decision deviation of the learner. This calculation function can be implemented using various mathematical models. For example, it can be a simple weighted summation function, i.e., "Decision Deviation = W1 × Instruction Matching + W2 × Timing Deviation + W3 × Attention Gap," where W1, W2, and W3 are the corresponding weight values. Alternatively, it can be a more complex nonlinear function to better simulate the interaction between different deviation types.

[0094] In this embodiment, to ensure the comparability and interpretability of the generated decision deviation data, it is necessary to standardize the weighted data using a deviation calculation function. Standardization transforms the raw weighted data, which may have different dimensions and ranges, into a unified, pre-defined scoring range (e.g., 0 to 100 or 0 to 1), allowing the decision deviation data to intuitively reflect the degree of deviation between trainees' performance and the standard clinical decision-making pathway. For example, a min-maximum normalization method can be used to map the calculated raw deviation score to a 0-100 percentage system, where 0 represents complete compliance with the standard and 100 represents the maximum deviation. This standardized scoring not only facilitates rapid assessment of trainee performance by instructors but also provides a unified benchmark for setting intervention thresholds and conducting long-term performance tracking.

[0095] In this embodiment, through the aforementioned technical solution, the system can assign different weight values ​​to different types of behavioral deviations (instruction matching data, temporal deviation data, and attention gap data) based on their actual importance in clinical practice or their impact on teaching objectives. This weighted processing mechanism enables the decision deviation data to more accurately and precisely reflect the key problems existing in the trainees' operations, avoiding the distortion of deviation information caused by simple fusion. For example, in emergency situations, temporal deviations may be given higher weights, thus appearing more significantly in the decision deviation data, prompting the system to prioritize intervention for this problem. Subsequently, the weighted data is standardized and scored using a deviation calculation function, ensuring that the final generated decision deviation data is a unified, quantifiable, and easily understood indicator. This not only improves the clinical relevance and accuracy of decision deviation assessment but also allows real-time teaching interventions to focus more on the key aspects that trainees most need to improve, thereby significantly enhancing the pertinence and effectiveness of medical simulation teaching and accelerating trainees' mastery of clinical skills.

[0096] In one feasible implementation, the method further includes: parsing a preset medical guideline document to obtain operational process node data; generating an initial clinical decision path based on the operational process node data; receiving student innovative operational data verified by experts, and dynamically updating the initial clinical decision path according to the student innovative operational data to obtain an updated standard clinical decision path.

[0097] In this embodiment, parsing a pre-defined medical guideline document to obtain operational flow node data refers to the system automatically identifying and extracting key operational steps, decision points, conditional judgments, and corresponding execution sequences from unstructured or semi-structured medical guideline documents (e.g., PDF format clinical pathway manuals, Word document treatment guidelines, or web page-based expert consensus) using natural language processing (NLP) techniques such as text mining, information extraction, or named entity recognition. This information is then structured into discrete operational flow node data; for example, each operational step can be defined as a node, with attributes including operation name, required tools, and expected results.

[0098] In this embodiment, generating an initial clinical decision path based on the operational process node data refers to organizing and connecting the discrete operational process node data obtained from the above analysis according to their inherent logical relationships and temporal dependencies, thereby constructing a logically coherent and executable initial flowchart or sequence. This can be achieved using graph theory algorithms, treating each operational process node as a vertex in the graph, and the sequential relationships or conditional branches between nodes as directed edges, thus forming a directed acyclic graph (DAG) or finite state machine model, which serves as a preliminary standard for evaluating trainees' operations.

[0099] In this embodiment, the system receives student innovative operation data verified by experts and dynamically updates the initial clinical decision-making path based on this data to obtain an updated standard clinical decision-making path. This means that the system records the student's behavioral data during medical simulation operations. When a student's operation sequence differs from the initial clinical decision-making path, but the operation result is good and clinically reasonable, this operation data is marked as "student innovative operation data." This innovative operation data does not directly update the path but is first submitted to medical experts for review and verification. Experts judge the effectiveness, safety, and universality of the innovative operation through manual review or in conjunction with auxiliary assessment tools. Once verified by experts, the system modifies, supplements, or optimizes the initial clinical decision-making path based on this approved student innovative operation data, using methods such as incremental learning, reinforcement learning, or rule-based expert systems. For example, new operation nodes can be added, the order of existing nodes can be adjusted, and the conditions or weights of decision branches can be modified to generate a more complete, more clinically relevant, and forward-looking updated standard clinical decision-making path.

[0100] In this embodiment, through the above technical solution, this application can automatically extract operational process information from authoritative medical guidelines and construct an initial clinical decision-making path, thereby providing trainees with an objective and standardized evaluation benchmark. More importantly, this application introduces an expert verification mechanism and can receive and integrate trainees' innovative operational data approved by experts, dynamically updating the initial clinical decision-making path. This makes the standard clinical decision-making path no longer static but able to reflect the latest medical advancements and expert experience in real time, effectively solving the problems of lagging standard path updates and inability to adapt to complex and changing clinical situations in traditional methods. During the learning process, trainees can not only follow the norms but also gain recognition through innovative operations, thereby promoting the cultivation of critical thinking and clinical adaptability, and improving the accuracy, timeliness, and practicality of teaching feedback.

[0101] In one feasible implementation, when the decision deviation data exceeds a preset threshold, the step of triggering the intervention instruction generation process and outputting real-time intervention instruction data includes: identifying the error type data with the largest deviation value in the decision deviation data; extracting the correction scheme template data corresponding to the error type data from the knowledge graph engine; and converting the correction scheme template data into an executable instruction format to generate real-time intervention instruction data.

[0102] In this embodiment, identifying the error type with the largest deviation value in the decision deviation data refers to analyzing each sub-item that constitutes the decision deviation data to determine the specific error category that causes the highest overall deviation. For example, if the decision deviation data is a weighted fusion of instruction matching data, timing deviation data, and missing attention data, the system will compare the deviation degree of each of these three types of data or their weighted contribution value to identify the most prominent error type. For example, if timing deviation data contributes the most to the total deviation, the identified error type is "operation sequence error"; if missing attention data contributes the most, the identified error type is "device operation omission".

[0103] In this embodiment, extracting correction scheme template data corresponding to the error type from the knowledge graph engine refers to using a pre-built knowledge graph engine to obtain standard correction guidance for specific error types. The knowledge graph engine is a structured knowledge base that stores a large amount of medical teaching knowledge, including various operational error types and their corresponding standard correction methods, demonstration procedures, or key tips. When the system identifies specific error type data, it sends a query request to the knowledge graph engine. The engine, based on its internal relationships, retrieves and returns the correction scheme template data that best matches the error type. These templates can be text descriptions, step lists, references to animation sequences, or links to specific teaching resources. For example, for "operation sequence error," the knowledge graph engine might provide a template of a standard operating procedure animation; for "equipment operation omission," it might provide a template highlighting the location of key equipment.

[0104] In this embodiment, converting correction scheme template data into executable instruction format to generate real-time intervention instruction data means transforming the abstract correction scheme obtained from the knowledge graph engine into specific operational instructions that the feedback execution terminal can directly understand and execute. Correction scheme template data is typically high-level teaching content, requiring parsing and conversion by an instruction generation module. This module maps the semantic information in the template into specific device control commands or display instructions based on different error types and the capabilities of the feedback execution terminal (such as a head-mounted augmented reality device or a feedback control unit integrated into a medical mannequin). For example, if the template is "display standard operating procedure animation," the instruction generation module will convert it into a 3D model or video stream instruction that the AR device can render, specifying the display position and timing; if the template is "highlight a specific device," it will convert it into graphic overlay instructions for the AR device or mannequin display, including color, flashing frequency, target coordinates, etc. These converted instruction data are ultimately encapsulated into real-time intervention instruction data and sent to the feedback execution terminal to execute real-time teaching intervention operations.

[0105] In this embodiment, through the above technical solution, this application can identify the error type data with the largest deviation value in the decision deviation data, and based on this, extract targeted correction scheme template data from the knowledge graph engine, and then convert it into an executable instruction format to generate real-time intervention instruction data. This mechanism makes teaching intervention no longer a general prompt, but can accurately locate specific errors in the trainee's operation and provide personalized and operable correction guidance. For example, when a trainee omits a specific step or makes a sequential error during operation, the system can immediately identify the error type and extract the corresponding standard operating procedure or equipment highlight prompt template from the knowledge graph, converting it into augmented reality display instruction data or targeted voice prompts. This greatly improves the pertinence and effectiveness of teaching feedback, helps trainees quickly understand and correct errors, and thus significantly improves the efficiency and training quality of medical simulation teaching.

[0106] In one feasible implementation, the step of converting the correction scheme template data into an executable instruction format to generate real-time intervention instruction data includes: when the error type data is an operation sequence error, generating standard process animation data containing step sequence number markers; when the error type data is a device operation omission, generating visual highlight data pointing to a specific location on the medical simulator; and encapsulating the standard process animation data or visual highlight data into augmented reality display instruction data to generate real-time intervention instruction data.

[0107] In this embodiment, when the system identifies an error in the sequence of a student's operations, this application proposes generating standard process animation data containing step number markers to provide clear and dynamic corrective guidance. This animation data can be pre-recorded or dynamically generated by a program, and its content details the execution order of the correct operation steps, key actions, and the usage methods of related equipment. The animation clearly marks the sequence number of each step, for example, through numbers, text descriptions, or dynamic arrows, ensuring that the student can intuitively understand the correct operation flow and sequence, thereby effectively correcting any deviations in their operation sequence.

[0108] In this embodiment, when the system detects that a trainee has missed a necessary operation on a key device during the operation, in order to immediately remind the trainee and guide their attention to the missed device, this application proposes to generate visual highlight data pointing to a specific location on the medical mannequin. This visual highlight data can be achieved using augmented reality technology, for example, highlighting the missed device in the trainee's field of vision with a flashing border, specific color changes, arrow indicators, or overlaid transparent layers. This requires the system to accurately obtain the three-dimensional spatial coordinates of each key device on the medical mannequin and perform real-time rendering based on the trainee's perspective, thereby directly guiding the trainee's attention to the device that needs to be operated.

[0109] In this embodiment, to ensure that the generated standard process animation data or visual highlight data can be effectively received and executed by the feedback execution terminal (e.g., a head-mounted augmented reality device), this application further proposes encapsulating this data into augmented reality display instruction data. The encapsulation process involves encoding information such as the coordinates, style, and duration of the animation sequence or highlighted area into a specific data format that can be parsed and rendered by the augmented reality device. This instruction data not only includes specific display content but may also contain metadata such as display trigger conditions and display priority, ensuring that intervention information can be presented to the learner in an immersive, intuitive, and seamless manner, achieving efficient real-time teaching intervention.

[0110] In this embodiment, through the above-described technical solution, this application can provide highly customized and intuitive real-time interventions for specific error types encountered by trainees during operation on the medical mannequin. When a trainee makes an error in the order of operations, the system no longer simply provides text or voice prompts, but instead generates standard process animation data containing step number markers to dynamically and visually demonstrate the correct operation process, greatly reducing the cognitive burden on trainees to understand and correct errors. Simultaneously, when a trainee misses the operation of key equipment, the system can generate visually highlighted data pointing to specific locations on the medical mannequin, directly highlighting the missed equipment in the trainee's field of vision using augmented reality technology, achieving immediate and precise attention guidance. These specific types of intervention data are encapsulated as augmented reality display instruction data, ensuring that feedback information is presented to trainees in an immersive and highly interactive manner, thereby significantly improving the effectiveness and learning efficiency of real-time teaching interventions, enabling trainees to identify and correct errors more quickly and master correct clinical operation skills.

[0111] In one feasible implementation, the feedback execution terminal includes a head-mounted augmented reality device and / or a feedback control unit integrated into a medical mannequin; the execution of real-time teaching intervention operations includes: displaying operation instructions through the head-mounted augmented reality device; and / or, overlaying an operation instruction layer on the vital signs display of the medical mannequin through the feedback control unit integrated into the medical mannequin; and / or, activating directional voice prompts within a preset timeout period for the student's operation through the feedback control unit integrated into the medical mannequin.

[0112] In this embodiment, the feedback execution terminal is a hardware device used to receive real-time intervention instruction data and convert it into a form that the trainee can perceive. Its main function is to transform the abstract instructions generated by the system into specific visual, auditory, or tactile feedback to guide the trainee in correcting their actions. This terminal can be a standalone device or a module integrated into the medical mannequin system.

[0113] In this embodiment, the head-mounted augmented reality device is a wearable display device that overlays virtual information onto the user's view of the real world. This device typically includes an optical display, sensors, and a processor, capable of sensing the user's head posture and environmental information in real time, and presenting operational guidance directly to the user's field of vision in the form of images, text, or 3D models, allowing them to receive guidance without interrupting their operation. For example, the device could be AR glasses or an AR helmet, using transparent lenses or projection technology to blend virtual images with the real scene. The feedback control unit integrated into the medical mannequin refers to a control module directly embedded or tightly connected to the medical mannequin itself. This unit is responsible for receiving intervention commands from the main control system and driving the medical mannequin's own output devices (such as displays and speakers) to perform feedback operations. This control unit typically includes a microprocessor, memory, and interfaces with various components of the mannequin, enabling modification of the mannequin's display content or playback of voice prompts.

[0114] In this embodiment, displaying operation guidance through a head-mounted augmented reality device refers to converting real-time intervention command data into visual information and presenting it directly to the trainee's field of vision using augmented reality technology. For example, when a trainee's operation is incorrect, the augmented reality device can highlight the correct operation area, indicate the correct instrument position, or demonstrate the correct operation steps in animation form. This method provides immersive, context-sensitive visual feedback, allowing trainees to receive immediate guidance without shifting their gaze. Overlaying an operation guidance layer on the vital signs display of the medical mannequin refers to using a feedback control unit integrated into the medical mannequin to overlay prompts in graphic or text form on the mannequin's built-in vital signs monitoring screen or other information display screen. For example, when a trainee fails to notice changes in key vital signs during a specific operation, the feedback control unit can highlight abnormal indicators on the vital signs display or overlay prompt text to guide the trainee's attention. This method utilizes the trainee's existing visual focus, providing non-invasive auxiliary information. Activating targeted voice prompts within a preset timeout period for student operations means that when a student performs an operation step beyond a pre-defined reasonable time range, the feedback control unit integrated into the medical mannequin will trigger a pre-set voice message. This voice prompt can be a pre-recorded instruction or a prompt generated through text-to-speech technology, such as "Please note, the next step should be to check the patient's airway." "Targeted" means that the voice prompt can provide specific and targeted guidance based on the type of error or the operational scenario, avoiding generalities and ensuring that the student receives timely auditory feedback at critical points.

[0115] In this embodiment, through the above-described technical solution, this application can provide multimodal, real-time, and non-invasive teaching intervention feedback. The use of a head-mounted augmented reality device allows operational guidance to be directly overlaid on the trainee's real field of vision. Trainees can receive immediate and intuitive visual guidance without interrupting their operations or shifting their gaze, significantly reducing cognitive load and maintaining the immersion and continuity of the simulation training. Simultaneously, by integrating the feedback control unit into the medical mannequin and overlaying the operational guidance layer on the mannequin's vital signs display, the system can utilize the trainee's existing visual focus to provide auxiliary prompts for key information, ensuring that trainees receive operational guidance while focusing on vital signs. Furthermore, activating directional voice prompts within a preset timeout period provides timely and targeted auditory feedback, particularly suitable for operational steps requiring strict time control, effectively preventing the accumulation of errors due to operational delays. This multi-sensory, contextualized feedback mechanism greatly improves the timeliness, accuracy, and effectiveness of teaching intervention, enabling trainees to more efficiently identify and correct operational deviations, thereby significantly improving the teaching quality of medical simulation training and the learning efficiency of trainees.

[0116] In the embodiments of this application, the real-time feedback teaching method using medical mannequins collects multi-source behavioral data generated by trainees operating the medical mannequins through a sensor array, performs feature extraction, and compares the data in real time to generate decision deviation data. When the deviation exceeds a threshold, an intervention command is triggered and a teaching intervention is executed. This method can instantly capture operational deviations and dynamically generate targeted guidance. By collecting and analyzing multi-source behavioral data in real time and comparing it with standard clinical decision-making paths, it can instantly detect trainees' operational deviations and automatically trigger teaching interventions when the deviations are too large. This improves the timeliness and accuracy of medical teaching feedback, prevents erroneous operations from being solidified, and enhances training efficiency.

[0117] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the real-time feedback teaching method of the medical manikin in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0118] This application also provides a real-time feedback teaching system using a medical mannequin, see reference. Figure 2 The medical simulator real-time feedback teaching system includes: a memory 10, a processor 20, and a medical simulator real-time feedback teaching program stored on the memory 10 and executable on the processor 20. The medical simulator real-time feedback teaching program is configured to implement the steps of the medical simulator real-time feedback teaching method.

[0119] The medical mannequin real-time feedback teaching system provided in this application, employing the medical mannequin real-time feedback teaching method described in the above embodiments, can improve the timeliness and accuracy of medical teaching feedback and enhance training efficiency. Compared with the prior art, the beneficial effects of the medical mannequin real-time feedback teaching system provided in this application are the same as those of the medical mannequin real-time feedback teaching method provided in the above embodiments, and other technical features of the medical mannequin real-time feedback teaching system are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0120] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0121] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. All equivalent structural transformations made under the technical concept of this application using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included within the scope of patent protection of this application.

Claims

1. A real-time feedback teaching method using a medical simulator, characterized in that, The method includes: The sensor array collects multi-source behavioral data generated by the trainee operating the medical mannequin. The multi-source behavioral data includes voice data, operation timing data, and eye movement trajectory data. The multi-source behavioral data is subjected to feature extraction processing to obtain a behavioral feature dataset; The behavioral feature dataset is compared with the standard clinical decision-making path in real time to generate decision deviation data. When the decision deviation data exceeds a preset threshold, the intervention instruction generation process is triggered, and real-time intervention instruction data is output. The real-time intervention instruction data is sent to the feedback execution terminal to perform real-time teaching intervention operations; The steps of comparing the behavioral feature dataset with the standard clinical decision-making path in real time to generate decision deviation data include: Extract speech keyword data, operation step sequence data, and visual attention area data from the behavioral feature dataset; The voice keyword data is matched with the medical terminology database of the current scene to generate instruction matching degree data. The sequence data of the operation steps is compared with the operation sequence of the standard clinical decision-making path in terms of time sequence alignment to generate time sequence deviation data. The visual attention area data is analyzed and processed to perform spatial overlap analysis with the preset key equipment area to generate attention missing data. The instruction matching data, the timing deviation data, and the attention missing data are weighted and fused to generate decision deviation data; The steps involved in aligning and comparing the sequence data of the operational steps with the operational sequence of the standard clinical decision-making pathway to generate time-series bias data include: Identify the start and end timestamps of each operation action in the operation step sequence data; Calculate the time interval data between adjacent operations based on the start timestamp and end timestamp; The time interval data is compared with the allowed time window of the standard clinical decision-making pathway. Count the number of operation steps that exceed the allowed time window and generate time series deviation data; The steps of performing spatial overlap analysis on the visual attention area data and the preset key equipment area to generate attention-missing data include: Mark the 3D region boundary data of key equipment in the spatial coordinate system of the medical simulator; The coordinate position relationship between the gaze point coordinates in the visual attention area data and the three-dimensional region boundary data is determined. Calculate the percentage of gaze points that are not covered by key equipment within a preset time period to generate attention gap data.

2. The real-time feedback teaching method using a medical simulator as described in claim 1, characterized in that, The steps for generating decision deviation data by weighted fusion of the instruction matching data, the time-series deviation data, and the missing attention data include: Assign a first weight value to the instruction matching data, assign a second weight value to the timing deviation data, and assign a third weight value to the missing data of concern; The weighted instruction matching data, weighted timing deviation data, and weighted attention missing data are input into the deviation calculation function for processing; The weighted data is standardized and scored using the deviation calculation function to generate decision deviation data.

3. The real-time feedback teaching method using a medical simulator as described in claim 1, characterized in that, The method further includes: Parse the pre-set medical guideline document to obtain operational process node data; Based on the operational process node data, an initial clinical decision path is generated; The system receives innovative operational data from trainees that has been verified by experts, and dynamically updates the initial clinical decision-making path based on this data to obtain an updated standard clinical decision-making path.

4. The real-time feedback teaching method using a medical simulator as described in claim 1, characterized in that, When the decision deviation data exceeds a preset threshold, the step of triggering the intervention instruction generation process and outputting real-time intervention instruction data includes: Identify the error type data with the largest deviation value in the decision deviation data; Extract correction scheme template data corresponding to the error type data from the knowledge graph engine; The correction scheme template data is converted into an executable instruction format to generate real-time intervention instruction data.

5. The real-time feedback teaching method using a medical simulator as described in claim 4, characterized in that, The steps of converting the correction plan template data into an executable instruction format to generate real-time intervention instruction data include: When the error type data is an operation sequence error, standard process animation data containing step sequence number markers is generated; When the error type data is a device operation omission, visual highlight data pointing to a specific location on the medical mannequin is generated; The standard process animation data or visual highlight data is encapsulated into augmented reality display instruction data to generate real-time intervention instruction data.

6. The real-time feedback teaching method using a medical simulator as described in claim 1, characterized in that, Feedback execution terminals include head-mounted augmented reality devices and / or feedback control units integrated into medical mannequins; The execution of real-time teaching intervention operations includes: Operation instructions are displayed via the aforementioned head-mounted augmented reality device; And / or, through the feedback control unit integrated into the medical mannequin, an operation guidance layer is overlaid on the vital signs display of the medical mannequin; And / or, through the feedback control unit integrated into the medical simulator, a directional voice prompt is initiated when the student's operation exceeds a preset time.

7. A real-time feedback teaching system for medical simulators, characterized in that, The medical simulator real-time feedback teaching system includes: a memory, a processor, and a medical simulator real-time feedback teaching program stored in the memory and executable on the processor, wherein the medical simulator real-time feedback teaching program is configured to implement the steps of the medical simulator real-time feedback teaching method as described in any one of claims 1 to 6.

Images