Practical training simulation method for fire-fighting skill training based on virtual reality technology

Abstract

The invention relates to the technical field of fire-fighting training information, and particularly discloses a fire-fighting skill training-oriented training simulation method based on a virtual reality technology, and the method sequentially comprises five core steps: learning behavior data collection, personalized learning course generation, supervision and AI question answering management, three-dimensional simulation practical operation, and interactive live broadcast and simulation machine examination. According to the method, firstly, learning behavior data of students are collected through a multi-source sensor, personalized courses are generated after analysis, then intelligent supervision and learning and real-time question answering are realized based on the learning data, and finally, high-simulation fire-fighting practical operation and intelligent evaluation are carried out according to weak spots. According to the method, data coherence and module collaboration in the whole training process are innovatively realized, the problems of single learning path, lack of real-time supervision, insufficient practical operation training, inaccurate evaluation and the like in traditional training are effectively solved, and the efficiency, pertinence and actual combat of fire-fighting skill training are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fire-fighting training information, in particular to a training simulation method for fire-fighting skill training based on virtual reality technology. BACKGROUND

[0002] The application of virtual reality technology in the field of fire-fighting skill training has become an important development direction of modern training methods. It enables trainees to conduct repetitive and highly simulated skill training in a virtual environment by constructing realistic fire scenes and emergency response processes. Some existing systems have attempted to combine three-dimensional simulation and interactive learning to improve the intuitiveness and engagement of training, gradually promoting the transformation of fire-fighting training from traditional teaching methods to digital and immersive methods.

[0003] However, there are still several key problems in existing virtual reality fire-fighting training systems. First, the system lacks comprehensive data collection and in-depth analysis of the learning process of trainees, making it difficult to achieve truly personalized teaching and leading to a mismatch between training content and the actual level of trainees. Second, there is a lack of effective real-time supervision and intelligent answering mechanism during training, which can lead to a lack of motivation and accumulation of problems during self-study. Third, the physical realism and operation feedback accuracy of the simulation scene are insufficient, limiting the effectiveness of skill training. Finally, the data between different training stages is fragmented, lacking unified and coherent data flow support, making it difficult to form a complete process from learning to practice and from evaluation to optimization, affecting the continuous improvement of overall training effectiveness. SUMMARY

[0004] (I) Technical problems solved

[0005] The present application provides a training simulation method for fire-fighting skill training based on virtual reality technology, which solves the problems mentioned in the background technology.

[0006] (II) Technical solutions

[0007] To achieve the above purpose, the present application is implemented by the following technical solutions: a training simulation method for fire-fighting skill training based on virtual reality technology, comprising the following steps:

[0008] S1, learning behavior data collection, including a data collection module, collecting raw learning behavior data of students through intelligent terminals, system logs, motion capture sensors, eye tracking sensors and heart rate monitoring sensors, the raw learning behavior data including login frequency, learning duration, course interaction records, practice completion rate, test scores, simulation operation scores, operation trajectory data, fixation point data and heart rate variability data, and cleaning the collected data to remove outliers and missing values, normalizing the data to unify the data format and dimension, converting the data to structured data through encoding, and storing the data in a distributed database, while establishing a dynamically updated learning profile for each student, the learning profile including learning habit analysis, knowledge mastery assessment and progress trend tracking, providing a data basis for subsequent steps;

[0009] S2, personalized learning course generation, including a course generation module, based on the processed learning behavior data output by the data collection module, analyzing individual differences and learning levels of students through personalized learning algorithms, the personalized learning algorithms including clustering algorithms and collaborative filtering algorithms, dynamically generating customized learning courses for each student, optimizing learning paths and improving efficiency, the output including course content, learning progress data and knowledge point mastery status data calculated from analysis of test scores and simulation operation scores;

[0010] S3, supervisor and AI question answering management, including a management module, based on the learning progress data and knowledge point mastery status data output by the course generation module, real-time monitoring of learning status and providing learning reminders, preparation suggestions and supervisor companion learning services through a supervisor algorithm, while integrating an AI question answering algorithm, using natural language processing technology to achieve instant reply and knowledge summary, generating question answering records and weak point identification;

[0011] S4, three-dimensional simulation practice, including a simulation module, based on the question answering records and weak point data output by the management module, constructing a fire fighting practice scene through a three-dimensional simulation algorithm, including fire simulation, equipment operation and emergency procedures, supporting AR and VR immersive training, and strengthening skill practice through real-time feedback of operation data through the Internet of Things;

[0012] S5, interactive live broadcast and simulation machine examination, including an interaction module, based on the simulation practice data output by the simulation module, providing real-time live broadcast interaction, on-site question answering and playback functions through an interactive live broadcast algorithm, enhancing the learning atmosphere, and conducting simulation examination practice through a simulation machine examination algorithm, including test question generation, automatic correction and post-examination analysis, helping students to find and fill gaps and familiarize themselves with the examination process.

[0013] As a further of the present application, the data acquisition module in step S1 further includes a data quality monitoring unit that identifies errors and biases in data acquisition in real-time through anomaly detection algorithms and performs data completion and verification to ensure data accuracy and integrity, while integrating data encryption and privacy protection mechanisms to prevent unauthorized access, the motion capture sensor specifically includes an inertial measurement unit and optical marker points for collecting body movement and operation posture data of the trainee in the virtual training environment, the eye tracking sensor uses infrared camera technology to collect visual attention distribution data of the trainee, and the heart rate monitoring sensor uses photoplethysmography to collect physiological stress response data of the trainee in the emergency drill, and uses Internet of Things transmission form to transmit data, so as to more comprehensively evaluate the learning state of the trainee.

[0014] As a further of the present application, the course generation module in step S2 specifically includes reading learning behavior data from a distributed database, grouping trainees by learning level using clustering algorithms, and analyzing individual learning preferences and weak links through collaborative filtering algorithms and machine learning models to dynamically generate customized course schedules containing videos, graphics, interactive exercises, and simulation tasks, real-time adjusting course order and difficulty based on learning feedback, and outputting personalized learning course data to the supervisor system, including course ID, learning objectives, estimated duration, key knowledge points, and progress indicators, to achieve precise matching of one person to one course and optimization of adaptive learning paths.

[0015] As a further of the present application, the course generation module in step S2 further includes a course optimization unit that continuously adjusts algorithm parameters through A / B testing and feedback loops, optimizes course content based on trainee performance improvement rates and satisfaction data, and supports multi-modal learning resource integration, including 3D models, interactive videos, and virtual laboratories, to enhance learning engagement and effectiveness.

[0016] As a further of the present application, the management module in step S3 specifically includes real-time monitoring of learning progress and knowledge point mastery status in personalized learning course data, identifying learning risk points based on historical data and prediction models through supervisor algorithms, automatically triggering learning reminder messages and exam preparation suggestions to be sent to trainee terminals, and providing exclusive academic mentors for accompanying learning services, while integrating an AI Q&A unit that uses pre-trained language models and natural language processing techniques to analyze trainee questions, retrieves or generates answers from a fire knowledge base and responds instantly, supports multi-turn conversations and context understanding, records Q&A interaction data and generates knowledge point summary reports, including weak point identification and learning suggestions, for subsequent simulation practice.

[0017] As a further of the present application, the management module in step S3 further comprises an emotion analysis unit, which analyzes the emotional state of the student in the interaction between the tutor and the student through natural language processing technology, adjusts the tutoring strategy and the answering tone to improve the support effect, and integrates the knowledge graph technology to associate the answering record with the fire safety knowledge system, and automatically recommends and individualizes the learning path.

[0018] As a further of the present application, the simulation module in step S4 specifically comprises selecting a corresponding fire operation scene according to the answering record and the weak point identification data, using a three-dimensional modeling engine and a physics engine to build a virtual environment, simulating real fire dynamics, smoke diffusion, device operation and evacuation process, the student conducts immersive training through the VR helmet, the system detects the operation correctness in real time through the motion capture sensor, and provides voice guidance and visual feedback, records the training data including the completion time, the error times, the skill score and the operation trajectory, generates a training report and synchronizes it to the central database for reinforcement practice and data analysis.

[0019] As a further of the present application, the interaction module in step S5 specifically comprises starting the interactive live broadcast system based on the training data output by the simulation module, the teacher conducts real-time teaching, demonstration and answering through the live broadcast platform, the student can interact through text, voice or video, the live broadcast content is automatically recorded and supports playback, at the same time, the simulation machine examination unit is called to randomly extract questions from the question bank to generate simulated test papers, the question types include multiple choice questions, true or false questions and scene operation questions, the system monitors the answering process and time, automatically corrects objective questions and evaluates subjective questions using a rule engine, generates a detailed post-examination analysis report including correct answers, knowledge point analysis and reinforcement exercise recommendations, and simulates the registration process to familiarize with the examination registration steps.

[0020] As a further of the present application, the interaction module in step S5 further comprises a learning effect evaluation unit, which generates a personal ability growth curve and a group ability distribution graph by comparing the simulation examination scores with the historical training data, provides data support for the teacher to adjust the teaching strategy, and recommends targeted reinforcement training content for the student.

[0021] As a further of the present application, the method further comprises a system integration and data flow management module, which integrates all modules through a central data bus and a microservice architecture, ensures seamless data flow and consistency between modules, uses real-time stream processing technology to monitor the data flow state, and uses load balancing and fault tolerance mechanism to ensure system high availability, at the same time, all output data is aggregated through a data analysis platform to generate an overall training effect report including learning efficiency improvement, skill mastery and examination pass rate, which is used for continuous optimization of fire training process.

[0022] (Three) beneficial effects

[0023] This invention provides a practical training simulation method for fire-fighting skills training based on virtual reality technology. It has the following beneficial effects:

[0024] (I) This practical training simulation method for fire-fighting skills training based on virtual reality technology establishes a multi-source data acquisition and intelligent analysis system. It comprehensively utilizes behavioral data such as students' login frequency, learning time, interaction records, and test scores, combined with motion capture, eye tracking, and physiological indicator monitoring, to construct a complete student profile. Based on this data, the system uses cluster analysis and collaborative filtering algorithms to accurately identify individual differences among students and dynamically generate fully personalized learning courses. This achieves precise matching of one course per student. Furthermore, it can not only automatically adjust the difficulty and order of courses based on real-time learning feedback, but also continuously optimize the learning path through a multi-dimensional evaluation model. Ultimately, it forms a data-driven personalized training program, effectively solving the problems of low efficiency caused by the single learning path and insufficient teaching focus in traditional fire-fighting training, and significantly improving training quality and learning outcomes.

[0025] (II) This practical training simulation method for fire protection skills training based on virtual reality technology monitors the learning status of trainees in real time through a learning risk prediction model, automatically identifies learning difficulties and triggers precise supervisory intervention, and integrates intelligent Q&A function. It adopts natural language processing and semantic understanding technology to achieve intelligent Q&A service with a response time of seconds. At the same time, it combines sentiment analysis technology to perceive the emotional state of trainees and dynamically adjust the interaction strategy. Furthermore, it deeply links learning records with the fire protection knowledge system through knowledge graph technology. The system can intelligently recommend learning content and fine-tune the learning path, forming a comprehensive learning support system that integrates supervision, learning companionship, Q&A, and emotional support. It effectively solves the problem of lack of real-time supervision and immediate learning support in traditional training, and significantly improves the continuity of learning and the solidity of knowledge mastery.

[0026] (III) This practical training simulation method for fire-fighting skills training based on virtual reality technology constructs realistic fire scenarios through precise physics engine and fluid dynamics simulation, providing trainees with an immersive practical training environment. Through motion capture and operation trajectory analysis, it achieves accurate assessment and real-time feedback of trainees' operational skills. At the same time, combined with intelligent test paper generation and project response theory, it provides a highly simulated computer-based examination experience and generates detailed analysis of weaknesses. In addition, by establishing a training effect evaluation model, the system can comprehensively track the trainees' ability development trajectory, providing data support for teaching optimization. Ultimately, it forms a complete link from theory to practice, from training to evaluation, effectively solving the problems of insufficient practical opportunities, inaccurate skills assessment, and insufficient exam preparation in traditional fire-fighting training, and greatly improving the practicality of training and the scientific nature of assessment. Attached Figure Description

[0027] Figure 1 This is a flowchart of the method of the present invention;

[0028] Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] First embodiment: as follows Figure 1 and Figure 2 As shown, the present invention provides a technical solution: a practical training simulation method for fire-fighting skills training based on virtual reality technology, comprising the following steps:

[0031] S1. Learning behavior data collection, including a data collection module, collects raw learning behavior data of students through smart terminals, system logs, motion capture sensors, eye-tracking sensors, and heart rate monitoring sensors. The raw learning behavior data includes login frequency, learning duration, course interaction records, exercise completion rate, test scores, simulated operation scores, operation trajectory data, fixation point data, and heart rate variability data. The collected data is cleaned to remove outliers and missing values, normalized to unify data format and units, and converted into structured data through encoding and stored in a distributed database. At the same time, a dynamically updated learning profile is established for each student. The learning profile includes learning habit analysis, knowledge mastery assessment, and progress trend tracking, providing a data foundation for subsequent steps.

[0032] S2. Personalized learning course generation, including a course generation module. Based on the processed learning behavior data output by the data acquisition module, it analyzes individual differences and learning levels of students through personalized learning algorithms, including clustering algorithms and collaborative filtering algorithms. It dynamically generates customized learning courses for each student, optimizes the learning path and improves efficiency. Its output includes course content, learning progress data and knowledge point mastery status data calculated by analyzing test scores and simulated operation scores.

[0033] S3, Learning Supervision and AI Q&A Management, includes a management module that uses learning progress data and knowledge point mastery status data output by the course generation module to monitor learning status in real time and provide learning reminders, exam preparation suggestions and learning companion services through the learning supervision algorithm. At the same time, it integrates an AI Q&A algorithm that uses natural language processing technology to achieve instant Q&A and knowledge point summaries, and generates Q&A records and weak point identification.

[0034] S4, 3D simulation practice, including a simulation module, based on the Q&A records and weak point data output by the management module, constructs fire protection practice scenarios through 3D simulation algorithms, including fire simulation, equipment operation and emergency procedures, supports AR and VR immersive training, and strengthens skills practice by providing real-time feedback of operation data through the Internet of Things;

[0035] S5, Interactive Live Streaming and Simulated Exams, includes an interactive module. Based on the simulation practice data output by the simulation module, it provides real-time live interaction, on-site Q&A, and replay functions through interactive live streaming algorithms to enhance the learning atmosphere. It also conducts simulated exam practice through simulated exam algorithms, including question generation, automatic grading, and post-exam analysis, to help students identify and fill knowledge gaps and familiarize themselves with the exam process.

[0036] The data acquisition module in step S1 also includes a data quality monitoring unit, which uses an anomaly detection algorithm to identify errors and deviations in data acquisition in real time, and performs data completion and verification to ensure data accuracy and integrity. It also integrates data encryption and privacy protection mechanisms to prevent unauthorized access. The motion capture sensor specifically includes an inertial measurement unit and optical markers to collect data on trainees' limb movements and operational postures in the virtual training environment. The eye-tracking sensor uses infrared imaging technology to collect data on trainees' visual attention distribution. The heart rate monitoring sensor uses photoelectric volumetric plethysmography to collect data on trainees' physiological stress response during emergency drills and transmits the data using IoT transmission to more comprehensively assess trainees' learning status.

[0037] The course generation module in step S2 specifically includes reading learning behavior data from a distributed database, grouping students according to their learning level using a clustering algorithm, analyzing individual learning preferences and weaknesses through collaborative filtering algorithms and machine learning models, dynamically generating a customized course schedule that includes videos, text and images, interactive exercises and simulated tasks, adjusting the course order and difficulty in real time based on learning feedback, and outputting personalized learning course data to the learning supervision system, including course ID, learning objectives, estimated duration, key knowledge points and progress indicators, to achieve precise matching of one course per student and adaptive optimization of learning paths.

[0038] The course generation module in step S2 also includes a course optimization unit, which continuously adjusts algorithm parameters through A / B testing and feedback loops, optimizes course content based on student performance improvement rate and satisfaction data, and supports the integration of multimodal learning resources, including 3D models, interactive videos and virtual laboratories, to enhance learning engagement and effectiveness.

[0039] Specifically, the data acquisition module first uses a smart terminal to collect data on students' system login frequency, learning duration, and number of course interactions. It then uses a motion capture sensor to record operation trajectory data, an eye-tracking sensor to collect gaze point data, and a heart rate monitoring sensor to acquire heart rate variability data. The raw learning behavior data is then cleaned and processed, specifically using a box plot method to identify and remove outliers. The formulas for calculating the upper and lower boundaries are as follows:

[0040]

[0041] Where Q1 is the first quartile, Q3 is the third quartile, and IQR is the interquartile range. Then, the missing values ​​are imputed using the K-nearest neighbor algorithm. The k most similar samples are selected based on Euclidean distance for mean imputation. The distance calculation formula is as follows:

[0042]

[0043] Then, data normalization is performed. The min-max normalization method is used to transform the data to the [0,1] interval. The calculation formula is as follows:

[0044]

[0045] The processed structured data is stored in a distributed database, and student learning profiles are created, including learning habit analysis, knowledge mastery assessment, and progress trend tracking. Then, the course generation module reads the learning behavior data from the database and uses the K-means clustering algorithm to group students according to their learning level, with the objective function being to minimize the squared error.

[0046]

[0047] Where k is the number of clusters and n is the number of samples. As cluster centers, the optimal k value is determined using the elbow rule. Then, a collaborative filtering algorithm is used to analyze learners' learning preferences, and similarity is calculated based on the user project rating matrix using the cosine similarity formula.

[0048]

[0049] Based on these analysis results, a customized course schedule is dynamically generated, including videos, text and images, interactive exercises, and simulated tasks. The course order and difficulty are adjusted in real time. Finally, a knowledge point mastery assessment algorithm is used to calculate the student's mastery of each knowledge point. The calculation formula is as follows:

[0050]

[0051] in For test scores, To score the simulated operation, To enhance interactive engagement, α, β, and γ are weighted coefficients, with α+β+γ=1. The output includes course content, learning progress data, and knowledge point mastery status data, enabling personalized learning matching. This embodiment effectively solves the problems of single learning paths and low efficiency through data-driven personalized course generation, improving the relevance and learning efficiency of fire protection training.

[0052] Second embodiment: as follows Figure 1 and Figure 2 As shown, the management module in step S3 specifically includes real-time monitoring of learning progress and knowledge mastery status in personalized learning course data; identifying learning risk points based on historical data and prediction models through a learning supervision algorithm; automatically triggering learning reminder messages and exam preparation suggestions to be sent to the student's terminal; providing dedicated learning assistant services; integrating an AI Q&A unit; using pre-trained language models and natural language processing technology to analyze student questions; retrieving or generating answers from the fire safety knowledge base and replying instantly; supporting multi-turn dialogue and contextual understanding; recording Q&A interaction data and generating knowledge point summary reports, including weak point identification and learning suggestions, for subsequent simulation practice.

[0053] The management module in step S3 also includes a sentiment analysis unit, which uses natural language processing technology to analyze the emotional state of learners during Q&A and supervision interactions, adjusts supervision strategies and Q&A tone to improve support effectiveness, and integrates knowledge graph technology to link Q&A records with the fire safety knowledge system, automatically recommending and fine-tuning personalized learning paths.

[0054] Specifically, based on the learning progress data and knowledge point mastery status data output by the first embodiment, a learning risk prediction model is first constructed using the supervisory algorithm in the management module. Then, multiple linear regression analysis is used to analyze the relationship between historical learning data and learning outcomes. The regression equation is expressed as follows:

[0055]

[0056] Where Y is the learning performance evaluation value. to Independent variables include learning duration, interaction frequency, and test scores. For the intercept term, to For regression coefficients, For the error term, the risk prediction value is obtained after estimating the parameters using the least squares method. When the predicted value is lower than the threshold θ, a learning reminder is automatically triggered. Then, in the AI ​​Q&A unit, a TF-IDF-based text similarity calculation algorithm is used to process student questions. First, the word frequency is calculated:

[0057]

[0058] in Let t represent the number of times word t appears in document d, and then calculate the inverse document frequency:

[0059]

[0060] Where N is the total number of documents, the TF-IDF weights are finally obtained:

[0061]

[0062] Calculate the matching degree between the question and the standard questions in the knowledge base using cosine similarity:

[0063]

[0064] When the similarity is higher than the threshold δ, a preset answer is returned directly; otherwise, a response is generated using a BERT-based semantic understanding model. Simultaneously, a dictionary-based sentiment analysis method is used in the sentiment analysis unit, with the sentiment score calculated using the following formula:

[0065]

[0066] in Weighting of sentiment words The emotional polarity value is used to adjust the learning strategy based on the emotional score. Finally, a fire safety knowledge association network is constructed using knowledge graph technology, and the PageRank algorithm is used to calculate the importance of knowledge points.

[0067]

[0068] Where d is the damping coefficient and N is the total number of nodes. For pointing A collection of pages, for The number of output chains is used to intelligently recommend personalized learning paths. This embodiment effectively solves the problem of lack of supervision and immediate support in the learning process through real-time learning supervision and intelligent Q&A, thereby enhancing the continuity of students' learning and the effect of knowledge consolidation.

[0069] Third embodiment: as follows Figure 1 and Figure 2As shown, the simulation module in step S4 specifically includes selecting the corresponding fire-fighting practice scenario based on the Q&A records and weak point identification data, constructing a virtual environment using a 3D modeling engine and a physics engine, simulating real fire dynamics, smoke diffusion, equipment operation, and evacuation procedures, and conducting immersive training through VR headsets. The system uses motion capture sensors to detect the correctness of operations in real time via data transmitted through the Internet of Things, and provides voice guidance and visual feedback. It records training data including completion time, number of errors, skill scores, and operation trajectories, generates training reports, and synchronizes them to the central database for reinforcement of practice and data analysis.

[0070] The interactive module in step S5 specifically includes launching the interactive live streaming system based on the training data output by the simulation module. Teachers conduct real-time lectures, demonstrations, and Q&A sessions through the live streaming platform. Students can interact via text, voice, or video. The live streaming content is automatically recorded and supports playback for review. At the same time, the simulation computer-based testing unit is invoked to randomly select questions from the question bank to generate mock exams. The question types include multiple-choice questions, true / false questions, and scenario-based operation questions. The system monitors the answering process and time, automatically grades objective questions, and uses a rule engine to evaluate subjective questions. A detailed post-exam analysis report is generated, including correct answers, knowledge point analysis, and recommended reinforcement exercises. The registration process is also simulated to familiarize students with the exam registration steps.

[0071] The interactive module in step S5 also includes a learning effect evaluation unit, which generates individual ability growth curves and group ability distribution maps by comparing simulated exam scores with historical training data. This provides data support for teachers to adjust their teaching strategies and recommends targeted reinforcement training content for students.

[0072] The method also includes a system integration and data flow management module, which integrates all modules through a central data bus and microservice architecture to ensure seamless data flow and consistency between modules. It uses real-time stream processing technology to monitor the data flow status and employs load balancing and fault tolerance mechanisms to ensure high system availability. At the same time, it aggregates all output data through a data analysis platform to generate an overall training effectiveness report, including learning efficiency improvement, skill mastery, and exam pass rate, for continuous optimization of the fire training process.

[0073] Specifically, based on the Q&A records and vulnerability data output from the second embodiment, a fire-fighting simulation scenario is first constructed using a 3D modeling engine through the simulation module. The Navier-Stokes equations are then used to simulate the smoke diffusion process, and the governing equations are expressed as follows:

[0074]

[0075] in Let ρ be density, v be velocity vector, p be pressure, τ be stress tensor, and g be gravitational acceleration. The finite element method is used to solve the heat conduction equation to simulate the temperature field distribution.

[0076]

[0077] in Let k be the specific heat capacity, k be the thermal conductivity coefficient, T be the temperature, and Q be the heat source term. Then, motion capture sensors record the student's operation data, and a dynamic time warping algorithm is used to calculate the similarity between the operation trajectory and the standard operation.

[0078]

[0079] Skill assessment scores are calculated based on operational data.

[0080]

[0081] Where A represents operational accuracy, B represents completion time score, C represents error count deduction, and D represents process completeness, with weighting coefficients satisfying the following: Then, the test paper is generated through the simulation testing system in the interactive module, and the test item parameters are calculated using item response theory. Its logistic model is as follows:

[0082]

[0083] in Let θ represent the probability that a candidate with ability θ answers the i-th question correctly. For differentiation, For difficulty, To estimate the parameters, the degree of weakness in knowledge mastery is calculated during the post-exam analysis phase:

[0084]

[0085] in Let i be the score rate for question i. The correlation index between test item i and knowledge point k is used as the basis for analysis. Finally, a multiple linear regression model is established through the learning effect evaluation unit to analyze the training effect.

[0086]

[0087] Where Y represents the overall ability score. For simulation training results, For computer-based test scores, For operational proficiency, To assess knowledge mastery, the model generates individual ability growth curves and group ability distribution maps, providing data support for adjusting teaching strategies. This embodiment effectively addresses the issues of insufficient skills practice and inadequate exam preparation through high-fidelity simulation and an intelligent assessment system, significantly improving the practicality and accuracy of fire safety training and assessment.

[0088] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0089] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A training simulation method for fire-fighting skill training based on virtual reality technology, characterized in that: The method comprises the following steps: S1, learning behavior data collection, including a data collection module, collecting raw learning behavior data of students through intelligent terminals, system logs, motion capture sensors, eye tracking sensors, and heart rate monitoring sensors, the raw learning behavior data including login frequency, learning duration, course interaction records, practice completion rate, test scores, simulation operation scores, operation trajectory data, fixation point data, and heart rate variability data, and cleaning the collected data to remove outliers and missing values, normalizing the data to unify the data format and dimension, converting the data to structured data through encoding, and storing the data in a distributed database, while establishing a dynamically updated learning profile for each student, the learning profile including learning habit analysis, knowledge mastery assessment, and progress trend tracking, providing a data basis for subsequent steps; S2, personalized learning course generation, including a course generation module, based on the processed learning behavior data output by the data collection module, analyzing individual differences and learning levels of students through personalized learning algorithms, including clustering algorithms and collaborative filtering algorithms, dynamically generating customized learning courses for each student, the output including course content, learning progress data, and knowledge point mastery status data calculated from analysis of test scores and simulation operation scores; S3, supervision and AI question answering management, including a management module, based on the learning progress data and knowledge point mastery status data output by the course generation module, monitoring learning status in real time and providing learning reminders, preparation suggestions, and supervision companion learning services through a supervision algorithm, while integrating an AI question answering algorithm, using natural language processing technology to achieve instant reply and knowledge summary, generating question answering records and weak point identification; S4, three-dimensional simulation practical operation, including a simulation module, based on the question answering records and weak point data output by the management module, constructing a fire fighting practical operation scene through a three-dimensional simulation algorithm, including fire simulation, device operation, and emergency process, supporting AR and VR immersive training, and strengthening skill practice through real-time feedback of operation data through the Internet of Things; S5, interactive live broadcast and simulation machine examination, including an interaction module, based on the simulation practice data output by the simulation module, providing real-time live broadcast interaction, on-site question answering, and playback functions through an interactive live broadcast algorithm, and conducting simulated examination practice through a simulation machine examination algorithm, including test question generation, automatic correction, and post-examination analysis, helping students find and fill gaps and familiarize themselves with the examination process. 2.The virtual reality technology-based training simulation method for fire-fighting skill training according to claim 1, characterized in that: The data collection module in step S1 further includes a data quality monitoring unit that uses an anomaly detection algorithm to identify errors and biases in data collection in real time, and performs data completion and verification to ensure data accuracy and integrity, while integrating data encryption and privacy protection mechanisms, the motion capture sensor specifically includes an inertial measurement unit and optical markers for collecting body movement and operation posture data of students in a virtual training environment, the eye tracking sensor uses infrared camera technology to collect visual attention distribution data of students, and the heart rate monitoring sensor uses photoplethysmography to collect physiological stress response data of students during emergency drills. 3.The virtual reality technology-based training simulation method for fire-fighting skill training according to claim 2, characterized in that: The course generation module in step S2 specifically includes reading learning behavior data from a distributed database, grouping students by learning level using clustering algorithms, and analyzing individual learning preferences and weaknesses through collaborative filtering algorithms and machine learning models. It dynamically generates customized course schedules containing videos, text, interactive exercises, and simulation tasks, adjusts course order and difficulty in real time based on learning feedback, and outputs personalized learning course data to the tutor system, including course ID, learning objectives, estimated duration, key knowledge points, and progress indicators.

4. The training simulation method for fire-fighting skill training based on virtual reality technology according to claim 3, characterized in that: The course generation module in step S2 also includes a course optimization unit that continuously adjusts algorithm parameters through A / B testing and feedback loops, optimizes course content based on student performance improvement rates and satisfaction data, and supports multi-modal learning resource integration, including 3D models, interactive videos, and virtual laboratories.

5. The training simulation method for fire-fighting skill training based on virtual reality technology according to claim 4, characterized in that: The management module in step S3 specifically includes real-time monitoring of learning progress and knowledge point mastery in personalized learning course data, identifying learning risk points based on historical data and predictive models through tutor algorithms, automatically triggering learning reminder messages and exam preparation suggestions to student terminals, and providing exclusive academic tutors for accompanying services. It also integrates an AI Q&A unit that uses pre-trained language models and natural language processing techniques to analyze student questions, retrieve or generate answers from a fire safety knowledge base, and respond instantly. It supports multi-turn conversations and context understanding, records Q&A interaction data, and generates knowledge point summary reports, including weak point identification and learning suggestions, for subsequent simulation practice.

6. The training simulation method for fire-fighting skill training based on virtual reality technology according to claim 5, characterized in that: The management module in step S3 also includes an emotional analysis unit that analyzes students' emotional states during Q&A and tutoring interactions using natural language processing techniques, and integrates knowledge graph technology to associate Q&A records with the fire safety knowledge system, automatically recommending and fine-tuning personalized learning paths.

7. The virtual reality technology-based practical simulation method for fire-fighting skill training according to claim 6, characterized in that: The simulation module in step S4 specifically includes selecting corresponding fire operation scenarios based on Q&A records and weak point identification data, using three-dimensional modeling engines and physics engines to build virtual environments that simulate real fire dynamics, smoke diffusion, device operation, and evacuation processes. Students can immerse themselves in training through VR headsets, and the system can detect operation correctness in real time through motion capture sensors, providing voice guidance and visual feedback. Training data is recorded, including completion time, error count, skill score, and operation trajectory, and a training report is generated and synchronized to the central database.

8. The training simulation method for fire-fighting skill training based on virtual reality technology according to claim 7, characterized in that: The interaction module in step S5 specifically includes starting an interactive live streaming system based on the training data output by the simulation module. Teachers can conduct real-time lectures, demonstrations, and Q&A through live streaming platforms, and students can interact through text, voice, or video. Live content is automatically recorded and supports playback, and a simulation exam unit is called to randomly select questions from a question bank to generate simulation test papers. Question types include multiple-choice, true / false, and scenario operation questions. The system monitors the answering process and time, automatically grades objective questions, and uses a rule engine to evaluate subjective questions. Detailed post-exam analysis reports are generated, including correct answers, knowledge point analysis, and recommended reinforcement exercises. The system also simulates the registration process to familiarize students with the exam registration steps.

9. The virtual reality technology-based training simulation method for fire-fighting skill training according to claim 8, characterized in that: The interaction module in the step S5 further includes a learning effect evaluation unit, which generates a personal ability growth curve and a group ability distribution graph by comparing the simulation test scores with historical training data, provides data support for teachers to adjust teaching strategies, and recommends targeted reinforcement training content for students.

10. The virtual reality technology-based training simulation method for fire-fighting skill training according to claim 8, characterized in that: The method further includes a system integration and data flow management module, which integrates all modules through a central data bus and a micro-service architecture, monitors the data flow state using real-time stream processing technology, and adopts load balancing and fault tolerance mechanisms. Meanwhile, through a data analysis platform, all output data are aggregated to generate an overall training effect report, including learning efficiency improvement, skill mastery, and examination pass rate.

Images