Edge data transmission method and system for metaverse scene synchronous interaction

Abstract

The present application relates to the technical field of metaverse scene synchronous interaction and edge computing communication, in particular to an edge data transmission method and system for metaverse scene synchronous interaction; containing pose collection, state determination, mode switching and pose reconstruction modules; the system performs frequency spectrum and dynamics analysis by collecting continuous pose data of users and virtual interactive objects; the core is to accurately determine the micro-motion standby state by combining micro-motion frequency band energy, main frequency, average acceleration and interactive distance parameters; when in this state, the system controls the terminal to stop sending high-frequency messages and sends keep-alive messages containing micro-motion parameters, and the edge end reconstructs the micro-motion pose until the action wake-up condition is met to restore high-frequency transmission; the present application realizes accurate distinction between natural micro-motion and fine operation, and can significantly reduce the communication data volume and edge end forwarding burden in the standby stage.

Description

Technical Field

[0001] This invention relates to the field of metaverse scene synchronous interaction and edge computing communication technology, specifically to an edge data transmission method and system for metaverse scene synchronous interaction. Background Technology

[0002] With the continuous development of multi-user virtual reality collaborative assembly, review and observation applications in the metaverse scenario, user terminals need to continuously synchronize the pose data of the head, hand end and the currently manipulated object to the edge side to ensure the consistency of actions and the continuity of interaction in the multi-user shared scene; how to reduce the high-frequency pose transmission burden during the standby stage while maintaining the scene synchronization realism and interaction timeliness has become a problem that needs to be solved in this type of synchronous interaction technology. Traditional multi-user virtual reality pose synchronization solutions currently rely mainly on the following methods: continuously uploading real poses at high frequency, judging the stationary state according to a fixed spatial threshold, and directly forwarding and synchronizing the received poses at the edge. However, continuous high-frequency uploading, fixed spatial threshold for judging stationary states, and direct forwarding synchronization all have certain drawbacks. For example, continuous high-frequency uploading will cause a large number of tiny pose updates to be sent even when the user is hovering, holding the device stably, or briefly stopping, resulting in a large number of communication packets and a heavy burden on edge forwarding. Fixed spatial threshold for judging stationary states makes it difficult to distinguish between natural micro-movements and real fine operations, and is prone to misjudgment when approaching the assembly target or performing fine-tuning actions. Direct forwarding synchronization, due to the lack of parameterized processing for tiny pose changes during the standby phase, is prone to causing high-frequency jitter in the proxy model in the shared scenario, and it is difficult to simultaneously ensure transmission efficiency and the timeliness of synchronization when real actions are restored. Summary of the Invention

[0003] The purpose of this invention is to provide an edge data transmission method and system for synchronous interaction in a metaverse scene, addressing the following technical problems: Existing virtual reality multi-user pose synchronization technologies suffer from significant shortcomings in handling high-frequency transmission burdens during standby, difficulty in distinguishing between natural micro-movements and fine operations, and high-frequency jitter in proxy models. There is an urgent need to propose a metaverse edge data transmission method and system that can balance synchronous realism and interactive timeliness, reduce transmission burdens, and smoothly reconstruct micro-movement poses. The objective of this invention can be achieved through the following technical solutions: The system collects the target user's pose data, including the target user's head, hand ends, and pose monitoring nodes on the virtual interactive object currently being controlled by the target user. For each pose monitoring node, it collects single-frame pose data at the corresponding sampling time and arranges them according to the sampling time to form continuous pose data. At the same time, it sends high-frequency pose messages to the edge server. Based on the continuous pose data, the edge server performs spectral analysis on the displacement data of each pose monitoring node within the sliding analysis window, calculates the average acceleration based on the continuous pose data, and analyzes the interaction distance between the pose monitoring node or the virtual interactive object and the interactive virtual objects in the scene to obtain the micro-motion frequency band energy ratio, micro-motion main frequency, displacement amplitude, initial phase, average acceleration, and interaction distance. When the energy ratio of the micro-motion frequency band meets the preset energy threshold condition, the micro-motion main frequency falls within the preset micro-motion frequency range, the average acceleration falls within the preset acceleration range, and the interaction distance meets the preset distance threshold condition, the target user is determined to be in a micro-motion standby state; the edge server maintains the normal synchronization state, the micro-motion standby state, and the wake-up transition state. When the target user is determined to have entered a micro-motion standby state, the system switches to the micro-motion standby state and controls the user terminal to stop sending the high-frequency pose message. Instead, it sends a connection keep-alive message according to the connection keep-alive cycle. The connection keep-alive message carries micro-motion parameters, which include at least the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time. The system continues sending the high-frequency pose message when the user is not in a micro-motion standby state. The edge server generates a reconstructed micro-motion pose of the target user based on the latest cached real pose and the micro-motion parameters, and updates the reconstructed micro-motion pose over time. The reconstructed micro-motion pose is merged with other user poses and synchronized; when an action wake-up detection is performed based on at least one of velocity, acceleration, and jerk and the action wake-up condition is met, the system switches to a wake-up transition state and controls the user terminal to resume sending high-frequency pose messages; after the edge server receives the first high-frequency pose message, the reconstructed micro-motion pose is replaced with the real pose corresponding to the first high-frequency pose message.

[0004] Preferably, the following steps are taken: collecting the three-dimensional position and posture parameters of the target user's head; collecting the three-dimensional position and posture parameters of the target user's hand controllers or gesture ends; collecting the pose data of the virtual interactive object currently being controlled by the target user; recording the sampling time information corresponding to each sampling point; and for any pose monitoring node, constructing a single frame of raw pose data composed of position parameters and posture parameters, and arranging it according to the sampling time to form the continuous pose data.

[0005] Preferably, a sliding analysis window consisting of a preset number of sampling points is established; using the position of the first frame of the sliding analysis window as a reference, the displacement data of each sampling point in the sliding analysis window relative to the position of the first frame is calculated; the displacement data is subjected to spectrum analysis to obtain the spectrum results corresponding to each analysis frequency point; the energy proportion of the micro-motion frequency band within the preset micro-motion frequency band is determined according to the spectrum results; the dominant frequency point is determined within the preset micro-motion frequency band, and the micro-motion dominant frequency, displacement amplitude, and initial phase are obtained.

[0006] Preferably, the velocity and acceleration are determined based on the changes in sampling points within the sliding analysis window, and the average acceleration is calculated; the interaction distance is obtained based on the distance between the outer bounding body of the virtual interactive object currently controlled by the pose monitoring node or the target user and the outer bounding body of the interactive virtual object in the scene; the energy ratio of the micro-motion frequency band, the micro-motion main frequency, the average acceleration, and the interaction distance are used as joint determination conditions; when the energy ratio of the micro-motion frequency band meets the preset energy threshold condition, the interaction distance meets the preset distance threshold condition, the micro-motion main frequency falls within the preset micro-motion frequency range, and the average acceleration falls within the preset acceleration range, the target user is determined to be in a micro-motion standby state.

[0007] Preferably, the system maintains a normal synchronization state, a micro-motion standby state, and a wake-up transition state. When it is determined that the target user is in a micro-motion standby state and the standby determination conditions are met continuously for a preset confirmation time, the system switches from the normal synchronization state to the micro-motion standby state. After entering the micro-motion standby state, the system stops sending high-frequency pose messages and only retains the transmission of connection keep-alive messages. The micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time are written into the connection keep-alive message to form the micro-motion parameters used by the edge server to generate and reconstruct the micro-motion pose.

[0008] Preferably, the main direction of micro-motion is determined based on the cumulative result of the position changes of each adjacent sampling point within the sliding analysis window, and the starting time of the parameter is the moment when the edge server begins to use the micro-motion parameter to generate a reconstructed micro-motion pose.

[0009] Preferably, the micro-motion parameters are extracted, including the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time. Based on the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time, a periodic micro-displacement of the pose monitoring node is generated on the basis of the latest real position corresponding to the micro-motion standby state. The periodic micro-displacement is superimposed on the latest real pose cached by the edge server to obtain the reconstructed micro-motion pose, wherein the pose is maintained as the latest real pose corresponding to the micro-motion standby state. The reconstructed micro-motion pose is merged with the real poses of other active users into a global scene state frame and synchronized with other users.

[0010] Preferably, an action wake-up index is determined based on at least one of the velocity, acceleration, and jerk of the current sampling point; when the action wake-up index meets the preset action activation threshold condition, the micro-motion standby state is switched to the wake-up transition state, and high-frequency pose message transmission is resumed in the next transmission cycle; after the edge server receives the first high-frequency pose message sent by the target user, the reconstruction of micro-motion pose generation is stopped; the reconstructed micro-motion pose is replaced with the real pose corresponding to the first high-frequency pose message, and a smooth interpolation approximation is used for transition to restore the synchronous output of the real pose.

[0011] Preferably, the preset distance threshold is determined based on the allowable spatial error tolerance of the current target interaction task. The preset distance threshold is obtained by multiplying the allowable spatial error tolerance by an amplification factor, and the amplification factor is preset according to the task accuracy level.

[0012] An edge data transmission system for synchronous interaction in a metaverse scene includes a user terminal, an edge server, and a communication interface between the user terminal and the edge server. The user terminal is equipped with a data acquisition module for acquiring the target user's raw pose data, constructing continuous pose data of the pose monitoring node, and sending high-frequency pose messages or connection keep-alive messages. The edge server is equipped with an analysis module, which performs displacement analysis, spectrum analysis, velocity and acceleration calculation, and interaction distance analysis on the continuous pose data based on a sliding analysis window to obtain the micro-motion standby state determination result; the edge server is equipped with a switching module, which controls the switching between normal synchronization state, micro-motion standby state, and wake-up transition state based on the micro-motion standby state determination result, and sends switching control information for sending high-frequency pose messages or connection keep-alive messages to the user terminal; The edge server is equipped with a synchronization module, which is used to generate a reconstructed micro-motion pose based on the micro-motion parameters in the connection keep-alive message, merge the reconstructed micro-motion pose with other user poses and synchronize them, and stop generating the reconstructed micro-motion pose and restore the real pose synchronization after receiving the real pose corresponding to the high-frequency pose message sent by the target user.

[0013] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention can accurately distinguish between natural micro-movements and real fine operations by performing spectral analysis on the pose, calculating the average acceleration, and combining it with the interaction distance for joint judgment, thus avoiding misjudgment of the state when approaching an interactive object; in the micro-movement standby state, the control terminal stops sending high-frequency pose messages and instead periodically sends connection keep-alive messages carrying micro-movement parameters, reducing the amount of communication data during the standby phase and the forwarding burden on the edge server; 2. The present invention generates and merges and synchronizes the reconstructed fine motion pose based on the latest true pose cache and fine motion parameters on the edge side, effectively avoiding the high-frequency jitter of the proxy model during standby; at the same time, through the action wake-up mechanism based on speed, acceleration, and jerk, it quickly switches to the wake-up transition state and resumes high-frequency pose transmission when a real action is detected, taking into account the transmission efficiency during standby and the synchronization timeliness when resuming real actions. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Through the following description of the embodiments of the present invention with reference to the accompanying drawings, the above and other objects, features, and advantages of the present invention will become clearer. In the drawings: Figure 1 It is a schematic flowchart of an edge data transmission method for metaverse scene synchronization interaction provided by an embodiment of the present application.

[0015] Figure 2 It is a schematic block diagram of an edge data transmission system for metaverse scene synchronization interaction provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with specific embodiments.

[0017] This embodiment is applied to a multi-user virtual reality collaborative assembly review scenario; the execution entities in the system include user terminals and edge servers; the user terminal is a local computing device supporting a wearable display device, and is used to collect the original pose data of the head, the ends of both hands, and the currently manipulated virtual interaction object of the target user; the edge server is used to receive pose messages, perform sliding window analysis, determine whether the target user enters the fine motion standby state, control the switching of the message sending method, and generate and synchronize the reconstructed fine motion pose to other session users during standby; The pose monitoring node in this embodiment refers to a pose sampling object set on the user's head, the end of the hand, or the currently manipulated virtual interaction object; the head pose corresponds to the spatial position and attitude of the user's perspective carrier, the end pose of the hand corresponds to the spatial position and attitude of the end of the controller or the end of the gesture, and the pose monitoring node on the virtual interaction object corresponds to the position and attitude of the object held or manipulated by the user in the scene coordinate system; the above poses are all objective spatial quantities, directly reflecting the user's observation, holding, and assembly behaviors in the virtual scene, rather than abstract data; As Figure 1 shown, the edge data transmission method for metaverse scene synchronization interaction includes: S1. The user terminal collects single-frame raw pose data from the pose monitoring nodes on the target user's head, hand ends, and / or the controlled virtual interactive object, arranges them into continuous pose data according to the sampling time, and sends high-frequency pose messages. S2. The edge server performs spectral analysis on the displacement based on continuous pose data and a sliding analysis window to obtain the energy proportion of the micro-motion frequency band and the micro-motion main frequency. It also calculates the velocity, average acceleration, and jerk, and analyzes the interaction distance between the pose monitoring node or the controlled virtual interactive object and the interactive virtual objects in the scene. When the energy proportion of the micro-motion frequency band meets the preset energy threshold condition, the micro-motion main frequency falls within the preset frequency range, the average acceleration falls within the preset acceleration range, and the interaction distance meets the preset distance threshold condition, it is determined to be in the micro-motion standby state; otherwise, it is determined not to be in the micro-motion standby state. S3. The edge server maintains normal synchronization, micro-motion standby, and wake-up transition states; when it is determined to be in micro-motion standby state, it switches to control the terminal to stop sending high-frequency pose messages and periodically send connection keep-alive messages carrying micro-motion parameters, including micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time; when not in this state, it maintains the sending of high-frequency pose messages. S4. The edge server generates and updates the reconstructed micro-motion pose based on the latest real pose cache and micro-motion parameters, and merges and synchronizes it with the poses of other users. When the action wake-up condition is met based on at least one of velocity, acceleration and jerk, it switches to the wake-up transition state, controls the terminal to resume sending high-frequency pose messages, and replaces the reconstructed micro-motion pose with the real pose corresponding to the first high-frequency pose message received.

[0018] Step S1 includes: Collect the three-dimensional position and pose parameters of the target user's head; Collect the three-dimensional position and posture parameters of the target user's two-hand controllers or gesture ends; Based on the current interaction state, collect the pose data of the virtual interactive object currently being controlled by the target user; Record the sampling time information corresponding to each sampling point; For any pose monitoring node, a single frame of raw pose data is constructed, consisting of position parameters, attitude parameters, and corresponding sampling time information, and then arranged according to the sampling time to form continuous pose data.

[0019] Step S2 involves performing a spectral analysis of the displacement, including: Establish a sliding analysis window consisting of a preset number of sampling points; Using the position of the first frame of the sliding analysis window as a reference, calculate the displacement data of each sampling point within the sliding analysis window relative to the position of the first frame; Spectral analysis was performed on the displacement data to obtain the spectral results corresponding to each analysis frequency point; Determine the energy percentage of the preset micro-motion frequency band based on the spectrum results; Within the preset micro-motion frequency band, determine the dominant frequency point and obtain the micro-motion main frequency, displacement amplitude, and initial phase.

[0020] In step S1, the user terminal continuously reads the original sampled values ​​of each pose monitoring node according to the sampling period of the tracking device; the monitoring endpoint in this embodiment and the corresponding embodiment is the aforementioned pose monitoring node; at any sampling time, the user terminal reads the three-dimensional position parameters and attitude parameters of the monitoring node, i.e. the monitoring endpoint, where the three-dimensional position parameters can be written as a position vector and the attitude parameters can be represented by quaternions. For the head, hands, and currently manipulated object, the user terminal constructs corresponding single-frame pose raw data. The single-frame pose raw data includes at least the spatial position, spatial orientation, and sampling time information corresponding to the sampling point of the monitoring node. The user terminal arranges the single-frame pose raw data of each sampling time in chronological order to form continuous pose data, and continuously sends high-frequency pose messages to the edge server through the communication interface. The high-frequency pose messages are pose update messages sent according to or close to the tracking sampling frequency, which are used to ensure the synchronization accuracy of real actions in multi-user scenarios. Under a set of optional parameters, the sampling frequency of the user terminal is preferably set within a range that matches the refresh capability of the virtual reality tracking device, so that the pose changes of the head, hands, and currently manipulated object can be continuously recorded; the number of window sampling points required for sliding analysis is calculated by the edge server according to the current sampling frequency, thereby ensuring that the time used for subsequent frequency band analysis is within the preset analysis duration; this analysis duration is preferably selected to take into account both frequency resolution and state response speed, so that the edge server can distinguish between low-frequency physiological micro-movements and real operating actions; In step S2, after receiving continuous pose data from the user terminal, the edge server establishes sliding analysis windows according to the monitoring nodes. The sliding analysis window is a time series segment composed of a preset number of sampling points. Every time the edge server receives a new pose sampling point, it pushes the sampling point into the current window and moves out the earliest sampling point, thereby repeatedly analyzing the data of adjacent time periods in continuous updates. If the number of window sampling points is less than the first preset threshold, low-frequency components are difficult to distinguish; if the number of window sampling points is greater than the second preset threshold, the switching response when the user switches from standby to operation will be slowed down. Therefore, it is preferable to take a value between the first preset threshold and the second preset threshold. For any pose monitoring node, the edge server uses the position of the first frame of the current sliding analysis window as a reference position and sequentially calculates the displacement data of other sampling points within the window relative to the position of the first frame. The displacement data preferably uses Euclidean displacement scalar, that is, the difference between the position vector of each sampling point and the position vector of the first frame is taken and then the L2 norm is taken. The resulting displacement sequence reflects the minute spatial movement of the monitoring node within the window, which can directly correspond to physical phenomena such as slight head shaking, slight tremor of the handheld part, or slight swing of the tool end. After obtaining the displacement sequence, since the displacement scalar is always non-negative, direct analysis will produce a non-zero DC bias. Therefore, the edge server calculates the arithmetic mean of the displacement sequence, subtracts the arithmetic mean from the displacement value of each sampling point to perform de-meaning processing and eliminate the DC component; the edge server performs spectral analysis on the de-meaned displacement sequence. The mean-reduced displacement values ​​are used as discrete sequence inputs for discrete spectrum transformation operations such as Fast Fourier Transform to obtain the complex spectrum results corresponding to each analysis frequency point; the edge server then establishes a mapping relationship between frequency points and real frequencies based on the sampling frequency and the number of sampling points in the window, thereby determining the real vibration frequency corresponding to each frequency point. The edge server is configured with a preset micro-motion frequency band to cover low-frequency micro-motions caused by breathing, heartbeat, and muscle micro-tension when the user hovers and observes or holds the device stably. In a specific system, this micro-motion frequency band is represented by a set of upper and lower limit frequency configuration items. For example, the lower limit frequency is set to 0.1 Hz and the upper limit frequency is set to 5 Hz to cover the main physiological micro-motion range. The edge server calculates the square of the modulus of the complex spectrum result at each analysis frequency point as the energy value of that frequency point. By traversing the frequency mapping table, the energy values ​​of all frequency points falling within the micro-motion band are summed to obtain the total energy of the micro-motion band; then, the energy values ​​of all frequency points within the full-frequency effective analysis range from 0 to the Nyquist frequency are summed to obtain the total effective energy; dividing the two yields the energy percentage of the micro-motion band. The specific calculation formula is as follows:

[0021] in, The complex spectrum result after discrete spectrum transformation. To analyze the actual frequency corresponding to the frequency point, This is the Nyquist frequency of the system; the higher this energy percentage, the more concentrated the current displacement change is within the preset micro-motion frequency band, and the closer it is to natural micro-vibration rather than active operation. The edge server continues to search for the frequency point with the largest amplitude within the preset micro-motion frequency band and determines this frequency point as the dominant frequency point. The micro-motion master frequency is obtained from the real frequency corresponding to the dominant frequency point, the displacement amplitude is obtained from the spectral amplitude of this frequency point, and the initial phase is obtained from the phase angle of this frequency point. The micro-motion master frequency represents the main periodic characteristics of the current micro-motion change of the monitoring node; the displacement amplitude represents the displacement magnitude of the micro-motion in space; the initial phase represents the phase state of the periodic micro-motion at the reference starting position. The above parameters together constitute the core low-dimensional parameter basis required for subsequent standby transmission and edge reconstruction. In addition to spectrum analysis, the edge server also calculates average acceleration based on continuous pose data. Specifically, the edge server first calculates the velocity vector according to the position changes of adjacent sampling points and the sampling period, and then calculates the acceleration vector according to the changes of adjacent velocity vectors and the sampling period. The acceleration magnitude of each sampling point is averaged within the sliding analysis window to obtain the average acceleration. The average acceleration is used to distinguish between real intentional actions and natural micro-movements. If the user is starting a hand movement, turning, or tentatively aligning, the acceleration changes within the window will usually exceed the range of natural micro-movements. If it is just stable observation or slight holding, the average acceleration is more likely to fall within the preset standby acceleration range. The edge server also analyzes the interaction distance; the interaction distance refers to the spatial proximity between the pose monitoring node or the currently manipulated virtual interactive object and the interactive virtual objects in the scene; the edge server can select the closest distance as the current interaction distance based on the minimum spatial distance between the outer bounding body of the monitoring node's object and the outer bounding bodies of each interactive virtual object in the scene; when the user approaches the assembly position, the insertion hole position, or the contact edge, this distance will decrease to below the preset distance threshold; incorporating this distance into the judgment condition can prevent the user from accidentally switching to the micro-motion standby state when approaching the precision interactive target; After calculating the energy proportion of the micro-motion frequency band, the micro-motion main frequency, the displacement amplitude, the initial phase, the average acceleration, and the interaction distance, the edge server performs a joint judgment. The prerequisite is that complete continuous pose data has been formed within the current window. The edge server compares the energy proportion of the micro-motion frequency band with a preset energy threshold, compares the micro-motion main frequency with a preset micro-motion frequency range, compares the average acceleration with a preset acceleration range, and compares the interaction distance with a preset distance threshold. The edge server determines that the target user is in micro-motion standby state only when the energy ratio of the micro-motion frequency band meets the energy threshold condition, the micro-motion main frequency falls within the preset micro-motion frequency range, the average acceleration falls within the preset acceleration range, and the interaction distance meets the preset distance threshold condition. This determination order allows frequency band characteristics, motion intensity, and assembly proximity to participate in the decision-making, thereby reducing the probability of misjudging real precision movements as standby. The above preset thresholds are all parameterizable. The micro-motion frequency band energy ratio threshold is used to ensure that the current motion energy is mainly concentrated in the low-frequency micro-motion band. Its preferred value is set within the range that can distinguish natural micro-motion samples from active micro-adjustment samples. The average acceleration threshold is used to filter out real actions such as starting actions, trial alignment, and attitude turning. Its preferred value is set between the upper limit of the average acceleration of natural micro-motion and the lower limit of the average acceleration of active actions. The preset distance threshold is used to limit the entry into standby mode when approaching the assembly target. Its preferred value is related to the allowable spatial error tolerance of the current interactive task, so that high-precision assembly tasks correspond to smaller distance thresholds, and rough observation tasks correspond to relatively larger distance thresholds. With the above parameter settings, the micro-motion standby state determination result output by the edge server not only comes from the pose data itself, but also from the real spatial relationship between the user and the interactive objects in the scene. Therefore, it can directly affect the physical performance in multi-user synchronization. In step S3, the edge server maintains a normal synchronization state, a micro-motion standby state, and a wake-up transition state. The normal synchronization state refers to the state in which the user terminal continuously sends high-frequency pose messages, and the edge server directly uses the real pose for multi-user synchronization. The micro-motion standby state refers to the state in which the edge server has determined that the user is in the natural micro-motion stage and controls the terminal to stop sending high-frequency pose messages and instead send connection keep-alive messages. The wake-up transition state refers to the intermediate state in which the system has detected that the user has started to act again, the terminal is preparing to resume high-frequency pose messages, and the edge server is waiting for the real pose to take over the output again. When the edge server determines that the target user has entered the micro-motion standby state based on the joint determination result of step S2, the edge server sends handover control information to the user terminal. After receiving the handover control information, the user terminal stops sending high-frequency pose messages and sends connection keep-alive messages according to the connection keep-alive cycle. The connection keep-alive message is a periodic message originally used to maintain the session connectivity between the user terminal and the edge server. In this embodiment, without changing its basic connectivity function, the received micro-motion parameters are written into its extended service payload for status confirmation. The micro-motion parameters include at least the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time. In this embodiment, the main micro-motion direction is the main spatial swing direction obtained by the edge server or user terminal based on the cumulative position change direction of adjacent sampling points within the window, which is used to indicate the direction along which the current natural micro-motion mainly unfolds; the parameter start time refers to the starting reference time when the edge server begins to generate the reconstructed micro-motion pose according to the set of micro-motion parameters. After the above parameters are put into the connection keep-alive message, the high-frequency pose point series that was previously continuously uploaded can be replaced with a small number of low-dimensional parameters; when the standby judgment condition is met, the high-frequency small displacement changes during the standby period are converted into parameter expressions reconstructed by the edge server, so as to reduce the number of communication messages and the amount of forwarded data of the edge nodes. When not in micro-motion standby mode, the edge server keeps the user terminal sending high-frequency pose messages. At this time, multi-user synchronization is still based on the real pose and no parameterization is performed. This ensures that when the user is in a state of normal observation, active hand operation or close to the assembly target, the system still transmits pose data according to the original high-frequency synchronization path without sacrificing the millimeter-level interaction response capability. In step S4, after the edge server enters the micro-motion standby state, it no longer waits for the high-frequency pose point list, but generates the reconstructed micro-motion pose of the target user based on the latest real pose in the cache and the micro-motion parameters in the connection keep-alive message; the latest real pose in the cache refers to the real position and real pose that the edge server received and confirmed once before the terminal stopped high-frequency reporting. Reconstructing micro-motion pose refers to the alternative pose calculated by the edge server based on the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction and parameter start time. The position part changes periodically and slightly over time, while the pose part preferably remains the latest real pose before entering standby or makes corresponding small changes when the pose parameters have been uploaded. Specifically, the edge server is based on the current time. Parameter start time Micro-frequency Displacement amplitude Initial phase and the unit vector of the principal direction of the micro-motion Through formula The periodic micro-displacement at the current moment is calculated. Then, this micro-displacement is superimposed onto the base position to obtain the simulated micro-motion position. The edge server repeats this calculation according to the scene synchronization cycle, thereby updating and reconstructing the micro-motion pose over time. Since this update is based on low-frequency periodic parameters rather than random sampled noise, the proxy model seen by other users appears as smooth and continuous natural micro-motions, rather than high-frequency scattered jitter. After obtaining the reconstructed micro-motion pose, the edge server merges it with the poses of other users into a global scene state frame and synchronizes it. Merging means that in the same session frame, the real pose is used for users in the active state, and the reconstructed micro-motion pose is used for users in the micro-motion standby state, and then uniformly publishes it to other user terminals. This preserves the natural presence of standby users and reduces the amount of pose transmission during standby and the burden of per-packet forwarding of edge nodes. In micro-motion standby mode, the edge server also needs to perform action wake-up detection; the prerequisite is that the user terminal still maintains session connectivity according to the connection keep-alive cycle and continuously or periodically provides sufficient detection basis to judge action changes; action wake-up detection is based on at least one of speed, acceleration, and jerk. The jerk is a quantity calculated from the change in adjacent accelerations and the sampling period, used to reflect the degree of abrupt change at the start of an action; the edge server or user terminal calculates and compares the velocity, acceleration and jerk within the current sampling point or the current detection period according to the preset action wake-up conditions; when at least one or a combination of indicators meets the action wake-up conditions, the system considers that the user has transitioned from natural micro-movements to the real action preparation stage; Once the action wake-up condition is met, the edge server switches to the wake-up transition state and controls the user terminal to resume sending high-frequency pose messages. After the user terminal resumes high-frequency pose reporting, the edge server waits to receive the first high-frequency pose message. Once the edge server receives the first high-frequency pose message, it replaces the previously output reconstructed micro-motion pose with the real pose corresponding to the message and stops generating fitted micro-motions for the user. Thus, the system completes the switch from micro-motion standby state to real action synchronization state. This switching path ensures that when the user starts millimeter-level insertion, fine twisting, or hole-aligning actions, the scene output immediately returns to real pose driving, and the synchronization of real actions is not delayed due to the standby optimization strategy. In this embodiment, the user terminal first collects a high-frequency pose stream that reflects the real physical space state. The edge server then uses a sliding analysis window to extract objective parameters such as micro-motion frequency band energy, dominant frequency, displacement amplitude and average acceleration from the pose stream. Combined with the interaction distance, it determines whether the user is far away from the precision operation point and is in a natural micro-motion state. Only when the above conditions are met will the system switch the high-frequency coordinate stream to low-dimensional parameter transmission, and then the edge server will reconstruct the smooth micro-motion pose based on the parameters. Since independent pose messages are no longer encapsulated for each tiny sampled displacement during standby, the number of messages processed per second and the forwarding burden of the edge server will decrease accordingly. At the same time, since other users see the smooth micro-motion trajectory generated continuously by the parameters, the jitter of the user agent model in the shared scene will also be reduced. In a typical scenario, a user terminal detects that the end of the hand and the virtual part being held are in a stable hovering state. The edge server calculates within a continuous sliding analysis window that: the energy of the displacement sequence is mainly concentrated in the preset micro-motion frequency band, the dominant frequency falls within the preset physiological micro-motion range, the average acceleration is at a low level, and the interaction distance between the currently held part and the nearest interactive part is greater than the distance threshold corresponding to the current task. At this time, the edge server determines that the user has entered a micro-motion standby state and controls the terminal to stop sending high-frequency pose messages, retaining only the connection keep-alive message carrying the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time. The edge server uses these parameters to generate periodic slight oscillations of the user agent model during standby and merges them with the real poses of other active users for synchronization. When the user begins to perform a real plug-in action, the velocity or acceleration detection value increases and triggers the action wake-up condition. The terminal resumes high-frequency pose reporting, and the edge server immediately stops the simulation micro-motion output and resumes real synchronization after receiving the first real pose. The whole process maintains the responsiveness before and after precise interaction and reduces invalid high-frequency packet transmission during standby. In this embodiment, the implementation further explains the determination rules for the micro-motion standby state, the state switching method, the micro-motion parameter generation method, and the specific process of the edge server regenerating and reconstructing the micro-motion pose. This implementation is also applied to multi-user virtual reality industrial collaboration scenarios. The execution entities include the edge server and the user terminal. The edge server is responsible for completing the determination logic and standby regeneration logic, while the user terminal is responsible for changing the message sending method after receiving the switching control. The acquisition of the determination result of whether the target user is in micro-switch standby state in step S2 includes: The velocity and acceleration are determined based on the changes in sampling points within the sliding analysis window, and the average acceleration is calculated. The interaction distance is obtained based on the distance between the bounding body of the virtual interactive object currently being controlled by the pose monitoring node or the target user and the bounding body of the interactive virtual object in the scene. The energy ratio of the micro-motion frequency band, the micro-motion main frequency, the average acceleration, and the interaction distance are used as joint judgment conditions. When the energy ratio of the micro-motion frequency band meets the preset energy threshold condition, the interaction distance meets the preset distance threshold condition, the micro-motion main frequency falls within the preset frequency range, and the average acceleration falls within the preset acceleration range, the target user is determined to be in micro-motion standby state.

[0022] Step S3 includes: Maintain normal synchronization, micro-motion standby, and wake-up transition states; When it is determined that the target user is in micro-motion standby mode and the standby determination conditions are met continuously for a preset confirmation time, the system switches from normal synchronization mode to micro-motion standby mode. After entering the micro-motion standby state, the transmission of high-frequency pose messages is stopped, and only the transmission of connection keep-alive messages is retained; The micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time received from the edge server are written into the connection keep-alive message to serve as the state confirmation parameters for the edge server to generate the reconstructed micro-motion pose.

[0023] The main direction of micro-motion is determined based on the cumulative result of the position changes of each adjacent sampling point within the sliding analysis window. The parameter start time is the moment when the edge server begins to use the micro-motion parameters to generate and reconstruct the micro-motion pose.

[0024] Step S4, which generates the reconstructed micro-motion pose of the target user based on the micro-motion parameters, includes: Extract the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time from the micro-motion parameters; Based on the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction and parameter start time, the periodic micro-displacement of the pose monitoring node is generated on the basis of the latest real position corresponding to the micro-motion standby state. The periodic micro-displacements are superimposed on the latest true pose cached on the edge server to obtain the reconstructed micro-motion pose, where the pose is maintained as the latest true pose corresponding to the micro-motion standby state. The reconstructed micro-motion pose is merged with the real poses of other active users into a global scene state frame and synchronized with other users.

[0025] In this embodiment, the preset energy threshold is a judgment threshold used by the edge server to determine whether the motion energy within the window is mainly concentrated in the preset micro-motion frequency band; the preset micro-motion frequency range is the range that the edge server recognizes as a natural micro-motion candidate frequency; the preset acceleration range is the acceleration limit range used to filter out real operation actions; and the preset distance threshold is the distance threshold used to limit the entry into standby mode when approaching the interactive target. The preset confirmation duration refers to the cumulative confirmation duration that must be reached when the standby judgment condition is continuously met. Only when this confirmation duration is reached will the edge server perform a state switch. The standby judgment is based on continuous spatial motion characteristics and interaction distance characteristics, not just on the results of a single sampling. In the determination process of the embodiment, the edge server calculates the velocity and acceleration based on the changes of sampling points within the sliding analysis window; specifically, the edge server calculates the difference between the position vectors of adjacent sampling points of the same pose monitoring node and divides it by the sampling period to obtain the velocity vector of the corresponding sampling point; then it calculates the difference between adjacent velocity vectors and divides it by the sampling period to obtain the acceleration vector of the corresponding sampling point. To avoid single-point abnormal changes directly affecting the judgment results, the edge server does not directly use the acceleration of a single sampling point, but instead statistically analyzes the acceleration magnitude across the entire sliding analysis window to obtain the average acceleration. The average acceleration reflects the overall motion intensity within that window. If the user is simply hovering to observe or holding the workpiece steady, the average acceleration will usually be small. If the user has already begun fine-tuning, rotating, or inserting, the average acceleration will deviate from the natural micro-motion range. The edge server simultaneously obtains the interaction distance based on the distance between the bounding volume of the virtual interactive object currently being manipulated by the pose monitoring node or the target user and the bounding volume of the interactive virtual objects in the scene; the bounding volume is a geometric boundary volume used to approximate the spatial occupancy of an object, preferably using an axial bounding box that is easy to calculate the minimum distance; the edge server maintains a set of interactive virtual objects in the current scene and calculates the minimum spatial distance between the currently monitored object and the bounding volumes of each virtual object in the set in each analysis cycle; Specifically, for any two axial bounding boxes, the system calculates the one-dimensional distance between their projection intervals on the X, Y, and Z coordinate axes respectively. If the intervals overlap, the one-dimensional distance is 0. The square root of the sum of the squares of the three-axis one-dimensional distances is taken to obtain the minimum distance in the bounding volume space between the two objects. The edge server calculates all distance results by traversing the above set and selects the minimum value as the interaction distance. This interaction distance represents the proximity between the user's current hand or held object and the potential interactive target. It is an objective physical quantity directly derived from the relationship between objects in three-dimensional space. Introducing the interaction distance into the determination can prevent the user from accidentally entering standby mode when approaching holes, joint surfaces, or edge positions due to small movement amplitude. After the above calculations are completed, the edge server uses the micro-motion frequency band energy ratio, micro-motion main frequency, average acceleration, and interaction distance as joint judgment conditions. In specific execution, the edge server first determines whether the micro-motion frequency band energy ratio meets the preset energy threshold condition, then determines whether the interaction distance meets the preset distance threshold condition, then determines whether the micro-motion main frequency falls within the preset micro-motion frequency range, and then determines whether the average acceleration falls within the preset acceleration range. Only when all four conditions are met simultaneously will the edge server determine that the target user is in micro-motion standby state. The frequency band characteristics, motion intensity, and proximity constraints are combined to make a multi-condition joint judgment to distinguish between natural micro-motions and real assembly actions. In a set of preferred settings, the preset energy threshold is selected to reflect the dominance of low-frequency components. Preferably, most results of natural micro-motion samples are above this threshold, while most results of real micro-adjustment samples are below this threshold. The preset acceleration range is selected to be a lower range of motion intensity to exclude initiation, tentative insertion, and active turning. The preset distance threshold is not fixed as a single constant, but can be adjusted according to the accuracy requirements of the current task. In specific engineering implementation, the preset energy threshold can be obtained by collecting the pose data of multiple test users during natural hovering to obtain the micro-motion energy ratio sequence, and taking the statistical average value minus one standard deviation as the calibration benchmark; the upper limit of the preset acceleration range can be calibrated by the lower edge value of the acceleration when the test user performs a slow starting action, and the lower limit is set to a smaller value covering the static drift value of the device; thus, when the system is applied to precision assembly review, the preset distance threshold will be tightened according to the allowable spatial error of the task; when the system is applied to long-distance observation, the preset distance threshold can be appropriately relaxed to improve the standby recognition coverage; In the state machine switching of the embodiment, the edge server maintains a normal synchronization state, a micro-motion standby state, and a wake-up transition state; the edge server performs a standby determination once in each analysis cycle and records the cumulative number of consecutive standby determination conditions met; when it is detected that the standby determination conditions are met in a certain analysis cycle, the edge server does not immediately switch the state, but instead adds the time length of the analysis cycle to the cumulative confirmation time. If the conditions for standby are met continuously in subsequent analysis cycles, the cumulative confirmation time will continue to increase. If the standby determination conditions are no longer met in any analysis cycle, the cumulative confirmation time will be reset to zero or restarted. Only when the cumulative confirmation time reaches the preset confirmation time will the edge server switch the user's status from normal synchronization to micro-motion standby. By using the preset confirmation time, the standby characteristics will be stably present in multiple consecutive sampling cycles or multiple consecutive analysis windows, and the user will be determined to have entered standby. After the edge server confirms the switch to micro-motion standby mode, the edge server sends state switching control information to the user terminal. The user terminal then stops sending high-frequency pose messages and only retains connection keep-alive messages. At this time, the user terminal or the edge server needs to form a set of micro-motion parameters for subsequent simulation micro-motion regeneration, including micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time. The above parameters are all derived from the sliding analysis results on which the standby determination is based, so there is no need to introduce new complex calculation paths. After they are written into the connection keep-alive message, the connection keep-alive message continues to play the role of maintaining connection survival on the one hand, and plays the role of transmitting micro-motion parameters during standby on the other hand, thereby realizing the reuse of the original network message channel. The embodiment further defines the method for determining the main direction of micro-motion and the starting time of the parameters; specifically, the main direction of micro-motion is determined based on the cumulative result of the position changes of each adjacent sampling point within the sliding analysis window; the edge server or user terminal first calculates the difference between the position vectors of each pair of adjacent sampling points within the window to obtain a set of position change vectors; To prevent the opposing change vectors from canceling each other out during the reciprocating micro-motion process, the system sets the position change vectors of the first and last sampling points of the window as the reference vector. For each vector in a set of position change vectors, its dot product with the reference vector is calculated. If the dot product is negative, the position change vector is flipped in the opposite direction to achieve flipping towards a unified reference hemisphere. The flipped change vectors are then vector-accumulated in chronological order to obtain the overall position change result. The overall position change result is then normalized to obtain the main direction of the micro-motion. Specifically, the normalization process is as follows: divide the vector of the overall position change result by its own L2 norm, i.e., the vector length; if the L2 norm is less than a preset minimum value such as 1e-6 to prevent division by zero errors, then set the main direction of the micro-motion to the direction of the previous effective analysis cycle or the default reference direction to avoid the calculation abnormality caused by the zero vector; this direction reflects the most important spatial direction of the micro-oscillation in the current standby phase. If the micro-motion within the window is mainly manifested as small reciprocating motions along a certain approximately fixed direction, then the accumulated results can stably represent this directional characteristic. The parameter start time is defined as the moment when the edge server begins to use this set of micro-motion parameters to generate the reconstructed micro-motion pose, that is, the reference start time when the edge server confirms that it has entered the micro-motion standby state and enabled parameterized regeneration. After setting this start time, the edge server can use the same time reference in each subsequent calculation of the simulated micro-displacement, ensuring the continuity of the output periodic displacement. In the edge regeneration process of this embodiment, after receiving or confirming the micro-motion parameters, the edge server first extracts the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time from the connection keep-alive message; after extraction, the edge server reads the latest real position corresponding to the micro-motion standby state as the base position, and reads the corresponding latest real pose as the base pose; the edge server generates the periodic micro-displacement of the pose monitoring node based on the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time; Its calculation logic is as follows: taking the parameter start time as the time reference, the periodic change rate is determined by the micro-motion main frequency, the initial waveform position is determined by the initial phase, the swing amplitude is determined by the displacement amplitude, and the swing direction is determined by the micro-motion main direction; the micro-displacement vector calculated in this way is a low-frequency displacement that changes continuously with the synchronization period. The edge server superimposes the periodic micro-displacement onto the position portion of the latest true pose in the cache to obtain the position result of the reconstructed micro-motion pose. For the pose portion, in this embodiment, the pose is maintained as the latest true pose corresponding to the micro-motion standby state, that is, no new high-frequency sampling update is performed on the pose during the standby period, and the true pose at the time of switching is directly used. The high-frequency minute position changes during standby are replaced by parameterization while maintaining the attitude, thus avoiding reconstruction errors caused by the lack of attitude micro-motion parameters. The edge server repeatedly executes the above micro-displacement calculation and pose superposition process in each scene synchronization cycle, thereby forming a reconstructed micro-motion pose updated over time. The edge server merges the reconstructed micro-motion pose with the real poses of other active users into a global scene state frame and synchronizes it with other users. Here, the real poses of other active users refer to the real spatial position and posture of the users who have not yet entered the micro-motion standby state and are still uploading pose data in a high-frequency manner. In this way, the system allows different users to use different synchronization sources within the same scene frame: active users use real coordinates, while standby users use parameterized coordinates generated from the edge; this will not affect the precise operation synchronization of active users, and will also reduce the high-frequency message transmission of standby users. The technical effects of this implementation are reflected in the following aspects: the standby determination no longer depends on the fixed spatial dead zone, but is determined by the frequency band characteristics, acceleration characteristics and interaction distance. Therefore, the system stops high-frequency transmission when the user is far away from the precision interaction target and only shows natural micro-movements; when the user approaches the assembly object or begins to actively move, it still maintains real synchronization. By setting a preset confirmation duration, the edge server avoids frequent state transitions between short pauses and short probing actions. Furthermore, during standby, the connection keep-alive message is used to carry micro-motion parameters instead of adding a new independent message type, which can reduce the number of high-frequency pose messages while maintaining the original session connectivity logic. The edge server regenerates and reconstructs the micro-motion pose by superimposing the basic real pose with periodic micro-displacements, so that the standby proxy model observed by other users presents a smooth and natural small-amplitude movement, reducing high-frequency jitter in the shared scenario. In a specific application scenario, a user holds a part to be assembled in a virtual assembly environment and observes it from a distance. The edge server establishes a sliding analysis window for the monitoring nodes of the user's hand end and the part held, and calculates within a continuous analysis cycle that: the frequency band energy of the displacement change is mainly distributed within the preset micro-motion frequency band, the dominant frequency falls within the natural micro-motion candidate range, the average acceleration is at a low level, and the interaction distance between the currently held part and the nearest interactive target is greater than the distance threshold corresponding to the task. The edge server further checks that these conditions are met within multiple consecutive analysis cycles and accumulate to the preset confirmation time, and then switches the user from the normal synchronization state to the micro-motion standby state. Afterwards, the user terminal stops sending high-frequency pose messages and only sends connection keep-alive messages containing the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time. The edge server calculates the periodic slight oscillation of the user agent model based on the latest real position and real attitude before the switch, and synchronizes the oscillation with the real pose of other active users. If the user moves the part closer to the assembly interface and begins to make real adjustments, the system will no longer maintain the standby judgment due to the reduced interaction distance or increased average acceleration, thus creating conditions for subsequent restoration of real synchronization. In this embodiment, the implementation describes the action wake-up exit rules in micro-motion standby state, the distance requirement threshold setting method driven by task tolerance, and the system structure and cooperation relationship between the modules to implement the method. This implementation is applicable to scenarios such as multi-person virtual assembly review, multi-person part observation, and multi-person synchronous review. In these scenarios, the premise of standby optimization is that it should not affect the synchronous timeliness when the user restarts the real action. Step S4, which involves exiting the micro-switch standby state when the action wake-up condition is met, includes: The action wake-up criterion is determined based on at least one of the velocity, acceleration, and jerk at the current sampling point. When the action wake-up indicator meets the preset action activation threshold condition, the micro-motion standby state is switched to the wake-up transition state, and high-frequency pose message transmission is resumed in the next transmission cycle. After the edge server receives the first high-frequency pose message sent by the target user, it stops reconstructing the micro-motion pose generation. Calculate the spatial deviation vector between the real position corresponding to the first high-frequency pose message and the current reconstructed micro-motion pose. Within a preset number of transition frames, perform smooth interpolation approximation of the pose based on this spatial deviation. After completing the visual smooth transition, fully restore the real pose and output synchronously.

[0026] The preset distance threshold is determined based on the allowable spatial error tolerance of the current target interaction task. The preset distance threshold is obtained by multiplying the allowable spatial error tolerance by the magnification factor, which is preset according to the task accuracy level.

[0027] like Figure 2 As shown, an edge data transmission system for synchronous interaction in a metaverse scene is used to implement the method in the embodiment, including a user terminal, an edge server, and a communication interface located between the user terminal and the edge server. The user terminal is equipped with a data acquisition module, which is used to collect the raw pose data of the target user, construct continuous pose data of the pose monitoring node, and send high-frequency pose messages or connection keep-alive messages. The edge server is equipped with an analysis module, which is used to perform displacement analysis, spectrum analysis, velocity and acceleration calculation, and interaction distance analysis on continuous pose data based on a sliding analysis window, and to obtain the micro-motion standby state determination result. The edge server is equipped with a switching module, which controls the switching between normal synchronization, micro-motion standby and wake-up transition states based on the micro-motion standby state determination result, and sends switching control information such as high-frequency pose message sending or connection keep-alive message sending to the user terminal. The edge server is equipped with a synchronization module, which is used to generate a reconstructed micro-motion pose based on the micro-motion parameters in the connection keep-alive message, merge the reconstructed micro-motion pose with other user poses and synchronize them, and stop generating the reconstructed micro-motion pose and restore the synchronization of the real pose after receiving the real pose corresponding to the high-frequency pose message sent by the target user.

[0028] In this embodiment, the motion wake-up indicator is a detection quantity used to determine whether the target user has transitioned from natural micro-movements to the initial stage of a real motion; the detection quantity comes from at least one of velocity, acceleration, and jerk at the current sampling point; velocity reflects the rate of position change, acceleration reflects the degree of velocity change, and jerk reflects the rate of acceleration change; While natural micro-movements can also cause these quantities to change, their range of change is usually small and relatively stable. However, when the user begins to perform real actions, especially when transitioning from hovering observation to plugging, aligning, twisting, or moving, these quantities will show a significant increase at the current sampling point or several consecutive sampling points. Therefore, this embodiment utilizes them to construct action wake-up indicators in order to promptly identify the start of real actions in the standby state. The determination of the action wake-up index can be done by single-item judgment or combination judgment. If the system has high requirements for the sensitivity of the start-up, any one of the velocity magnitude, acceleration magnitude, or jerk magnitude of the current sampling point can be directly used as the action wake-up index. If the system wants to take into account both jitter suppression and action recognition stability, the velocity, acceleration, and jerk can be combined according to preset weights to obtain a comprehensive action wake-up index. Since velocity, acceleration, and jerk have different dimensional characteristics, before combining them, the system uses a preset maximum reference value to perform dimensionless normalization on the above three indicators, and then performs a weighted summation of their respective dimensionless values ​​according to preset weights to form a unified comprehensive action wake-up indicator. Specifically, the normalization process involves adjusting the velocity magnitude of the current sampling point. acceleration modulus and accelerometer magnitude Divide by the maximum reference speed calibrated by the system respectively Maximum reference acceleration and maximum reference jerk Obtain the dimensionless value and combine it with the normalization condition. Each item corresponds to a preset weight , and Through formula Calculate the comprehensive arousal index ; The above weight allocation is set based on the motion detection sensitivity requirements: if the task needs to prevent missed detections of long-distance translational displacements, then a velocity weight is set. A relatively large proportion; if the task requires capturing subtle changes at the start, the weight of accelerator should be increased accordingly. The proportion of; Regardless of the method used, the prerequisite is that the edge server or user terminal can obtain the corresponding velocity, acceleration, and jerk values ​​based on the continuous pose data of the current sampling point and adjacent sampling points. The jerk can be obtained by dividing the difference between the current acceleration vector and the acceleration vector of the previous sampling point by the sampling period, thus reflecting the trend of motion change. Because consecutive differential operations in digital acquisition amplify high-frequency hardware sampling noise, to prevent such noise from dominating the wake-up index and causing frequent false wake-ups of the standby state in the control link, the system performs low-pass filtering on the received pose raw position sequence before calculating speed, acceleration, and jerk. Specifically, a second-order Butterworth low-pass filter is used, and its discrete-time difference equation is as follows:

[0029] in, This refers to the sequence number of discrete-time sampling points. The original pose position sequence before filtering is at the 1st... The input component values ​​of each sampling point For the filtered corresponding position sequence at the th The output component values ​​at each sampling point, with a preset cutoff frequency of 5 Hz, and the coefficients of the second-order Butterworth low-pass filter. , , , , The preset cutoff frequency and the current tracking sampling frequency of the system are jointly determined by bilinear transformation calculation, and the filtering process is applied to the three orthogonal coordinate components of the position vector independently; this makes the basic data fed into the differential calculation smooth, thereby ensuring that the action wake-up index truly reflects the physical human body's starting action. When the action wake-up index meets the preset action activation threshold condition, the system switches from micro-motion standby state to wake-up transition state. The preset action activation threshold condition is a judgment threshold set according to the difference between natural micro-motion samples and real starting action samples. Its setting principle is to be higher than the upper limit of the action wake-up index that may occur when pure natural micro-motion, while being lower than the action wake-up index level that can usually be achieved by real starting action. In practice, a preset action activation threshold is used. The calculation formula is:

[0030] in, and These are the historical average and standard deviation of the comprehensive motion awakening index corresponding to natural micro-motion samples collected during the system's preheating and calibration phase. The current sampling point... When the preset action activation threshold condition is met, it is determined that the condition is met. The detection of action activation in this embodiment and corresponding examples has the same meaning as satisfying the action wake-up condition in the aforementioned action wake-up detection; the preset action activation threshold condition is a quantitative judgment threshold that constitutes the aforementioned action wake-up condition; After entering the wake-up transition state, the user terminal resumes high-frequency pose message transmission in the next transmission cycle. The reason for not continuing to maintain the connection keep-alive message mode is that once the user starts real action, subsequent multi-user synchronization needs to continuously receive the real pose point list, and can no longer rely on the parameterized regeneration result. After the edge server switches from the micro-motion standby state to the wake-up transition state, it does not immediately discard the current reconstructed micro-motion pose, but continues to maintain the pose as a temporary output until it receives the first high-frequency pose message sent by the target user. The wake-up transition state is a brief transition phase in which the standby regeneration ends but the real pose has not yet been re-arrived. The edge server continues to output before receiving the real pose to avoid discontinuity in the proxy state of the user in the perspective of other users. Once the first high-frequency pose message arrives, the edge server immediately stops reconstructing the micro-motion pose generation and replaces the reconstructed micro-motion pose with the real pose corresponding to the first high-frequency pose message, restoring the real pose synchronous output. In this way, the user's transition from standby to action is completed in the order of detecting action activation, restoring high-frequency reporting, and the edge side taking over the real pose after receiving the first packet. This avoids the user being processed according to standby parameters during the initial stage and also avoids the synchronous output being interrupted at the switching point. In this embodiment, the distance requirement threshold is not fixed, but is determined based on the allowable spatial error tolerance of the current target interaction task; the distance requirement threshold in this embodiment is the preset distance threshold used in the aforementioned standby judgment condition; the allowable spatial error tolerance refers to the range of spatial deviations allowed for the current assembly, alignment, gap inspection, or insertion task, which directly reflects the current task's requirements for interaction accuracy; allowable spatial error tolerance The minimum physical assembly gap of the bounding volumes of the two virtual objects involved in the target interaction task The system determines this by reading scene attributes. The mapping relationship is as follows ; The edge server obtains the distance requirement threshold, i.e. the preset distance threshold, based on the product of the allowable spatial error tolerance and the magnification factor, where the magnification factor is preset according to the task accuracy level. To facilitate the transformation from abstract task levels to specific computational parameters, the edge server maintains a quantitative mapping table between precision levels and amplification factors. Tasks are divided into three discrete precision levels: high, medium, and low. For example, for a socket assembly task that requires close matching, a high precision level is set, and the system looks up the table to obtain and set a smaller amplification factor value, such as 1.0 to 1.5. For surface alignment operations, a medium precision level is set, and an intermediate amplification factor value, such as 1.5 to 3.0, is obtained. For the overall layout review, a low precision level is set to obtain a relatively large amplification factor, such as 3.0 to 5.0. If the task precision level is high, it means that the allowable spatial error tolerance is small and the amplification factor obtained from the table is small. Therefore, the calculated distance requirement threshold is reduced accordingly, so that the system is only allowed to enter standby when the user is significantly away from the precise target position. If the task precision level is low, it means that the allowable spatial error tolerance can be relatively relaxed and the amplification factor obtained from the table is large. Therefore, the distance requirement threshold is increased accordingly, so that the system can trigger standby optimization within the range of positions exceeding the preset tolerance threshold. The accuracy requirements of the target interaction task determine whether to pause the high-frequency real pose upload; when the current task requires high-precision hole alignment or precise insertion and the spatial position is close to the interaction target, it does not enter standby mode; while in the observation phase far from the target, as long as the user does not take any active action, the standby judgment range can be appropriately expanded; by associating the distance requirement threshold with the allowable spatial error tolerance, the system can adapt the same set of standby judgment logic to tasks with different accuracy levels, without having to design completely different judgment processes for each type of task. In one optional scenario, if the current task is a high-precision virtual assembly, the edge server reads the allowable spatial error tolerance corresponding to the task and calculates a smaller distance requirement threshold based on a preset amplification factor. In this case, when the end of the hand or the held virtual part approaches the interactive part, even if the displacement frequency band result and average acceleration within the sliding window meet the natural micro-motion characteristics, as long as the interaction distance does not exceed the smaller threshold, the edge server will not determine that the user has entered a micro-motion standby state, thus maintaining high-frequency real synchronization. Conversely, if the current task is only overall observation or rough position confirmation, the allowable spatial error tolerance is larger, and the distance requirement threshold will also increase accordingly, allowing the system to enter standby optimization at an earlier stage. The system of the embodiment includes a user terminal, an edge server, and a communication interface between the user terminal and the edge server; the user terminal is equipped with a data acquisition module, and the edge server is equipped with an analysis module, a switching module, and a synchronization module; the above modules are not abstract logical concepts, but program units or combinations of programs and hardware resources running on corresponding hardware devices, and their execution results directly affect the generation, transmission, switching, and global scene synchronization of pose messages; The acquisition module is deployed on the user terminal to collect the target user's raw pose data, construct continuous pose data for the pose monitoring node, and send high-frequency pose messages or connection keep-alive messages according to the current state. The input of the acquisition module comes from the head-mounted display tracker, the two-hand controller or the gesture end-point tracking interface, and the binding relationship of the currently manipulated virtual interactive object. The output is pose message data with position, posture and sampling time information. When in normal synchronization state, the acquisition module continuously outputs high-frequency pose messages. When in micro-motion standby state, the acquisition module changes to output connection keep-alive messages carrying micro-motion parameters according to the control information issued by the switching module. The analysis module is deployed on an edge server and is used to perform displacement analysis, spectrum analysis, velocity and acceleration calculation, and interaction distance analysis on continuous pose data based on a sliding analysis window to obtain the micro-motion standby state determination result. The input of the analysis module comes from the pose data stream of the user terminal received by the communication interface, and its output includes the energy ratio of the micro-motion frequency band, the micro-motion main frequency, the displacement amplitude, the initial phase, the average acceleration, the interaction distance, and the standby determination result determined by these quantities. The objects processed by the analysis module are all real pose data reflecting the spatial state of the head, hand, or virtual workpiece, so its output results can directly determine the subsequent communication method and synchronization method. The switching module is also deployed on the edge server. It is used to control the switching between normal synchronization state, micro-motion standby state and wake-up transition state based on the micro-motion standby state determination result, and to send switching control information to the user terminal by sending high-frequency pose messages or connection keep-alive messages. After receiving the judgment output from the analysis module, the switching module first determines whether the state switching condition has been met. If the condition for entering standby is met and the preset confirmation time is satisfied, the switching control for entering standby is issued. If the action wake-up indicator meets the preset action activation threshold condition, the state is switched from micro-motion standby to wake-up transition, and the terminal is controlled to resume high-frequency pose message transmission in the next transmission cycle. The output result of the switching module directly changes the message transmission path of the user terminal. The synchronization module is deployed on the edge server and is used to generate a reconstructed micro-motion pose based on the micro-motion parameters in the connection keep-alive message. The reconstructed micro-motion pose is then merged with the poses of other users and synchronized. Upon receiving the real pose corresponding to the high-frequency pose message sent by the target user, the module stops generating the reconstructed micro-motion pose and resumes real pose synchronization. The synchronization module extracts the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time from the connection keep-alive message and calculates the reconstructed micro-motion pose by combining it with the latest real pose cached in the cache. The pose is then written into the global scene state frame and sent to other user terminals through the communication interface. If the synchronization module detects the first recovered high-frequency pose message from the target user, it terminates parameterized regeneration and replaces the user pose in the global scene state frame with the real pose. With the above system structure, the user terminal is responsible for front-end pose acquisition and selecting the message sending mode according to the state, the edge server is responsible for intermediate analysis, state switching and standby pose regeneration, and the communication interface is responsible for transmitting high-frequency pose messages, connection keep-alive messages and state switching control information between the two. The modules work together in the following order: the acquisition module first generates and sends continuous pose data; the analysis module then makes a standby decision based on the received data; the switching module changes the message sending method according to the decision result; and the synchronization module outputs the real pose or reconstructs the micro-motion pose according to the current state. The system constructed in this way can rewrite the high-frequency pose transmission during standby into low-dimensional micro-motion parameter transmission without changing the basic connection relationship of the multi-user session, and switch back to real high-frequency synchronization when the action is restored. In one application scenario, multiple engineers access the same edge server through their respective user terminals and participate in virtual assembly review. When a user observes a component while holding it at a location far from the target interface, the acquisition module continuously collects the raw pose data of the user's hand end and the virtual interactive object being held, and sends high-frequency pose messages. After performing sliding window analysis on the user's continuous pose data, the analysis module determines that the user meets the micro-motion standby conditions. The switching module then controls the user terminal to switch to sending connection keep-alive messages after a preset confirmation time is met. After receiving the micro-motion parameters in the connection keep-alive messages, the synchronization module generates the reconstructed micro-motion pose of the user during the standby period and merges and synchronizes it with the real poses of other active users. When the user begins to move the part closer to the assembly position and makes a clear starting action, the analysis module determines that the action wake-up index has reached the preset action activation threshold condition based on speed, acceleration, or jerk. The switching module controls the terminal to resume sending high-frequency pose messages. After receiving the first restored real pose message, the synchronization module replaces the original reconstructed micro-motion pose with the real pose and continues to output synchronously according to the real pose. In this way, in the whole system, high-frequency invalid packets are reduced during standby, while timely takeover of the real pose is maintained when the action is restored.

[0031] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An edge data transmission method for synchronous interaction in a metaverse scene, characterized in that, include: S1. The user terminal collects single-frame raw pose data from the pose monitoring nodes on the target user's head, hand ends, and / or the controlled virtual interactive object, arranges them into continuous pose data according to the sampling time, and sends high-frequency pose messages. S2. The edge server performs spectral analysis on the displacement based on continuous pose data and a sliding analysis window to obtain the energy proportion of the micro-motion frequency band and the micro-motion main frequency. It also calculates the velocity, average acceleration, and jerk, and analyzes the interaction distance between the pose monitoring node or the controlled virtual interactive object and the interactive virtual objects in the scene. When the energy proportion of the micro-motion frequency band meets a preset energy threshold condition, the micro-motion main frequency falls within a preset frequency range, the average acceleration falls within a preset acceleration range, and the interaction distance meets a preset distance threshold condition, it is determined to be in a micro-motion standby state; otherwise, it is determined not to be in a micro-motion standby state. S3. The edge server maintains normal synchronization, micro-motion standby, and wake-up transition states; when it is determined to be in micro-motion standby state, it switches to control the terminal to stop sending high-frequency pose messages and periodically send connection keep-alive messages carrying micro-motion parameters, including micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time; when not in this state, it maintains the sending of high-frequency pose messages. S4. The edge server generates and updates the reconstructed micro-motion pose based on the latest real pose cache and the micro-motion parameters, and merges and synchronizes it with the poses of other users; when the action wake-up condition is met based on at least one of velocity, acceleration and jerk, it switches to the wake-up transition state, controls the terminal to resume sending high-frequency pose messages, and replaces the reconstructed micro-motion pose with the real pose corresponding to the first high-frequency pose message received.

2. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 1, characterized in that, Step S1 includes: Collect the three-dimensional position and pose parameters of the target user's head; Collect the three-dimensional position and posture parameters of the target user's two-hand controllers or gesture ends; Based on the current interaction state, collect the pose data of the virtual interactive object currently being controlled by the target user; Record the sampling time information corresponding to each sampling point; For any pose monitoring node, a single frame of raw pose data consisting of position parameters, attitude parameters, and corresponding sampling time information is constructed, and the continuous pose data is formed by arranging them according to the sampling time.

3. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 1, characterized in that, Step S2 involves performing a spectral analysis of the displacement, including: Establish a sliding analysis window consisting of a preset number of sampling points; Using the position of the first frame of the sliding analysis window as a reference, calculate the displacement data of each sampling point within the sliding analysis window relative to the position of the first frame; The displacement data is subjected to spectral analysis to obtain the spectral results corresponding to each analysis frequency point; The energy percentage of the preset micro-motion frequency band within the preset micro-motion frequency band is determined based on the spectrum results. Within the preset micro-motion frequency band, the dominant frequency point is determined, and the micro-motion main frequency, displacement amplitude, and initial phase are obtained.

4. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 3, characterized in that, The acquisition of the determination result of whether the target user is in micro-switch standby state in step S2 includes: The velocity and acceleration are determined based on the changes in sampling points within the sliding analysis window, and the average acceleration is calculated. The interaction distance is obtained based on the distance between the outer bounding body of the virtual interactive object currently controlled by the pose monitoring node or the target user and the outer bounding body of the interactive virtual object in the scene. The energy ratio of the micro-motion frequency band, the micro-motion main frequency, the average acceleration, and the interaction distance are used as joint determination conditions; When the energy ratio of the micro-motion frequency band meets the preset energy threshold condition, the interaction distance meets the preset distance threshold condition, the micro-motion main frequency falls within the preset frequency range, and the average acceleration falls within the preset acceleration range, the target user is determined to be in micro-motion standby state.

5. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 1, characterized in that, Step S3 includes: Maintain normal synchronization, micro-motion standby, and wake-up transition states; When it is determined that the target user is in micro-motion standby mode and the standby determination conditions are met continuously for a preset confirmation time, the system switches from normal synchronization mode to micro-motion standby mode. After entering the micro-motion standby state, the transmission of high-frequency pose messages is stopped, and only the transmission of connection keep-alive messages is retained; The micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time received from the edge server are written into the connection keep-alive message to serve as the state confirmation parameters for the edge server to generate the reconstructed micro-motion pose.

6. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 5, characterized in that, The main direction of the micro-motion is determined based on the cumulative result of the position changes of each adjacent sampling point within the sliding analysis window, and the starting time of the parameter is the moment when the edge server begins to use the micro-motion parameter to generate the reconstructed micro-motion pose.

7. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 1, characterized in that, Step S4, which generates the reconstructed micro-motion pose of the target user based on the micro-motion parameters, includes: Extract the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction, and parameter start time from the micro-motion parameters; Based on the micro-motion main frequency, displacement amplitude, initial phase, micro-motion main direction and parameter start time, the periodic micro-displacement of the pose monitoring node is generated on the basis of the latest real position corresponding to the micro-motion standby state. The periodic micro-displacement is superimposed on the latest true pose cached by the edge server to obtain the reconstructed micro-motion pose, wherein the pose is maintained as the latest true pose corresponding to the micro-motion standby state. The reconstructed micro-motion pose is merged with the real poses of other active users into a global scene state frame, and then synchronized with other users.

8. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 1, characterized in that, Step S4, which involves exiting the micro-motion standby state when the action wake-up condition is met, includes: The action wake-up criterion is determined based on at least one of the velocity, acceleration, and jerk at the current sampling point. When the action wake-up indicator meets the preset action activation threshold condition, the micro-motion standby state is switched to the wake-up transition state, and high-frequency pose message transmission is resumed in the next transmission cycle. After the edge server receives the first high-frequency pose message sent by the target user, it stops reconstructing the micro-motion pose generation. Calculate the spatial deviation vector between the real position corresponding to the first high-frequency pose message and the current reconstructed micro-motion pose. Within a preset number of transition frames, perform smooth interpolation approximation of the pose based on this spatial deviation. After completing the visual smooth transition, fully restore the real pose and output synchronously.

9. The edge data transmission method for synchronous interaction in a metaverse scene according to claim 4, characterized in that, The preset distance threshold is determined based on the allowable spatial error tolerance of the current target interaction task. The preset distance threshold is obtained by multiplying the allowable spatial error tolerance by an amplification factor, which is preset according to the task accuracy level.

10. An edge data transmission system for synchronous interaction in a metaverse scene, used to implement the method described in any one of claims 1 to 9, characterized in that, It includes a user terminal, an edge server, and a communication interface located between the user terminal and the edge server; The user terminal is equipped with a data acquisition module, which is used to acquire the original pose data of the target user, construct continuous pose data of the pose monitoring node, and send high-frequency pose messages or connection keep-alive messages. The edge server is equipped with an analysis module, which is used to perform displacement analysis, spectrum analysis, velocity and acceleration calculation, and interaction distance analysis on the continuous pose data based on a sliding analysis window, and to obtain the micro-motion standby state determination result. The edge server is equipped with a switching module, which is used to control the switching between normal synchronization, micro-motion standby and wake-up transition states based on the micro-motion standby state determination result, and to send switching control information to the user terminal via high-frequency pose message sending or connection keep-alive message sending. The edge server is equipped with a synchronization module, which is used to generate a reconstructed micro-motion pose based on the micro-motion parameters in the connection keep-alive message, merge the reconstructed micro-motion pose with other user poses and synchronize them, and stop generating the reconstructed micro-motion pose and restore the real pose synchronization after receiving the real pose corresponding to the high-frequency pose message sent by the target user.

Images