Digital media immersive education interaction system

Abstract

The invention relates to the technical field of computer application, and particularly discloses a digital media immersive education interaction system. The system comprises an immersive interaction terminal, a space cooperation server, a process analysis server and a teacher guiding terminal. The space collaboration server processes collaboration interaction through conflict arbitration and space guidance prompt, so that the sense of presence of space is enhanced; the process analysis server records and analyzes the interaction log in multiple dimensions, and generates a structured evaluation report; and the teacher guiding terminal generates a personalized guiding instruction according to the report. According to the invention, a closed loop of deep space cooperation support and teaching process intelligent evaluation guidance is realized. A mechanism supporting deep space cooperation is constructed through cooperative work of the space cooperation server, the conflict arbiter of the space cooperation server and the space presence enhancement module.

Description

Technical Field

[0001] This invention belongs to the field of computer application technology, and specifically relates to a digital media immersive educational interactive system. Background Technology

[0002] Digital media technology, as an important branch of the information age, is profoundly changing the way knowledge is disseminated and skills are acquired. Its core lies in building an efficient information transmission and cognitive construction environment through digital audiovisual and interactive means.

[0003] Immersive educational interactive systems aim to use technologies such as virtual reality and augmented reality to create highly realistic and deeply engaging learning environments for learners, thereby improving the efficiency of knowledge comprehension and skill mastery.

[0004] Existing technologies typically achieve immersive learning by constructing 3D virtual scenes and providing basic interactive functions. However, these systems face significant challenges in supporting collaborative learning in small groups: First, the online collaboration functions provided by these systems are mostly limited to voice communication and simple object manipulation, lacking the sense of presence and spatial collaborative perception required for collaborative exploration and construction of complex 3D concepts, resulting in insufficient depth of collaboration; second, the systems struggle to structurally record and analyze the dynamic interaction processes of multiple members within a group, making it impossible for teachers to objectively assess the collaboration quality and cognitive progress of each group in real time, thus hindering the implementation of effective personalized guidance and intervention.

[0005] Therefore, how to build a digital media immersive educational interactive system that can support deep spatial collaboration and realize intelligent assessment and guidance of group learning processes has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] This invention provides a digital media immersive educational interactive system to address the technical contradictions in existing immersive educational systems, such as insufficient presence and spatial collaborative perception capabilities when supporting group collaboration, and a lack of structured recording and intelligent evaluation capabilities for dynamic group interaction processes.

[0007] The technical solution of the present invention is a digital media immersive educational interactive system, the system comprising an immersive interactive terminal, a spatial collaboration server, a process analysis server, and a teacher guidance terminal.

[0008] An immersive interactive terminal, deployed at each learner's location, generates and presents a shared immersive 3D learning scene while simultaneously collecting learners' multimodal interaction data. The immersive interactive terminal includes a scene rendering engine, a multimodal perception module, and a local interaction agent. The scene rendering engine renders a visually and aurally consistent 3D virtual environment locally based on unified scene state data issued by the spatial collaboration server. The multimodal perception module continuously collects learners' head pose data, hand movement data, eye-tracking data, and speech data. The local interaction agent processes learners' direct operation commands and encapsulates the operation intentions into atomic interaction events, which are then uploaded to the spatial collaboration server along with the multimodal perception data.

[0009] The spatial collaboration server, as the core coordination hub of the system, maintains a globally consistent 3D scene state and processes all collaborative interaction requests from immersive interactive terminals, enabling deep spatial collaboration. The spatial collaboration server includes a scene state manager, a conflict arbitrator, and a spatial presence enhancement module. The scene state manager maintains a global 3D scene data model, organized in a scene graph structure, recording the spatial pose, physical attributes, and logical state of all interactive objects in the scene. The conflict arbitrator receives atomic interaction events from different immersive interactive terminals; these events include the target object identifier, operation type, and the initiator's spatial perspective information. The conflict arbitrator processes concurrent events according to preset arbitration rules based on spatiotemporal consistency.

[0010] For non-conflict events, the global scene data model is directly updated and the state change is broadcast; for conflict events, a collaborative resolution process is initiated. The spatial presence enhancement module receives the collaborative events output by the conflict arbitrator and the spatial perspective information of relevant learners, and generates spatial guidance prompts. The global scene data model visualizes and makes audible the operational intentions and focus of attention of other collaborators in three-dimensional space by calculating and overlaying virtual gaze indicators, highlighting focus areas, and generating spatial audio prompts, thereby constructing a shared attention space.

[0011] A process analysis server receives and stores group interaction logs transferred from the spatial collaboration server, and performs multi-dimensional analysis on these logs to generate a structured evaluation report on group collaboration quality and cognitive progress. The process analysis server includes a process log database, a multi-dimensional analysis engine, and an evaluation report generator. The process log database stores the complete interaction process of each group in time series format. Each record includes a timestamp, interaction event type, initiator identifier, target object identifier, snapshots of the object's state before and after the event, and relevant spatial presence augmentation data.

[0012] The multi-dimensional analysis engine performs real-time streaming analysis and periodic batch processing analysis of process logs based on a pre-defined analysis dimension model. The analysis dimension model includes at least three dimensions: social interaction, cognitive operation, and spatial collaboration. The social interaction dimension quantifies the group's communication efficiency and participation balance by analyzing voice activity patterns, interactive event response chains, and changes in spatial distance between virtual avatars. The cognitive operation dimension tracks the group's cognitive construction process of learning objectives by identifying the sequence of operations on key conceptual objects, operational complexity, and the correctness of operational results. The spatial collaboration dimension evaluates the group's spatial collaboration effectiveness by analyzing the trigger frequency, acceptance rate, and average time spent collaboratively resolving conflicts related to spatial guidance prompts. The evaluation report generator receives the quantitative indicators for each dimension output by the multi-dimensional analysis engine and, based on pre-defined evaluation thresholds and weighting models, generates a structured evaluation report including a comprehensive score, dimension-specific scores, key event retrospectives, and trend analysis.

[0013] The teacher guidance terminal provides teachers with a system management interface and real-time monitoring dashboard, and receives structured assessment reports pushed by the process analysis server, thereby generating and issuing personalized guidance and intervention instructions. The teacher guidance terminal includes a panoramic monitoring module, a report parsing module, and an intervention instruction generation module. The panoramic monitoring module obtains a global scene thumbnail view and real-time activity heatmap of each group from the spatial collaboration server, allowing teachers to have a macro-level understanding and control.

[0014] The report parsing module receives structured assessment reports and uses natural language processing technology to extract key risk points and outstanding progress from the reports. The intervention instruction generation module generates specific guidance instructions based on the parsing results and a pre-set intervention strategy library. Guidance instructions are divided into system-level instructions and content-level instructions. System-level instructions are directly sent to the spatial collaboration server to dynamically adjust the group's collaboration rules or resource permissions. Content-level instructions are transformed into specific spatial guidance prompts or learning resource pushes through the spatial collaboration server and directly embedded into the target group's immersive scenario.

[0015] Furthermore, the specific execution process of the arbitration rules based on spatiotemporal consistency, upon which the conflict arbitrator operates, is as follows: First, each atomic interaction event is appended with a timestamp accurate to the millisecond level and the three-dimensional spatial coordinates of the event initiator. When multiple events are detected targeting the same object, the conflict arbitrator compares the timestamps of each event and determines the event with the earliest timestamp as the priority event. If the difference in timestamps among multiple events is less than a preset threshold of 50 milliseconds, it is determined to be a concurrent conflict event. For concurrent conflict events, the conflict arbitrator calculates the spatial distance between each event initiator and the target object and sets the event corresponding to the event initiator with the closest spatial distance as the initial solution. Simultaneously, all relevant information of the concurrent conflict event, including the spatial perspectives of each alternative solution and its initiator, is packaged into a collaborative task package and sent to the spatial presence enhancement module.

[0016] Furthermore, the specific method for the spatial presence enhancement module to generate spatial guidance prompts is as follows: After receiving the collaborative task package, the spatial presence enhancement module generates a virtual, semi-transparent 3D indicator arrow for each alternative solution in the task package. The 3D indicator arrow originates from the perspective of the virtual avatar of the solution initiator and points to a specific operation point on the target object. At the same time, a pulsating bright halo is generated around the operation point of the target object. In addition, according to the location of the initiator of each alternative solution, a voice segment with spatial effects is generated in the 3D scene. The content of the voice segment is a brief intention description and played from the virtual sound source at the location of its corresponding initiator. All visual and auditory prompts are only presented to the team members participating in the collaborative task.

[0017] Furthermore, the cognitive operation dimension analysis process in the multi-dimensional analysis engine adopts a cognitive state modeling method based on operation chains. This cognitive state modeling method first extracts all operation events for predefined key concept objects from the process log and arranges them in chronological order to form an operation chain. Then, it encodes each event in the operation chain using a preset cognitive operation coding table, with coding categories including exploratory operations, verification operations, constructive operations, and erroneous operations. Next, it applies a state transition probability model to analyze the transition patterns between operations of different coding categories and calculates the continuous length and proportion of constructive operation sequences. Finally, combined with the verification of the correctness of the operation results, the operation chain is mapped to discrete cognitive progress states, with state levels divided into initial contact, concept exploration, relationship construction, and proficient application.

[0018] Furthermore, the intervention strategy library in the intervention instruction generation module is constructed using a hybrid rule-and-case-driven approach. The rules are pre-defined by domain education experts, defining standard intervention actions to be triggered when a specific combination of assessment indicators falls below or exceeds a certain threshold. The case studies, on the other hand, use machine learning methods to learn the mapping relationship between assessment report patterns and corresponding effective intervention instructions from historical successful intervention cases. When the report parsing module outputs key risk points, the intervention instruction generation module first matches them against the rule library. If no perfectly matching rule is found, the case reasoning engine is activated to find the three historical cases with the most similar assessment patterns in the case library, and their intervention instructions are weighted and fused to generate new personalized guidance instructions.

[0019] Furthermore, the system operates within a layered synchronization architecture, which includes an event layer, a state layer, and an evaluation layer. The event layer handles high-frequency atomic interaction events and conflict arbitration, ensuring real-time and smooth interaction; its synchronization cycle is 100 milliseconds. The state layer maintains and synchronizes the global 3D scene data model; its state synchronization cycle is 500 milliseconds, ensuring consistency in the performance of all terminal scenes. The evaluation layer performs multi-dimensional analysis and report generation; its analysis cycle can be dynamically configured according to the teaching stage, ranging from 1 minute to 10 minutes, enabling monitoring and evaluation of the collaboration process at different granularities.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs a mechanism supporting deep spatial collaboration through the collaborative work of a spatial collaboration server, its conflict arbitrator, and a spatial presence enhancement module. The conflict arbitrator processes concurrent interactions based on spatiotemporally consistent rules, transforming conflicts into tasks requiring collaborative resolution. The spatial presence enhancement module then transforms abstract conflict intentions into visual and audible guidance prompts in three-dimensional space. This technical approach fundamentally enhances the spatial collaborative perception capabilities among team members, enabling collaborative discussions and operations of complex three-dimensional concepts to possess a realistic sense of presence and direction, effectively solving the problem of insufficient collaboration depth in existing systems.

[0021] 2. This invention utilizes a process analysis server to perform full log recording and multi-dimensional quantitative analysis of group interactions, enabling structured and objective evaluation of the learning process. The three-dimensional analysis model—social interaction, cognitive operation, and spatial coordination—can accurately depict the quality of group collaboration and cognitive progress from multiple perspectives, including communication behavior, cognitive construction, and spatial cooperation. The evaluation report not only provides a comprehensive score but also includes traceable key events and trend analyses, allowing teachers to move beyond superficial activity observations and gain insight into the micro-processes of group dynamics and cognitive development, providing a solid data foundation for precise intervention.

[0022] 3. This invention achieves an intelligent closed loop of assessment and intervention through a teacher-guided terminal. The system can automatically analyze assessment reports, identify risks and progress, and generate personalized guidance instructions based on a hybrid-driven intervention strategy library. These instructions can be directly converted into system rule adjustments or embedded spatial guidance and resource pushes within immersive scenarios, achieving a seamless connection from process analysis to real-time guidance. This frees teachers from the burden of tedious process monitoring and preliminary judgment, allowing them to focus on higher-level instructional design and personalized guidance, significantly improving the timeliness and effectiveness of teaching interventions, and truly realizing intelligent and personalized collaborative learning support. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention; Figure 2 This is a schematic diagram of the core principle framework of the collaborative work between conflict arbitration and spatial presence enhancement based on spatiotemporal consistency in this invention; Figure 3 This is a flowchart illustrating the logical flow of the multi-dimensional analysis engine in this invention analyzing the group interaction process. Figure 4 This is a schematic diagram of the hierarchical synchronization architecture of the event layer, state layer and evaluation layer in this invention; Figure 5 This is a logical flowchart of the intelligent closed loop of assessment and intervention implemented by the teacher guidance terminal in this invention. Detailed Implementation

[0024] Example 1: Please refer to the appendix Figure 1 To be continued Figure 5 The digital media immersive educational interactive system proposed in this invention consists of four core components: an immersive interactive terminal, a spatial collaboration server, a process analysis server, and a teacher guidance terminal. These components are connected via a high-speed, low-latency network, forming a closed-loop immersive educational environment that supports real-time collaboration and intelligent assessment. The system's operational logic strictly follows a hierarchical synchronous architecture of event layer, state layer, and assessment layer, as shown in the attached diagram. Figure 4 As shown, this ensures that the system maintains stability, consistency, and timeliness of teaching interventions even in high-concurrency interactive scenarios.

[0025] First, the immersive interactive terminal is deployed at each learner's location, serving as the sole entry point into the immersive 3D learning environment. This terminal comprises three main functional modules: a scene rendering engine, a multimodal perception module, and a local interaction agent. The scene rendering engine receives unified scene state data from a spatial collaboration server. This data, organized in a scene graph structure, precisely describes the spatial pose, physical properties (such as rigid body mass, friction coefficient, and collision volume), and logical states (such as on / off status and content loading progress) of all interactive objects in the 3D virtual environment. Based on this data, the scene rendering engine locally constructs and updates a visually and aurally consistent 3D virtual environment in real time. Visual rendering employs a physically based lighting model and anti-aliasing technology, maintaining a stable frame rate above 90 Hz. Auditory rendering utilizes the HRTF (Head-Related Transfer Function) algorithm to achieve spatial audio localization, ensuring a strict correspondence between the sound source and its position in 3D space.

[0026] The multimodal perception module continuously collects four types of key interaction data from learners: head pose data, hand motion data, eye tracking data, and speech data. Head pose data is acquired through the fusion of a six-DOF inertial measurement unit and an optical tracking system, with a sampling frequency of 250 Hz and an accuracy better than 0.1 mm and 0.1 degrees, respectively. Hand motion data is collected through joint angle sensors and accelerometers embedded in the data glove or controller, represented as skeletal animation, with each hand containing 21 degrees of freedom. Eye tracking data is captured by an infrared camera array, capturing the pupil center and corneal reflection point to calculate the gaze direction and gaze point, with a sampling rate of 120 Hz and a spatial resolution of 0.5 degrees. Speech data is collected by a directional microphone array, and after local noise reduction and speech activity detection, it is encoded at a sampling rate of 16 kHz and a depth of 16 bits. All multimodal data includes nanosecond-level timestamps generated by a local high-precision clock and are strictly aligned with the current rendering frame.

[0027] The local interaction agent handles direct operation commands issued by learners via controllers, gestures, or voice. The local interaction agent parses the raw input into atomic interaction events with clear semantics. Each atomic interaction event includes the following fields: event type (e.g., grab, rotate, scale, label, ask), globally unique identifier of the target object, operation parameters (e.g., rotation axis vector, scaling factor), initiator learner identifier, event generation timestamp (milliseconds), and the initiator's current 3D spatial coordinates and orientation. The local interaction agent packages the atomic interaction events with synchronously acquired multimodal perception data and uploads them to the spatial collaboration server via UDP in a non-blocking manner, with transmission latency controlled within 20 milliseconds.

[0028] The spatial collaboration server serves as the core coordination hub of the system, and its internal structure is shown in the attached figure. Figure 2As shown, the system includes a scene state manager, a conflict arbitrator, and a spatial presence enhancement module. The scene state manager maintains a globally unique 3D scene data model, organized as a directed acyclic graph (DAG). Each node represents a scene object, and edges represent parent-child hierarchies or spatial constraints. The model uses a version number mechanism for state tracking; each valid state change generates a new version number and records a change log. The scene state manager broadcasts the latest version of the scene state data to all immersive interactive terminals every 500 milliseconds, ensuring that the virtual environment presented on each terminal remains macroscopically consistent.

[0029] The conflict arbitrator monitors the atomic interaction event streams from various immersive interactive terminals in real time. When multiple events targeting the same object are received, an arbitration rule based on spatiotemporal consistency is initiated. This arbitration rule first extracts the timestamp and initiator spatial coordinates of each event. If the difference between the timestamps of any two events is less than 50 milliseconds, it is determined to be a concurrent conflict event. For non-conflict events, the conflict arbitrator directly applies its operation to the global scene data model and marks the object as "updated," awaiting the next state synchronization cycle broadcast. For concurrent conflict events, the conflict arbitrator calculates the Euclidean distance between each event initiator and the target object: in, For the first The spatial coordinates of the event initiator The coordinates are the centroid coordinates of the target object. The event with the smallest distance is tentatively designated as the initial solution, but not executed immediately. The conflict arbitrator encapsulates all concurrent events and their corresponding initiator's perspective matrix, operation intent summary, spatial coordinates, and other information into a collaborative task package, and pushes it to the spatial presence enhancement module.

[0030] After receiving the collaborative task package, the spatial presence enhancement module generates multi-channel spatial guidance prompts for each alternative solution in the task package. Specifically, the spatial presence enhancement module generates a semi-transparent blue indicator arrow for each solution in the 3D scene. The arrow's starting point is located at the eye position of the solution initiator's virtual avatar, and its ending point points to a specific operation point on the target object (such as the center of a knob or the surface of a button). The arrow is implemented using a particle system, exhibiting a slight pulsating effect, and its transparency is set to 70% to avoid obscuring key content.

[0031] A bright, ring-shaped area with a diameter of 30 cm is generated around the operation point of the target object. The area is yellow, and its brightness modulates sinusoidally over time at a frequency of 1 Hz. Furthermore, based on the relative positions of each initiator in three-dimensional space, the system uses binaural audio synthesis technology to generate short voice prompts, such as "I suggest rotating clockwise" or "This needs to be marked." This voice prompt is played from the virtual sound source location of the corresponding initiator and is only audible to team members participating in the collaborative task. The volume automatically adjusts to the ambient background noise level. All prompts automatically disappear after the collaborative task is completed or after a timeout (default 15 seconds).

[0032] The process analysis server is responsible for the structured recording and intelligent analysis of the group's interaction process. Its logical flow is shown in the attached figure. Figure 3 As shown. The process analysis server includes a process log database, a multi-dimensional analysis engine, and an evaluation report generator. The process log database adopts a time-series database architecture, storing interaction logs by team. Each log record includes: a timestamp accurate to milliseconds, event type, initiator identifier, target object identifier, object state snapshot before the event (including pose, attributes, and logical state), object state snapshot after the event, whether spatial presence enhancement was triggered, and the type and acceptance status of enhancement prompts (e.g., whether other members followed the operation). The log write latency is less than 10 milliseconds, supporting a throughput of tens of thousands of events per second.

[0033] The multi-dimensional analysis engine performs dual analysis on logs based on a pre-defined analysis dimension model: real-time streaming analysis for immediate feedback and periodic batch analysis for in-depth evaluation. The analysis dimension model includes three orthogonal dimensions: social interaction, cognitive operation, and spatial collaboration. The social interaction dimension quantifies communication efficiency and participation balance by calculating metrics such as the overlap rate of voice activities, the median response delay of interactive events, and the rate of change in the average distance between virtual avatars. For example, if a member does not initiate any interactive events for 5 consecutive minutes and their average distance from other members is greater than 3 meters, they are marked as having "low participation."

[0034] The cognitive operation dimension employs a cognitive state modeling method based on operation chains. This method first filters all operation events targeting predefined key concept objects (such as molecular structure models, circuit components, and historical maps) from the logs and sorts them chronologically to form operation chains. Then, a cognitive operation coding table is used to classify and encode each event, with categories including: exploratory operations (such as rotating observation, zooming in), verification operations (such as dragging and dropping test connections, clicking to verify answers), constructive operations (such as assembling parts, drawing relationship diagrams), and error-prone operations (such as incorrect connections, deleting key components). Next, a Markov state transition matrix is ​​constructed to calculate the transition probability from exploratory operations to constructive operations. Finally, combining the correctness of the operation results (verified by the built-in knowledge graph), the entire operation chain is mapped to discrete cognitive progress states: initial contact (only exploratory operations), concept exploration (including verification operations but construction failure), relationship construction (successfully completing at least one constructive operation sequence), and proficient application (three or more consecutive correct constructions without error operations).

[0035] The spatial collaboration dimension focuses on collaboration efficiency indicators, including the trigger frequency of spatial guidance prompts (number of times per minute), acceptance rate (the proportion of other members performing the same or similar operations within the next 5 seconds), and the average time spent collaboratively resolving conflicts. If the acceptance rate is consistently below 30% and the conflict resolution time exceeds 20 seconds, it is judged as low spatial collaboration efficiency.

[0036] The assessment report generator integrates quantitative indicators from three dimensions and generates a structured assessment report based on preset assessment thresholds and weighting models (e.g., social interaction accounts for 30%, cognitive operation accounts for 50%, and spatial collaboration accounts for 20%). The report includes a comprehensive score (0-100 points), sub-scores for each dimension, key event retrospectives (e.g., "10:15:23, the group successfully completed DNA double helix assembly"), risk warnings (e.g., "Member A did not participate in the interaction for 8 consecutive minutes"), and trend analysis charts (e.g., the evolution curve of cognitive state over time).

[0037] The teacher guidance terminal provides teachers with panoramic monitoring and intelligent intervention capabilities, and its closed-loop logic is shown in the attached figure. Figure 5 As shown in the diagram, the terminal includes a panoramic monitoring module, a report parsing module, and an intervention instruction generation module. The panoramic monitoring module subscribes to global scene thumbnails of each group from the spatial collaboration server, displaying the distribution of virtual avatars and activity heatmaps of all groups from a bird's-eye view. The heatmap colors gradient from blue (low activity) to red (high activity), with a refresh rate of 1 Hz. Teachers can click on any group's view to switch to that group's real-time first-person or third-person perspective for in-depth observation.

[0038] The report parsing module receives structured evaluation reports pushed by the process analysis server and extracts key information using natural language processing technology. The report parsing module has a built-in named entity recognition model, which can identify learner names, key concept object names, time points, and risk types in the report. For example, it parses "Member B incorrectly connected the power supply polarity 3 times consecutively in the circuit building task" as: Subject = Member B, Task = Circuit building, Error type = Reverse power supply polarity, Frequency = 3 times.

[0039] The intervention instruction generation module generates guidance instructions based on the analysis results and a hybrid-driven intervention strategy library. The strategy library includes rules and cases. The rules are preset by education experts, for example: "If the cognitive operation dimension score is below 60 and the percentage of incorrect operations is above 40%, then push basic concept review resources." The cases are trained from historical intervention data through supervised learning. The feature vectors include the values ​​of each dimension's indicators, the teaching stage, group size, etc., and the labels are valid intervention instruction texts. When the report analysis module outputs risk points, the system first matches them in the rule library. If no match is found, the case reasoning engine is activated to calculate the cosine similarity between the current assessment mode and historical cases. The three cases with the highest similarity are selected, and their intervention instructions are weighted and averaged (weight is proportional to similarity) to generate new personalized instructions.

[0040] The generated guidance instructions are divided into two categories: system-level instructions and content-level instructions. System-level instructions, such as "elevate member C's operation privileges" or "lock the current scene to prevent accidental operation," are directly issued to the spatial collaboration server, which modifies the access control list of scene objects or freezes the physics engine. Content-level instructions, such as "display power polarity indicators next to the circuit board" or "play a 5-second audio prompt about Ohm's Law," are transformed into specific spatial elements by the spatial collaboration server: the former generates a floating panel with text labels, anchored near the target object; the latter generates a spatial audio clip, played at a specified location. All instructions have an expiration date and scope of application to ensure the accuracy and timeliness of the intervention.

[0041] The entire system operates efficiently under a layered synchronous architecture. The event layer handles atomic interaction events and conflict arbitration, with a synchronization cycle of 100 milliseconds to ensure immediate response to interactions; the state layer maintains global scene consistency, with a synchronization cycle of 500 milliseconds to balance network load and visual continuity; the evaluation layer executes analysis tasks, with a configurable cycle of 1 minute (for classroom exercises) to 10 minutes (for project exploration), adapting to different teaching paces. The three layers are decoupled through message queues, ensuring independent scaling of each layer under high load.

[0042] In summary, this embodiment constructs a digital media immersive educational interactive system with strong presence, high collaboration, and deep insight by employing multimodal data acquisition from an immersive interactive terminal, spatiotemporal consistency arbitration and presence enhancement from a spatial collaboration server, multidimensional structured evaluation from a process analysis server, and an intelligent intervention closed loop from a teacher-guided terminal. This digital media immersive educational interactive system not only solves the core problems of insufficient collaboration depth and lack of process evaluation in existing technologies, but also achieves full-process automation from data acquisition and intelligent analysis to precise intervention, providing solid technical support for future intelligent collaborative learning.

[0043] Example 2: Based on the previous examples, this example further optimizes the prompt generation mechanism of the spatial presence enhancement module and the cognitive modeling method of the multi-dimensional analysis engine to adapt to more complex learning tasks and more refined teaching needs.

[0044] In terms of spatial presence enhancement, this embodiment introduces a dynamic prompt intensity adjustment mechanism based on intent confidence. When the conflict arbitrator generates a collaborative task package, in addition to including each alternative solution, it also includes an intent confidence score calculated by the local interaction agent. This intent confidence score is derived from multimodal perception data fusion: if the initiator has a prolonged gaze at the target object (>2 seconds), stable hand movements (acceleration standard deviation <0.5m / s²), and clear voice instructions (voice recognition confidence >90%) before proposing the operation, then the confidence score is high. The spatial presence enhancement module dynamically adjusts the visual and auditory intensity of the prompts based on this score. For example, the high-confidence solution uses a green arrow with increased transparency to 85%, and the voice prompt volume increases by 3 decibels; the low-confidence solution uses an orange arrow with reduced transparency to 50%, and the voice prompt is presented only as a text bubble. This effectively guides team members to prioritize high-reliability proposals, improving collaborative decision-making efficiency.

[0045] In terms of cognitive operation dimension analysis, this embodiment expands the original discrete state model into a continuous cognitive progress trajectory model. This continuous cognitive progress trajectory model no longer relies on preset state levels, but instead maps the operation chain to a two-dimensional cognitive plane: the horizontal axis represents operation complexity (calculated based on the entropy of the number of objects involved in the operation and the combination of operation types), and the vertical axis represents operation accuracy (based on knowledge graph verification results). Each operation event is projected as a point on the plane, and the entire operation chain forms a trajectory. The multi-dimensional analysis engine calculates the slope, curvature, and endpoint position of this trajectory, thereby more precisely depicting the cognitive development path. For example, a trajectory with a large slope and small curvature indicates that the learner quickly transitions from simple operations to complex and correct operations, demonstrating efficient cognitive progress; while a trajectory that oscillates repeatedly in the low accuracy region indicates conceptual confusion. The evaluation report generator then generates a dynamic cognitive heatmap, visually displaying the evolution of the group's cognitive state.

[0046] Furthermore, this embodiment enhances the diversity of intervention instructions on the teacher-guided terminal. A new "scaffolding guidance" strategy has been added to the intervention strategy library. When the system detects that a group has stalled for more than a preset threshold (e.g., 3 minutes) at a critical step, the intervention instruction generation module no longer provides the answer directly, but instead generates a series of progressive prompts. For example, in a geometry proof task, the graphical representation of the relevant theorem is first highlighted in the scene. If no progress is made after one minute, the textual description of the theorem is displayed. If the problem is still not solved after another minute, the location of the auxiliary line to be used is suggested. This step-by-step information release strategy provides necessary support while preserving learners' exploration space, aligning with constructivist teaching principles.

[0047] The above optimization measures significantly improve the system's adaptability and teaching effectiveness while maintaining the core system architecture, making it particularly suitable for complex inquiry-based learning scenarios at the high school level and above.

Claims

1. A digital media immersive educational interactive system, characterized in that, include: An immersive interactive terminal, deployed at each learner's location, is used to generate and present a shared immersive 3D learning scene and collect learners' multimodal interaction data. The spatial collaboration server is used to maintain a globally consistent 3D scene state and handle all collaborative interaction requests from immersive interactive terminals. The process analysis server is used to receive and store the group interaction process logs flowing from the spatial collaboration server, and to perform multi-dimensional analysis on the group interaction process logs to generate a structured evaluation report on the group collaboration quality and cognitive progress. The teacher guidance terminal provides teachers with a system management interface and real-time monitoring dashboard, and receives structured evaluation reports pushed by the process analysis server, thereby generating and issuing personalized guidance and intervention instructions.

2. The digital media immersive educational interactive system according to claim 1, wherein the immersive interactive terminal includes a scene rendering engine, a multimodal perception module, and a local interactive agent; The scene rendering engine renders a visually and aurally consistent 3D virtual environment locally based on the unified scene state data issued by the spatial collaboration server. The multimodal perception module continuously collects learners' head pose data, hand movement data, eye tracking data, and speech data; The local interaction agent processes the learner's direct operation commands and encapsulates the operation intentions into atomic interaction events, which are then uploaded to the spatial collaboration server along with the multimodal perception data.

3. The digital media immersive educational interactive system according to claim 1, wherein the spatial collaboration server includes a scene state manager, a conflict arbitrator, and a spatial presence enhancement module; The scene state manager maintains a global 3D scene data model organized in a scene graph structure. The global 3D scene data model records the spatial pose, physical attributes and logical state of all interactive objects in the scene. The conflict arbitrator receives atomic interaction events from different immersive interactive terminals. The atomic interaction events include the target object identifier, the operation type, and the spatial perspective information of the initiator. The conflict arbitrator processes concurrent events according to preset arbitration rules based on spatiotemporal consistency. For non-conflict events, it directly updates the global scene data model and broadcasts the state change. For conflict events, it initiates a collaborative resolution process. The spatial presence enhancement module receives the events to be coordinated output by the conflict arbitrator and the spatial perspective information of the relevant learners, and generates spatial guidance prompts. The spatial guidance prompts visualize and make audible the operational intentions and focus of attention of other collaborators in three-dimensional space by calculating and overlaying virtual line-of-sight indicators, highlighting focus areas, and generating spatial audio prompts.

4. The digital media immersive educational interactive system according to claim 1, wherein the process analysis server includes a process log database, a multi-dimensional analysis engine, and an evaluation report generator; The process log database stores the complete interaction process of each group in time series. Each record includes a timestamp, interaction event type, initiator identifier, target object identifier, object state snapshot before and after the event, and related spatial presence augmentation data. The multi-dimensional analysis engine performs real-time streaming analysis and periodic batch processing analysis on process logs based on a preset analysis dimension model. The analysis dimension model includes at least the social interaction dimension, the cognitive operation dimension, and the spatial collaboration dimension. The social interaction dimension quantifies the group's communication efficiency and participation balance by analyzing voice activity patterns, interactive event response chains, and changes in spatial distance between virtual avatars. The cognitive operation dimension tracks the group's cognitive construction process of the learning objectives by identifying the operation sequence, operation complexity, and operation result correctness of key concept objects. The spatial collaboration dimension evaluates the team's spatial collaboration effectiveness by analyzing the trigger frequency, acceptance rate, and average time spent collaboratively resolving conflicts in spatial guidance prompts. The evaluation report generator receives quantitative indicators for each dimension from the multi-dimensional analysis engine and generates a structured evaluation report that includes a comprehensive score, dimension-specific scores, key event retrospectives, and trend analysis based on preset evaluation thresholds and weight models.

5. The digital media immersive educational interactive system according to claim 1, wherein the teacher guidance terminal includes a panoramic monitoring module, a report parsing module, and an intervention instruction generation module; The panoramic monitoring module obtains a global scene thumbnail view and real-time activity heat map of each group from the spatial collaboration server. The report parsing module receives the structured assessment report and uses natural language processing technology to extract key risk points and outstanding progress from the report; The intervention instruction generation module generates specific guidance instructions based on the parsing results and a preset intervention strategy library; The guidance instructions are divided into system-level instructions and content-level instructions; the system-level instructions are directly sent to the space collaboration server to dynamically adjust the group's collaboration rules or resource permissions. The content-level instructions are transformed into specific spatial guidance prompts or learning resource pushes through the spatial collaboration server and directly embedded into the immersive scene of the target group.

6. The digital media immersive educational interactive system according to claim 1, characterized in that, The specific execution process of the arbitration rule based on spatiotemporal consistency is as follows: Attach a timestamp accurate to milliseconds and the three-dimensional spatial coordinates of the event initiator to each atomic interaction event; When multiple events are detected targeting the same object, the conflict arbitrator compares the timestamps of each event and determines the event with the earliest timestamp as the priority event. If the timestamp difference of multiple events is less than the preset threshold of 50 milliseconds, it is determined to be a concurrent conflict event; For concurrent conflict events, the conflict arbitrator calculates the spatial distance between each event initiator and the target object, and sets the event corresponding to the event initiator with the closest spatial distance as the initial solution; All relevant information regarding concurrent conflict events, including the spatial perspectives of each alternative and its initiator, is packaged into a collaborative task package and sent to the spatial presence enhancement module.

7. A digital media immersive educational interactive system according to claim 2, characterized in that, The specific method by which the spatial presence enhancement module generates spatial guidance prompts is as follows: After receiving the collaborative task package, the spatial presence enhancement module generates a virtual, semi-transparent 3D indicator arrow for each alternative solution in the task package. The arrow is emitted from the perspective of the virtual avatar of the solution initiator and points to a specific operation point on the target object. Generate a pulsating, bright halo around the operation point of the target object; In addition, based on the location of the initiator of each alternative plan, a voice segment with spatial effect is generated in the three-dimensional scene. The content of the voice segment is a brief intention statement, which is played from the virtual sound source at the location of its corresponding initiator. All visual and auditory cues are presented only to team members participating in the collaborative task.

8. The digital media immersive educational interactive system according to claim 1, characterized in that, The analysis process of the cognitive operation dimension in the multi-dimensional analysis engine adopts a cognitive state modeling method based on operation chains. The cognitive state modeling method first extracts all operation events for predefined key concept objects from the process log and arranges them in chronological order to form an operation chain; Each event in the operation chain is encoded using a pre-defined cognitive operation encoding table. The encoding categories include exploratory operations, confirmatory operations, constructive operations, and error-prone operations. Next, the state transition probability model is applied to analyze the transition patterns between different coding categories of operations, and the continuous length and proportion of the constructive operation sequence are calculated; By verifying the correctness of the operation results, the operation chain is mapped to discrete cognitive progress states, with state levels divided into initial contact, concept exploration, relationship construction, and proficient application.

9. A digital media immersive educational interactive system according to claim 1, characterized in that, The intervention strategy library is constructed using a hybrid rule-and-case approach. The rules section is pre-set by domain education experts, defining the standard intervention actions that should be triggered when a specific combination of assessment indicators is below or above a certain threshold; The case study section uses machine learning methods to learn the mapping relationship between assessment report patterns and corresponding effective intervention instructions from historical successful intervention cases; When the report parsing module outputs key risk points, the intervention instruction generation module first performs matching in the rule base. If there is no completely matching rule, the case reasoning engine is activated to find the three historical cases with the most similar evaluation patterns in the case base, and the intervention instructions of these cases are weighted and fused to generate new personalized guidance instructions.

10. A digital media immersive educational interactive system according to claim 1, characterized in that, The system operates on a layered synchronization architecture, which includes an event layer, a state layer, and an evaluation layer. The event layer is responsible for handling high-frequency atomic interaction events and conflict arbitration to ensure the real-time and smoothness of the interaction, with a synchronization period of 100 milliseconds. The state layer is responsible for maintaining and synchronizing the global 3D scene data model. Its state synchronization cycle is 500 milliseconds to ensure the consistency of scene performance across all terminals. The evaluation layer is responsible for performing multi-dimensional analysis and report generation. Its analysis cycle can be dynamically configured according to the teaching stage, ranging from 1 minute to 10 minutes, to achieve monitoring and evaluation of different granularities of the collaborative process.

Images