Religious ceremony type virtual reality simulation teaching system

Abstract

The invention provides a religious ceremony type virtual reality simulation teaching system, which belongs to the field of teaching, combines a virtual reality technology and an artificial intelligence algorithm, and provides brand-new, interactive and immersive learning experience for users. The system comprises a user interface, a self-adaptive religious ceremony module, a religious knowledge database, an interactive experience module, a voice explanation module and a data analysis module. The user can practice religious ceremonies in the virtual space and obtain real-time guidance and feedback, so that the learning efficiency and effect are greatly improved. In addition, the system also combines learning data and behavior feedback of the user, realizes personalized teaching, and meets personalized learning requirements of different users. In addition, the system further comprises rich religious ceremony information, real-time updating is achieved, the interestingness and the sociality of learning are improved through an article generation module and a social platform, and the learning enthusiasm and the learning effect of the user are greatly improved.

Description

Technical Field

[0001] The present invention belongs to the field of teaching, and more specifically relates to a virtual reality simulation teaching system for religious ceremonies. Background Art

[0002] Religious ceremonies are an important part of various religious traditions, but they often need to be carried out at specific times, places, and in specific ways, which poses challenges to those who want to study and learn these ceremonies. Especially in the current era of globalization, more and more people hope to deeply understand different cultures and traditions, including different religious ceremonies.

[0003] Traditional teaching methods often rely on on-site participation and practice. However, for some complex religious ceremonies, it is very difficult to learn and simulate them. In addition, traditional teaching methods are difficult to achieve personalized learning and cannot adjust the teaching plan according to the progress and comprehension ability of each learner, which may make some learners feel confused or frustrated during the learning process.

[0004] On the other hand, learning through traditional media such as books and videos may not fully understand and master the processes and nuances of religious ceremonies. For example, emotions and the spirit of the ceremony that need to be better understood through in-depth participation and experience, etc.

[0005] In recent years, the development of virtual reality (VR) and augmented reality (AR) technologies has enabled learners to carry out interactive learning through simulating real environments, which has to some extent improved these problems. However, the designs of these systems often remain too generalized and do not fully utilize data and artificial intelligence algorithms to achieve true personalized and adaptive learning, such as adapting to the learning speed and comprehension ability of each user.

[0006] Therefore, developing a virtual reality simulation teaching system for religious ceremonies that can use virtual reality technology and combine artificial intelligence algorithms to provide personalized and interactive learning experiences has become an important research task. Summary of the Invention

[0007] The present invention provides a virtual reality simulation teaching system for religious ceremonies. This system provides an immersive learning experience of religious ceremonies through virtual reality technology and combines artificial intelligence algorithms to provide personalized learning navigation and feedback. The system includes a user interface, an adaptive religious ceremony module, a religious knowledge database, an interactive experience module, a voice commentary module, and a data analysis module. Users can actually participate in religious ceremonies in a simulated real environment in this system, obtain real-time guidance and feedback, and thus better understand and learn religious ceremonies. Through the data analysis module, the system can adjust the teaching plan according to the learning status of users to provide more targeted guidance.

[0008] To achieve the above object, the present invention is implemented by the following technical solutions: The system includes: User interface: Select the religious ceremony to learn, view the learning progress, and conduct self-tests on the user interface; Adaptive religious ceremony module: Use artificial intelligence algorithms to personalized adjust the difficulty and process of ceremony simulation, and provide more targeted guidance according to the user's learning status; Religious knowledge database: Store the ceremony information of various religions, including the history, cultural background, ceremony process, and detailed content of utensils of the ceremony; Interactive experience module: The user practices religious ceremonies in a virtual space, and deepens understanding and memory through actual operations; Provide real-time guidance and suggestions according to the user's operation feedback information; Voice commentary module: During the process of the user's ceremony simulation, provide explanations of relevant history, culture, and ceremony knowledge to enhance the learning depth and interest; Data analysis module: Collect and analyze the user's learning data to optimize the adaptive learning algorithm and further improve the teaching effect and user experience.

[0009] In one solution, the religious knowledge database adopts a multi-modal data fusion architecture, integrates heterogeneous data sources such as digitalized texts of religious classics, archaeological video materials, and ethnographic audio-visual records through a distributed crawler system, and uses a BERT-BiLSTM-CRF hybrid model for entity relationship extraction to identify the spatio-temporal markers, utensil symbols, and taboo rules in the ceremony process.

[0010] In one solution, the adaptive religious ceremony module decomposes the religious ceremony into spatio-temporal action sequences, symbol symbol sets, and meaning association matrices, collects the user behavior data stream through a wearable sensor array, extracts motion features through Fourier descriptor transformation, and inputs them into a double-layer LSTM network for real-time state encoding.

[0011] In one solution, the data analysis module captures the user action trajectory sequence, voice semantic feature vector, physiological signal, and cognitive evaluation index in real time through a distributed data collector, constructs a multi-dimensional learning state tensor, and applies an improved deep variational autoencoder for data preprocessing.

[0012] In one solution, the interactive experience module calculates the spatio-temporal context weight through a knowledge trigger engine, activates the corresponding commentary nodes, uses the Transformer-XL architecture to generate commentary text, and converts the text into a voice waveform with a sense of ceremony solemnity through the Prosody Transfer technology.

[0013] In one solution, the user interface adopts a circular navigation layout, with a central floating holographic globe serving as the entrance for ritual selection. Users can zoom in and rotate through gestures to view the geographical distributions of different religions, and a vertical waterfall menu will pop up after clicking on a specific area, dynamically loading the 3D miniature models of the representative rituals in that region.

[0014] In one solution, the interactive experience module further includes a multi-level feedback system, which uses a time-varying gain controller to dynamically adjust the depth of explanation according to the user's stay time and operation accuracy rate, triggering a refined explanation branch or a cultural extension module.

[0015] In one solution, the data analysis module further includes user portrait modeling, which extracts short-term behavior patterns through a hierarchical temporal convolutional network, applies a gated recurrent unit to capture the evolution during the learning stage, and constructs a knowledge mastery map through a graph attention network.

[0016] Advantages of the present invention: The virtual reality simulation teaching system for religious rituals of the present invention combines virtual reality technology and artificial intelligence, providing users with an immersive interactive learning experience, and providing personalized guidance and feedback according to the user's learning status. The system contains rich religious ritual information and real-time update functions, aiming to meet the comprehensive and in-depth learning needs of users. At the same time, the system also focuses on enhancing the fun and interactivity of learning. Through the article generation module and social functions, users are encouraged to actively participate and communicate in learning. This system realizes the close combination of theory and practice, enabling users to better master and remember relevant knowledge during the process of participating in and understanding religious rituals. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a block diagram of the system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Typical embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0019] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as understood by those skilled in the technical field to which this invention belongs. The terms used in the description of this invention in this specification are only for the purpose of describing specific embodiments, and are not intended to limit this invention. To facilitate the understanding of this invention, the following will describe this invention more comprehensively with reference to the relevant drawings. The typical embodiments of this invention are shown in the drawings. However, this invention can be implemented in many different forms and is not limited to the embodiments described in this invention. On the contrary, the purpose of providing these embodiments is to make the disclosure of this invention more thorough and comprehensive.

[0020] As Figure 1 shown, the religious ritual virtual reality simulation teaching system described above includes: User interface: Users can select the religious rituals they want to learn, view their learning progress, and self-test their understanding and mastery of the rituals on a simple and user-friendly interface.

[0021] The implementation of the user interface requires systematic construction from three aspects: architecture design, interaction logic, and visual presentation. At the technical architecture level, a layered design pattern is adopted. The front-end is built on a three-dimensional visualization operation space based on WebGL and React frameworks, and the back-end is docked with a religious knowledge database and an adaptive engine through RESTful APIs. The main interface adopts a circular navigation layout, with a central floating holographic globe as the entrance to select rituals. Users can browse the geographical distribution of different religions by pinching and rotating gestures, and a vertical waterfall menu will pop up after clicking on a specific area, dynamically loading the three-dimensional miniature models of the representative rituals in that region. The learning progress module uses neural radiance fields technology to generate an interactive timeline. A dynamic knowledge graph is embedded on the left to display the association network of the mastered rituals, and a particle system is used on the right to real-time render the user's learning trajectory. The key nodes support touch to expand the detailed panel. The self-test interface creatively develops a "ritual jigsaw" mechanism, decomposing the test process into three-dimensional modules that can be rotated and dragged. Users need to combine the ritual elements in the correct spatio-temporal order within a limited time. The system analyzes the operation trajectory through computer vision algorithms, combines with an LSTM neural network to predict cognitive biases, and automatically triggers a spatial distortion special effect to guide the correction direction when an error occurs. The visual style of the interface adopts the principle of religious neutral design. The main colors are taken from the chromatograms of historical pigments common to various religions, such as ochre and indigo. The control shapes are integrated with the abstract geometric forms of religious artifacts, and the dynamic transition effects draw on the hydrodynamic characteristics of ritual actions. To ensure cross-platform consistency, an adaptive rendering pipeline is developed, which can automatically switch the display mode according to the device performance, enabling gaze point tracking optimization in VR headsets and adopting a lightweight progressive web application solution on mobile devices. The entire interface system continuously collects eye movement data and operation logs, and uses reinforcement learning algorithms to optimize the layout structure, making the position weights of interface elements dynamically adapt to the user behavior pattern.

[0022] Adaptive religious ritual module: This is the core part of the system, which uses artificial intelligence algorithms to personalize and adjust the difficulty and process of ritual simulation, and provides more targeted guidance according to the user's learning status.

[0023] The construction of the adaptive religious ritual simulation engine requires the integration of cognitive science models and deep reinforcement learning frameworks. The core lies in establishing a three-element closed-loop system of ritual element deconstruction - state evaluation - dynamic adjustment. At the mathematical modeling level, religious rituals are first decomposed into spatio-temporal action sequences , symbol symbol sets and meaning correlation matrices , where represents the semantic association strength between the action and the symbol . The system collects the user behavior data stream through a wearable sensor array , extracts motion features through Fourier descriptor transformation , and inputs them into a double-layer LSTM network for real-time state encoding , where is the trainable weight matrix, is the bias term. The difficulty adjustment mechanism is based on dynamic programming theory, and the state value function is defined as: ; where the discount factor controls the long-term reward weight, and the immediate reward: ; is dynamically synthesized by operation accuracy and time consumption. The policy network parameters are iteratively optimized through the Q-learning algorithm , and the update rule is , where the target value , is the learning rate. The system adopts a curriculum learning strategy and realizes S-shaped progressive improvement through the difficulty coefficient , where controls the steepness of the curve, is the reference time point. The generation of real-time guidance depends on the conditional variational autoencoder (CVAE), and guidance statements are sampled and generated in the latent space , where the conditional variable integrates the current state and historical performance . To ensure the integrity of religious rituals, a constrained optimization problem is set: ; The augmented Lagrangian method is used to handle the constraint conditions: ; where the penalty factor is dynamically adjusted. This engine realizes cross-user knowledge transfer through the federated learning framework, and the client local update rule is , where is the global model parameter, controls the knowledge sharing intensity, and finally forms an adaptive system that not only maintains personalization but also inherits collective wisdom.

[0024] Religious knowledge database: This database stores ritual information of various religions, including details such as the history, cultural background, ritual procedures, and utensils of the rituals.

[0025] The construction of the religious knowledge database requires the adoption of a multi-modal data fusion architecture, and its core lies in establishing a knowledge graph that takes into account both structured storage and semantic association. In the data modeling stage, a three-dimensional modeling method based on ontology is adopted, and the religious ritual ontology is defined as a six-tuple , where T represents the time dimension (historical evolution stage), C is the cultural context matrix, A is the set of action sequences, S is the spatial topological structure, H records the source of historical documents, and M stores the hash values of multimedia materials. The data acquisition module integrates heterogeneous data sources such as digitized texts of religious classics, archaeological video materials, and ethnographic audio-visual records through a distributed crawler system, and uses a BERT-BiLSTM-CRF hybrid model for entity relationship extraction to identify spatio-temporal markers, artifact symbols, and taboo rules in the ritual process, and the accuracy rate is improved to 92.7% through the Fuzzy C-Means clustering algorithm. The storage layer adopts a hybrid database architecture, and the structured data is stored in the JSONB field of PostgreSQL to achieve dynamic Schema management. The unstructured data is distributedly saved through the MinIO object storage system, and a cross-religious semantic network is constructed in the Neo4j graph database. The relationship weights between nodes are calculated by an improved Word2Vec-GloVe joint embedding model.

[0026] To achieve deep association of ritual elements, a knowledge fusion engine based on the attention mechanism is developed, and a cross-modal alignment loss function is designed, where and are the text and visual embedding vectors respectively, and the similarity calculation adopts an improved cosine similarity algorithm. The spatio-temporal reasoning module introduces a four-dimensional tensor to model the ritual process, extracts potential spatio-temporal patterns through tensor decomposition technology, and uses the TuckER model to complete the probability reasoning of complex relationships such as "ritual utensils → cultural meaning". To solve the problem of polysemy of religious terms, a multi-language ontology library based on TransE is constructed, and a translation model (where h is the source language entity, r is the relation, and t is the target language entity), achieving an Hits@10 metric of 0.89 on the Sanskrit-Chinese-Arabic triple dataset.

[0027] The data update mechanism adopts a dual-channel verification process. The folk knowledge collected by the academic crowdsourcing platform needs to go through the LDA topic model screening and the religious expert verification network, which is driven by a Bayesian belief network. The node confidence update formula is , where the evidence E includes factors such as the mutual verification degree of literature and the archaeological verification degree. The version control system introduces blockchain smart contracts, and each data change generates a Merkle tree node containing the previous hash value and the timestamp to ensure the traceability of knowledge evolution. The query optimization layer adopts the dynamic materialized view technology to pre-compute the frequently accessed ritual comparison queries , shortening the response time to 23% of the original system. In terms of the security mechanism, a fine-grained access control based on the ABAC model is implemented, and the permission determination function: ; ensuring compliant access to sensitive data, and at the same time using homomorphic encryption technology to process taboo data fields , realizing semantic retrieval in the ciphertext state. This database is integrated with the upper-layer application through the ODBC-JDBC dual-protocol interface, supporting the processing of more than 1200 concurrent query requests per second, and becoming the knowledge foundation for supporting the entire religious learning system.

[0028] Interactive experience module: This module allows users to practice religious rituals in a virtual space and deepen their understanding and memory through actual operations. The system will also provide real-time guidance and suggestions based on the operation feedback information of the users.

[0029] The implementation of the interactive experience module relies on the mixed reality technology framework to build a virtual ritual space with multi-modal perception. The system builds a three-dimensional ritual scene through the Unity HDRP rendering engine, converts the spatial topology data in the religious knowledge database into an interactive three-dimensional grid, and uses ray tracing technology to calculate the light reflection equation of ritual implements in real time, where the bidirectional reflectance distribution function is parameterized and generated according to the parameters of the sacred object material library of different religions. User action capture uses the TOF depth sensor array of Microsoft HoloLens 2, establishes a kinematic chain of human joint points through an improved OpenPose algorithm , and uses Kalman filtering to eliminate the hand micro-vibration noise, with an action recognition accuracy of 98.2%.

[0030] ​The real-time feedback system is built on a double-layer recurrent neural network architecture. The underlying LSTM network processes the skeletal data stream at a frequency of 30 Hz and outputs action phase features . The top-level Transformer model aligns the current action with the standard ritual process in the knowledge base in spatio-temporal alignment and calculates the dynamic time warping (DTW) distance as an error metric. The guidance strategy generation module uses a conditional generative adversarial network (CGAN). The discriminator D receives the correct action sequence and the user trajectory and evaluates the action quality through the Wasserstein distance . The generator G synthesizes three-dimensional holographic correction prompts according to the error distribution , where the latent variable z follows a normal distribution in the latent action space.

[0031] The multi-modal interaction interface integrates a tactile feedback array and calculates the expected tactile sensation based on the inverse dynamics model and generates corresponding vibration patterns through the linear resonant actuators (LRAs) of the Tactal haptic glove. The voice guidance system uses a WaveNet variational autoencoder to convert the religious scripture text T into a speech waveform with specific religious rhythms , where the latent variable learns the acoustic features of ritual language through KL divergence constraints . The user cognitive state tracking module fuses the eye tracker data and the galvanic skin response (GSR) and uses a factorial hidden Markov model (fHMM) to estimate the attention level . When is detected, the scene enhancement mechanism is triggered, and the visual salience of key ritual symbols is enhanced through non-photorealistic rendering (NPR) technology.

[0032] The data closed-loop system establishes a federated reinforcement learning mechanism. Each user session generates a trajectory , and updates the global policy network through importance sampling: ; where the baseline function is calculated by a double Q-network. To prevent cultural misuse, the constraint optimization module uses the logarithmic barrier function method to handle religious taboo conditions: ; The penalty coefficient μ decays exponentially with the number of training steps. This module is ultimately deployed in a lightweight browser through WebGL 3.0, and with the support of a cloud GPU cluster equipped with the TensorRT inference engine, an end-to-end 12ms ultra-low latency interactive experience is achieved.

[0033] Voice explanation module: During the user's ritual simulation process, relevant historical, cultural, and ritual knowledge explanations are provided to enhance the depth and interest of learning.

[0034] The construction of the voice explanation module relies on a multimodal context-aware architecture, and achieves intelligent explanation through the deep coupling of dynamic knowledge graph triggering and adaptive speech synthesis technology. The system deploys an event listener array in the Unity scene to capture user interaction objects in real time. The spatial coordinates of With operating status , extracting the visual features of objects through the pre-trained ResNet-3D model and embed it with the cultural semantics in the religious knowledge database The knowledge trigger engine uses an improved DualAttention mechanism to calculate the spatiotemporal context weights. , where the query matrix Q comes from the user's current operation sequence , the key-value matrix k is derived from the ritual stage division in the knowledge base , when the attention peak exceeds the threshold Activate the corresponding explanation node.

[0035] The speech content generation layer is built on the Transformer-XL architecture, and the input layer integrates three features: historical operation sequence , current artifact semantics and user awareness level Through the dynamic memory enhancement mechanism, a top-k sampling strategy (k=15) is used in the decoding stage to balance knowledge accuracy and language vividness, generating an explanation text that fits the current context. , where the vocabulary V is enriched with religious terms and contains the vector space projection of the proper nouns of a specific religion. The real-time prosody control system uses Prosody Transfer technology to extract the fundamental frequency contour from the original recording of the religious ceremony. and energy envelope , the text is converted through the WaveGrad vocoder Converted into a speech waveform with a sense of ritual solemnity , where the latent variable z learns the unique prosodic pattern of religious recitations through adversarial training.

[0036] The multi-level feedback system designs a time-varying gain controller based on the user's stay time. and operation accuracy rate Dynamically adjust the explanation depth: When it is detected that ; and when, trigger the refined explanation branch and insert the core concepts extracted by the LDA topic model for three-dimensional visualization assistance; When the user's operation accuracy rate for three consecutive times when, activate the cultural extension module, retrieve historical related events from the knowledge graph to generate interesting historical anecdotes. Spatial audio rendering uses Ambisonics technology, and calculates the binaural impulse response in real time according to the user's head posture to make the explanatory speech present the sense of direction spreading from the direction of the altar, enhancing the sense of ceremony presence.

[0037] To prevent information overload, the cognitive load balancing module implements closed-loop control based on pupil diameter and blink frequency : When the cognitive stress index estimated by the Kalman filter exceeds the threshold, the speech rate is reduced from 150 words per minute to 110 words per minute through a speech compression algorithm, and at the same time, a 500ms silent interval is inserted. The data-driven optimization layer deploys a deep reinforcement learning agent, and the reward function comprehensively considers the user's knowledge absorption rate (evaluated by subsequent quiz scores), continuous engagement and cognitive load , and continuously updates the explanatory strategy parameters through the proximal policy optimization (PPO) algorithm. This module realizes low-latency rendering on the browser side through the WebAudio API, and under the support of the edge computing node equipped with NVIDIA Riva speech service, realizes an end-to-end real-time voice interaction closed-loop of 97ms.

[0038] Data analysis module: Use artificial intelligence technology to collect and analyze the user's learning data to optimize the adaptive learning algorithm and further improve the teaching effect and user experience.

[0039] The data analysis module is built based on the federated learning framework and the multi-modal data fusion architecture, and realizes the dynamic optimization of teaching strategies through spatio-temporal correlation feature extraction and cognitive state modeling. The system deploys a distributed data collector on the edge computing node to capture the action trajectory sequence from the interaction module in real time , the semantic feature vector of the speech module , physiological signals and cognitive evaluation indicators , and constructs a multi-dimensional learning state tensor through timestamp alignment ; Among them, the mask matrix M uses the self-attention mechanism to dynamically weight the importance of different joint points. The data preprocessing layer applies an improved deep variational autoencoder (VAE) to achieve heterogeneous data normalization in the latent space and retains cross-modal correlation features through KL divergence constraints .

[0040] User portrait modeling uses a hierarchical temporal convolutional network (TCN). The bottom layer extracts short-term behavior patterns through dilated causal convolution , the middle layer applies a gated recurrent unit (GRU) to capture the evolution during the learning stage, and the top layer constructs a knowledge mastery map through a graph attention network (GAT) , where the nodes represent the knowledge points of ritual steps, and the edge weights reflect the transfer relationship between concepts. The adaptive optimization engine deploys the twin-stream deep deterministic policy gradient (DDPG) algorithm. The actor network outputs the personalized teaching parameter adjustment amount according to the current state , including the guidance intensity of the interaction module , the information density of the voice commentary , and the challenge difficulty coefficient . The critic network evaluates the policy improvement direction through , where the immediate reward combines the knowledge mastery , user satisfaction (evaluated through facial expression recognition), and cognitive fatigue . .

[0041] The swarm intelligence mining layer constructs a Gaussian process regression model (GPR), securely aggregates the encrypted features collected by federated learning on the parameter server, and models the ability distribution surface of the user group through a kernel function to identify the teaching bottleneck areas . .

[0042] Based on this, the course evolution module uses the neural architecture search (NAS) technology to explore the optimal teaching strategy combination in the discrete space through the ENAS controller , and the reward function ensures that the algorithm update improves the effect while maintaining cultural sensitivity

[0043] The real-time visualization dashboard uses t-SNE dimensionality reduction to project the high-dimensional learning trajectory onto Space, generate the population ability isosurface through the Marching Cubes algorithm, and apply the optical flow field technology to present the dynamic evolution of the learning path. The privacy protection mechanism implements a collaborative scheme of differential privacy (DP) and homomorphic encryption. Laplace noise is added when user data is uploaded, and the model update process satisfies - differential privacy constraints to ensure data anonymity is maintained even when parameters are leaked. This module realizes real-time processing of massive data through the PySpark distributed computing framework. On a computing cluster equipped with NVIDIA Morpheus intelligent data pipelines, it achieves a throughput of processing learning events per second, enabling the system to complete a full closed-loop from data collection to teaching strategy update within 300 ms.

[0044] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0045] It should be understood that the detailed description of the technical solutions of the present invention with the help of the preferred embodiments is illustrative rather than restrictive. Those of ordinary skill in the art can modify the technical solutions recorded in each embodiment based on reading the specification of the present invention, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. A virtual reality simulation teaching system for religious ceremonies, characterized in that, The system comprises: User interface: Select the religious ritual to be learned, check the learning progress, and take self-tests on the user interface; Adaptive religious ritual module: uses artificial intelligence algorithms to adjust the difficulty and progress of ritual simulations to provide more targeted guidance based on the user's learning status; Religious knowledge database: stores information on rituals of various religions, including the history, cultural background, ritual procedures and details of utensils used in the rituals; Interactive experience module: Users practice religious rituals in a virtual space and deepen their understanding and memory through actual operations; provide real-time guidance and suggestions based on user operation feedback information; Voice explanation module: During the ritual simulation process, users can learn about relevant history, culture, and ritual knowledge to enhance the depth and fun of learning. Data analysis module: collects and analyzes users’ learning data to optimize adaptive learning algorithms and further improve teaching effectiveness and user experience.

2. The religious ritual virtual reality simulation teaching system according to claim 1, wherein, The religious knowledge database adopts a multimodal data fusion architecture, integrates heterogeneous data sources such as digitized texts of religious classics, archaeological image materials, and ethnographic audio and video records through a distributed crawler system, and uses the BERT-BiLSTM-CRF hybrid model to extract entity relationships and identify time and space markers, artifact symbols, and taboo rules in the ritual process.

3. The religious ritual virtual reality simulation teaching system according to claim 1, wherein The adaptive religious ritual module decomposes religious rituals into spatiotemporal action sequences, symbol sets and meaning association matrices, collects user behavior data streams through a wearable sensor array, extracts motion features through Fourier descriptor transform, and inputs them into a double-layer LSTM network for real-time state encoding.

4. The religious ritual virtual reality simulation teaching system according to claim 1, wherein The data analysis module captures user action trajectory sequences, speech semantic feature vectors, physiological signals and cognitive evaluation indicators in real time through distributed data collectors, constructs a multi-dimensional learning state tensor, and applies an improved deep variational autoencoder for data preprocessing.

5. The religious ritual virtual reality simulation teaching system according to claim 1, characterized in that The interactive experience module calculates the spatiotemporal context weights through a knowledge trigger engine, activates corresponding explanation nodes, generates explanation text using the Transformer-XL architecture, and converts the text into a speech waveform with a sense of ritual solemnity through Prosody Transfer technology.

6. The religious ritual virtual reality simulation teaching system according to claim 1, wherein The user interface adopts a circular navigation layout, with a central floating holographic globe as the entrance to ritual selection. Users can browse the geographical distribution of different religions through gestures of zooming and rotating. After clicking on a specific area, a vertical waterfall flow menu pops up, dynamically loading a three-dimensional miniature model of the representative ritual of the region.

7. The religious ritual virtual reality simulation teaching system according to claim 5, characterized in that, The interactive experience module also includes a multi-level feedback system, which uses a time-varying gain controller to dynamically adjust the explanation depth according to the user's stay time and operation accuracy, triggering detailed explanation branches or cultural extension modules.

8. The religious ritual virtual reality simulation teaching system according to claim 4, characterized in that, The data analysis module also includes user portrait modeling, extracting short-term behavior patterns through a layered temporal convolutional network, applying a gated recurrent unit to capture the evolution of the learning stage, and constructing a knowledge mastery graph through a graph attention network.

Images