A digital human driving-based physical model interactive display method and system

Abstract

The application discloses a kind of based on digital person drive's real model interactive display method and system, belong to intelligent interaction technical field.For motion capture, voice drive, rendering synchronization and model servo control etc.multimodal signal link configuration uniform time base reference source and synchronization deviation allowable interval, along each link for signal data additional time stamp mark;Through the deviation value of actual time stamp and expected time stamp of each link measurement mark time base deviation event;Execute isochronous section continuous window slice division, based on event occurrence intensity value definition signal disturbance concentrated section;Determine the intersection of any two link disturbance sections to determine the time base correlation;Extract amplitude value track sequence of associated link pair and distinguish strong correlation and weak correlation link pair based on waveform similarity;Strong correlation link pair is extracted to statistical characteristic parameter, constructs disturbance characteristic vector to determine disturbance mode and executes adaptive correction.The application realizes the real-time monitoring, correlation analysis and automatic correction of the time base deviation of multiple signal links.

Description

Technical Field

[0001] This invention relates to the field of intelligent interactive technology, specifically to a method and system for interactive display of physical models driven by digital humans. Background Technology

[0002] Interactive displays of physical models driven by digital humans are a novel display method that integrates virtual digital human technology with physical model control technology. It is widely used in scenarios such as digital displays of cultural heritage, virtual assembly demonstrations of industrial products, and interactive experiences for science education. Its typical workflow is as follows: motion capture equipment captures the body movements and facial expressions of the person driving the display in real time. This data is then processed by the digital human driving platform to generate motion and voice signals for the virtual digital human. Simultaneously, the motion commands are transmitted to the servo control system of the physical model, driving the physical model to make coordinated and consistent motion responses with the virtual digital human, achieving an immersive interactive experience that blends the virtual and real worlds.

[0003] The aforementioned interactive display system typically includes multiple heterogeneous signal links: a motion capture signal link is responsible for collecting and driving the movement data of the personnel; a voice-driven signal link is responsible for collecting and synthesizing voice commands; a rendering synchronization signal link is responsible for generating the visual image of the virtual digital human; and a model servo control signal link is responsible for driving the physical model to perform mechanical movements. Each signal link originates from different signal source devices, travels through different transmission paths, and ultimately converges at a unified display terminal. Because each signal link relies on different clock sources, has different transmission delay characteristics, and different processing loads, time synchronization deviations between the driving signals of each link are unavoidable.

[0004] Currently, the synchronous control of digital human-driven physical model interactive display systems mainly relies on the following methods: First, each signal source device operates independently, relying on its own clock crystal oscillator to drive signal acquisition and output. The lack of a unified time reference between different devices leads to inherent and unavoidable cumulative time base deviations between the drive signals of each link. Operators can only visually observe whether the virtual digital human's movements are synchronized with the physical model's movements, making quantitative deviation monitoring impossible.

[0005] Second, when existing systems experience deterioration such as audio-visual asynchrony or delayed actions, they typically rely on manual troubleshooting to check each signal link one by one. This troubleshooting process is time-consuming and depends on the experience of the operators, making it difficult to detect and locate the source of synchronization deviation in real time during interactive displays.

[0006] Third, the time base deviations of different signal links often do not occur independently. When a signal source device experiences clock drift, a transmission link experiences network jitter, or a processing unit experiences computational overload, the synchronization performance of multiple signal links may be affected simultaneously. However, existing technologies lack systematic analytical methods for the coupling relationship of time base deviations between different signal links, making it impossible to determine whether the simultaneous misalignment of multiple links is caused by the same root cause or by independent factors, and also making it impossible to take targeted compensation measures accordingly.

[0007] Fourth, the existing system lacks adaptive capability in handling synchronization deviations. Every time a synchronization problem occurs, manual intervention is required for adjustment. It cannot automatically identify deviation patterns and match corresponding correction schemes based on historical experience, resulting in low system operating efficiency. Summary of the Invention

[0008] The purpose of this invention is to provide a method for interactive display of physical models based on digital human-driven technology, so as to solve the problems mentioned in the background art.

[0009] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A method for interactive display of physical models based on digital human-driven approaches, comprising the following steps: Step S1: Construct a multimodal signal link for interactive display of digital human-driven physical model, configure a unified time base reference source and synchronization deviation allowable range for each signal link, add timestamp marks to the signal data of the signal source device along each signal link, and establish the signal timestamp sequence of each signal link; Step S2: During the interactive display operation, the instantaneous time base deviation value of the timestamp mark actually transmitted in each signal link is continuously calculated based on the time base reference source, and the time base offset event is marked based on the instantaneous time base deviation value to form a time base offset event log for each signal link. Step S3: Perform continuous window slicing of the time base offset event log for equal time periods, count the event occurrence intensity value of each window slice, and mark the window slice as an active slice; merge and define the active slices arranged continuously in the same signal link as the signal disturbance concentration section; Step S4: For any two signal links with concentrated signal disturbances, if there is a segment crossing on the time base reference source, it is determined that there is a time base correlation between the two signal links; based on the time base correlation, the crossing duration of the concentrated signal disturbances is counted, and a crossing duration distribution record is generated. Step S5: For signal link pairs identified as having a time base correlation, extract the instantaneous time base deviation value change trajectory in each signal disturbance concentration segment from the time base offset event log of each signal link in the signal link pair, and form the first and second amplitude trajectory sequences respectively; based on the first and second amplitude trajectory sequences, analyze the waveform similarity value of the signal link pair to distinguish between strongly correlated link pairs and weakly correlated link pairs; Step S6: For strongly correlated link pairs, extract the statistical feature parameters of the two signal links in each amplitude trajectory sequence, and combine them to construct the perturbation feature vector of the strongly correlated link pairs to determine the perturbation mode and guide the signal data adjustment method; for weakly correlated link pairs, no signal data adjustment is performed.

[0010] Furthermore, the specific implementation process of constructing the multimodal signal link for the interactive display of the digital human-driven physical model in step S1 includes: A unified addressing is implemented for the signal source devices in the interactive display process of digital human-driven physical models. The signal source devices include motion capture sensor arrays, voice pickup terminals, graphics rendering servers, and model servo drivers. Each signal source device is assigned a dedicated signal link, and all signal links share the same time base reference source. Specifically, motion capture sensor arrays are assigned motion capture signal links, voice pickup terminals are assigned voice drive signal links, graphics rendering servers are assigned rendering synchronization signal links, and model servo drivers are assigned model servo control signal links. A timestamp marker is deployed at the output port of each signal link. The timestamp marker is used to add a timestamp marker to each frame of signal data based on the time base reference source when the signal data of the driving signal source device is injected into the distribution network. A synchronization deviation allowable range is set for each signal link. The width of the synchronization deviation allowable range is determined based on the service quality requirements of the signal data transmitted by the corresponding signal link. The allowable range width of the motion capture signal link and the voice drive signal link is narrower than the allowable range width of the rendering synchronization signal link and the model servo control signal link.

[0011] Furthermore, the specific implementation process of labeling the time base offset event based on the instantaneous time base deviation value in step S2 includes: At the start of the interactive display cycle, the expected timestamp sequence of each signal link is constructed using the integer time of the time base reference source as the anchor point. For each signal link, the actual driving signal frame and its attached signal timestamp sequence transmitted by each signal link are received in real time. The difference between the timestamp marker in the signal timestamp sequence and the anchor point in the expected timestamp sequence of the corresponding sequence number of the driving signal frame is calculated to obtain the instantaneous time base deviation value of the driving signal frame. If the instantaneous time base deviation value of any signal link exceeds the synchronization deviation allowable range of that signal link, the drive signal frame corresponding to the instantaneous time base deviation value is marked as a time base offset event, and the instantaneous time base deviation value and occurrence time of the time base offset event are recorded; if the instantaneous time base deviation value falls back to within the synchronization deviation allowable range, the time base offset event is marked as terminated, and the duration after the termination of the time base offset event is recorded.

[0012] Furthermore, the specific implementation process for defining the concentrated signal disturbance section in step S3 includes: Set the window slice width and window step length, and advance the sliding window on the time base offset event log by the window step length to divide the time base offset event log into several continuous window slices of equal width. For each window slice of each signal link, the count of time base offset events falling within the window slice is accumulated, and the ratio of the count to the width of the window slice is used as the event occurrence intensity value of each window slice; if the event occurrence intensity value exceeds a preset intensity threshold, then each window slice is marked as an active slice. For active slices that are continuously arranged in the same signal link, merge and consolidate them, and define the consolidated continuous time period as the concentrated signal disturbance section of the same signal link.

[0013] Furthermore, the specific implementation process for identifying the existence of a time base correlation in step S4 includes: For any two signal links, obtain the set of signal disturbance concentration segments for each of the two signal links; traverse and compare the signal disturbance concentration segments in the two sets of signal disturbance concentration segments. If any pair of signal disturbance concentration segments forms a segment intersection on the time base reference source, it is determined that there is a time base correlation relationship between the two signal links. For signal link pairs that are identified as having a time base correlation, the interactive display operation cycles of all segments with crossover are counted to generate a crossover cycle record set for the signal link pairs; the crossover duration length of the concentrated signal disturbance segment of any two signal links in each crossover cycle is counted to generate a crossover duration distribution record for the signal link pairs.

[0014] Furthermore, the specific implementation process for distinguishing between strongly correlated link pairs and weakly correlated link pairs in step S5 includes: For signal link pairs with time base correlation, retrieve the cross-time distribution record of the signal link pair, calculate the average cross-duration length within the range of the cross-duration distribution record, and filter out the signal disturbance concentration segments corresponding to the cross-duration length being greater than or equal to the average cross-duration length. The filtered signal disturbance concentration segments are taken as valid cross-duration periods, and the unfiltered signal disturbance concentration segments are taken as invalid cross-duration periods. For one signal link in the signal link pair, the instantaneous time base deviation value change trajectory within each effective cross period is extracted from the time base offset event log of that signal link to form a first amplitude trajectory sequence; for the other signal link in the signal link pair, a corresponding second amplitude trajectory sequence is formed. Using the first amplitude trajectory sequence as a reference, the time base reference source is scaled and matched on the second amplitude trajectory sequence to find the optimal matching path between the first amplitude trajectory sequence and the second amplitude trajectory sequence, and the cumulative deformation cost along the optimal matching path is calculated. The inverse normalized value of the cumulative deformation cost is used as the waveform similarity value of the signal link pair. The higher the waveform similarity value, the more consistent the offset amplitude change pattern of the two links is. A preset threshold for waveform similarity is set. If the waveform similarity value is not lower than the threshold, the signal link pair is classified as a strongly correlated link pair. If the waveform similarity value is lower than the threshold, the signal link pair is classified as a weakly correlated link pair.

[0015] Furthermore, the specific implementation process for determining the disturbance mode in step S6 includes: For signal link pairs classified as strongly correlated link pairs, statistical feature parameters of the two signal links in each amplitude trajectory sequence are extracted. The statistical feature parameters include the mean, dispersion, and extreme values ​​of the instantaneous time base deviation. The statistical feature parameters of the two signal links are combined and encoded into the perturbation feature vector of the signal link pair. The perturbation feature vector is input into a preset perturbation pattern classifier and matched with the perturbation feature vectors corresponding to different perturbation patterns stored in the perturbation pattern classifier, and the perturbation pattern with the best matching result is output. Based on the determined disturbance mode, the corresponding drive parameter correction scheme is retrieved from the preset compensation strategy library, which stores mapping entries between each disturbance mode and the correction operation; the retrieved drive parameter correction scheme is executed to adjust the signal data generated by the signal link pair when driving the signal source device; For signal link pairs classified as weakly correlated link pairs, no drive parameter correction operation is triggered.

[0016] A digital human-driven interactive display system for physical models, comprising: One or more processors; a digital human-driven platform for scheduling one or more computer programs, which, when executed by the one or more processors, enable the digital human-driven platform to implement the steps of the digital human-driven physical model interactive display method; and a computer-readable storage medium for carrying a computer program that can be executed by a processor to implement the steps of the digital human-driven physical model interactive display method.

[0017] Furthermore, the digital human-driven platform includes: The multimodal signal link construction module is used to construct multimodal signal links for interactive display of digital human-driven physical models. It configures a unified time base reference source and synchronization deviation allowable range for each signal link, adds timestamp markers to the signal data of the signal source device along each signal link, and establishes the signal timestamp sequence of each signal link. The time base offset event log generation module is used to continuously calculate the instantaneous time base offset value of the timestamp mark actually transmitted in each signal link based on the time base reference source during the interactive display operation, and mark the time base offset event based on the instantaneous time base offset value to form the time base offset event log of each signal link. The signal disturbance concentration section definition module is used to divide the time base offset event log into continuous window slices of equal time periods, count the event occurrence intensity value of each window slice, and mark the window slice as an active slice; and merge and define the active slices arranged continuously in the same signal link as a signal disturbance concentration section. The time base correlation determination module is used to determine that if there is a segment intersection on the time base reference source for any two signal links with concentrated signal disturbances, then the two signal links are identified as having a time base correlation relationship; based on the time base correlation relationship, the intersection duration of the concentrated signal disturbance segment is calculated and an intersection duration distribution record is generated; The link correlation differentiation module is used to extract the instantaneous time base deviation value change trajectory in each signal disturbance concentration segment from the time base offset event log of each signal link in the signal link pair that is identified as having a time base correlation relationship, and form a first and a second amplitude trajectory sequence respectively; based on the first and second amplitude trajectory sequences, the waveform similarity value of the signal link pair is analyzed to distinguish between strongly correlated link pairs and weakly correlated link pairs; The disturbance mode determination and adjustment module is used to extract the statistical feature parameters of the two signal links in each amplitude trajectory sequence for strongly correlated link pairs, and combine them to construct the disturbance feature vector of the strongly correlated link pairs to determine the disturbance mode and guide the signal data adjustment method; for weakly correlated link pairs, no signal data adjustment is performed.

[0018] The beneficial effects achieved by this invention are: (1) By configuring a unified time base reference source for all signal links in the digital human-driven physical model interactive display system and deploying time stamp markers at the output ports of each signal link to add time stamp markers to each frame of signal data, the quantifiable synchronous monitoring of multi-mode signal links under a unified time base is realized, overcoming the technical defects of independent operation of each signal source device and lack of a unified time base reference in the existing technology.

[0019] (2) In view of the coexistence of occasional signal jitter and persistent time base drift in the interactive display of digital human-driven physical model, the event occurrence intensity analysis of continuous window slices of equal time period can effectively identify and filter out occasional offsets caused by instantaneous equipment disturbances, random network jitters, etc., and effectively extract the concentrated segment of signal disturbances reflecting persistent time base misalignment, providing a reliable time base for subsequent link correlation analysis.

[0020] (3) In view of the characteristic that the time base deviation between different signal links may have the same source coupling, a waveform similarity analysis method based on time axis elastic matching is introduced. The inverse normalized value of the cumulative deformation cost is used as the waveform similarity measure of the link pair. It can quantify the consistency of the amplitude change pattern of the two signal links, effectively distinguish between strongly correlated link pairs and weakly correlated link pairs, and avoid misjudging the concurrent offset caused by independent factors as the same source coupling.

[0021] (4) By extracting the statistical feature parameters of strongly correlated link pairs and constructing the perturbation feature vector, and matching and identifying it with the preset perturbation mode classifier, the perturbation mode category to which the current time base perturbation belongs can be automatically determined, and the corresponding driving parameter correction scheme can be matched from the compensation strategy library according to the identification result, thus realizing automatic diagnosis and adaptive correction of synchronization misalignment. Attached Figure Description

[0022] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0023] Figure 1 This is a schematic diagram illustrating the steps of a digital human-driven interactive display method for physical models according to the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] This invention provides a method and system for interactive display of physical models driven by digital humans, addressing the technical problems in existing digital human-driven physical model interactive display systems, such as the lack of unified time base monitoring for multiple signal links, difficulty in automatically identifying time base deviation coupling relationships between links, and inability to achieve adaptive correction. Specifically, this invention configures a unified time base reference source and a synchronization deviation allowable interval for each signal link, and uses timestamp marking to quantify and monitor the time base deviation of each link's driving signal; it employs event intensity analysis using continuous window slices of equal time intervals to filter out occasional disturbances and extract concentrated signal disturbance segments; it utilizes a time axis elastic matching method to quantify the similarity of the amplitude changes between links, distinguishing between strongly correlated and weakly correlated link pairs; and it identifies disturbance patterns and performs adaptive correction through disturbance feature vector matching. Based on the above technical framework, this invention achieves the technical effects of real-time monitoring, coupling analysis, and adaptive correction of time base deviations in multimodal signal links within a digital human-driven physical model interactive display system.

[0026] Please see Figure 1 In this first embodiment, a method for interactive display of physical models driven by digital humans is provided, which includes the following steps S1 to S6.

[0027] Step S1: Construct a multimode signal link: Construct a multimodal signal link for interactive display of digital human-driven physical models, configure a unified time base reference source and synchronization deviation allowable range for each signal link, add timestamp markers to the signal data of the signal source device along each signal link, and establish a signal timestamp sequence for each signal link.

[0028] The specific execution process for this step is as follows: First, a unified addressing is implemented for the signal source devices in the interactive display process of digital human-driven physical model. The signal source devices include motion capture sensor arrays, voice pickup terminals, graphics rendering servers, and model servo drivers, which correspond to four types of signal sources: motion capture sensor arrays are used to collect limb movement data of the driving personnel, voice pickup terminals are used to collect voice commands of the driving personnel, graphics rendering servers are used to generate visual images of virtual digital humans, and model servo drivers are used to drive physical models to perform mechanical actions.

[0029] Each signal source device is assigned a dedicated signal link. Specifically, the motion capture sensor array is assigned a motion capture signal link, the voice pickup terminal is assigned a voice drive signal link, the graphics rendering server is assigned a rendering synchronization signal link, and the model servo driver is assigned a model servo control signal link. All signal links share the same time base reference source, which is provided by a high-precision clock generator with a clock accuracy better than ±1ppm, ensuring that all signal links maintain a unified time base reference during long-term operation.

[0030] Then, timestamp markers are deployed at the output ports of each signal link. The timestamp markers are used to add timestamps to each frame of signal data based on the time base reference source when the signal data driving the signal source device is injected into the distribution network. Specifically, when a frame of signal data is output from its signal source device in any signal link and is ready to be injected into the distribution network, the timestamp marker reads the count value of the current time base reference source and adds this count value as the timestamp of the signal data frame to the header field of the signal frame.

[0031] Finally, a synchronization deviation allowable range is set for each signal link. The width of the allowable range is determined based on the quality of service requirements of the signal data transmitted by the corresponding signal link. Specifically, the signals carried by the motion capture signal link and the voice-driven signal link have extremely high real-time requirements. Motion capture data needs to be transmitted and mapped within milliseconds, and the voice-driven signal also needs to be precisely synchronized with the lip-sync animation. If the deviation exceeds the allowable range, it will lead to perceptible audiovisual asynchrony. Therefore, the allowable range width for the motion capture signal link and the voice-driven signal link is set to ±5ms. The signals carried by the rendering synchronization signal link and the model servo control signal link have relatively relaxed real-time requirements. The frame rate fluctuation of the graphics rendering screen has a high visually perceptible threshold, and the mechanical response of the physical model servo control itself has inertial delay. Therefore, the allowable range width for the rendering synchronization signal link and the model servo control signal link is set to ±20ms.

[0032] Step S2: Generate time base offset event log: During the interactive demonstration, the instantaneous time base deviation value of the timestamp mark actually transmitted in each signal link is continuously calculated based on the time base reference source, and the time base offset event is marked based on the instantaneous time base deviation value to form the time base offset event log of each signal link.

[0033] The specific execution process for this step is as follows: At the start of the interactive display cycle, the desired timestamp sequence for each signal link is constructed using the integer timescale of the time base reference source as the anchor point. The integer timescale is a discrete time point marked on the time base reference source according to a fixed time scale. Let the master clock period of the time base reference source be... (For example =10ms), then the integer time-stamped time sequence is 0ms, 10ms, 20ms, 30ms... and so on. For any signal link, the expected timestamp of its nth frame of signal data is n× The system pre-constructs the desired timestamp sequence for each signal link according to the above rules when the system starts its operating cycle.

[0034] For each signal link, the actual driving signal frame and its attached signal timestamp sequence transmitted by the signal link are received in real time. The difference between the timestamp marker in the signal timestamp sequence and the anchor point in the expected timestamp sequence of the corresponding sequence number of the driving signal frame is calculated to obtain the instantaneous time base deviation value of the driving signal frame.

[0035] Taking the motion capture signal link as an example, the specific calculation process is explained. Assume the expected timestamp of the nth frame signal data in the motion capture signal link is... =n×10ms (expected to arrive at n×10ms); the timestamp marker reads the actual timestamp based on the time base reference source when the signal data of this frame is actually injected into the distribution network. =n×10ms+8ms, then the instantaneous time base deviation of the signal frame is ΔT(n)=T_act(n)-T_exp(n)=8ms.

[0036] If the instantaneous time base deviation value of any signal link exceeds the preset synchronization deviation allowable range of the signal link, the drive signal frame corresponding to the instantaneous time base deviation value is marked as a time base offset event, and the instantaneous time base deviation value and occurrence time of the time base offset event are recorded; if the instantaneous time base deviation value falls back to within the synchronization deviation allowable range, the time base offset event is marked as terminated, and its duration is recorded.

[0037] Continuing with the motion capture signal link example above, the allowed synchronization deviation range for this signal link is [-5ms, +5ms]. If the instantaneous time base deviation of the signal data in the nth frame is +8ms, exceeding the upper limit of the allowed range of +5ms, then the nth frame is marked as the starting frame of a time base offset event, with an offset amplitude of 8ms recorded, and the occurrence time is the actual arrival time of the nth frame (i.e., the actual timestamp). (corresponding time); thereafter, if the instantaneous time base deviation value of the (n+1)th frame is +9ms, the value of the (n+2)th frame is +7ms, and the value of the (n+3)th frame is +4ms (falling within the allowable range), then the time base offset event spans from the nth frame to the (n+2)th frame, and the event ends at the (n+3)th frame; the system records the start time (the time the nth frame arrives), the end time (the time the (n+3)th frame arrives), and the duration (the end time minus the start time) of the time base offset event.

[0038] For each signal link, all marked time base offset events are arranged in chronological order to form a time base offset event log for that signal link. Each log record contains the following fields: event number, start frame number, end frame number, offset amplitude sequence (instantaneous time base offset values ​​of each frame during the duration of the event), start time, end time, and duration.

[0039] Step S3: Define the concentrated area of ​​signal disturbance. The time base offset event log is divided into continuous window slices of equal time period, and the event occurrence intensity value of each window slice is counted to mark the window slice as an active slice; the active slices arranged continuously in the same signal link are merged and defined as the signal disturbance concentration section.

[0040] The specific execution process for this step is as follows: Set the window slice width W and window step length L. Advance the sliding window on the time base offset event log using the window step length L to divide the time base offset event log into several continuous window slices of equal width. The window step length L is 1 / 3 of the window slice width W. The window slice width W is determined based on the signal link type: for motion capture signal links and voice-driven signal links, the time base offset is usually caused by fluctuations in signal acquisition frequency or transmission delay jitter, so the window slice width is set to 500ms; for rendering synchronization signal links, the time base offset is usually caused by fluctuations in graphics rendering load, so the window slice width is set to 1000ms; for model servo control signal links, the time base offset is usually caused by backlog in the servo driver's control command queue, so the window slice width is set to 800ms.

[0041] For each window slice of each signal link, the count of time base offset events falling within that window slice is accumulated. Specifically, if the start time of a certain time base offset event falls within the time range of a certain window slice, the event is counted in the offset event count of that window slice, and the ratio of the count to the width of the window slice is used as the event occurrence intensity value of that window slice.

[0042] Taking a motion capture signal link as an example, let the window slice width W = 500ms, the window step length L ≈ 167ms, and the time range of the k-th window slice be []. , +500ms]; Count the number of starting frames marked as time base offset events within this time range, assuming it is... =12. Therefore, the event occurrence intensity value for this window slice is... = / W=12 / 500ms=24 items / second.

[0043] A preset intensity threshold θ is set based on the statistical distribution of event intensity of each signal link in historical normal operation batches. Specifically, the system collects time base offset event logs for 20 consecutive interactive display cycles under normal operation, calculates the event intensity value of each window slice, and takes the 90th percentile of these intensity values ​​as the intensity threshold of that signal link. If the intensity exceeds the preset intensity threshold θ, the window slice is marked as an active slice; otherwise, it is not marked.

[0044] For consecutive active slices in the same signal link, merge and consolidate them. Specifically, if the k-th, k+1-th, and k+2-th window slices are all active slices and they are consecutively arranged on the time axis (i.e., the end time of the k-th window slice is the same as the start time of the k+1-th window slice, and so on), then these three slices are merged into a continuous active time period. The consolidated continuous time period is defined as the signal disturbance concentration segment of the signal link, and the start and end boundary times of the signal disturbance concentration segment are extracted.

[0045] During the interactive display of digital human-driven physical models, the occurrence of time base offset events has two modes: one is an occasional offset caused by instantaneous network jitter, occasional device interruptions, etc., which is sparsely distributed on the time axis, and the event occurrence intensity value of each window slice is lower than the threshold; the other is a continuous offset caused by time base drift of signal source devices, continuous congestion of transmission links, continuous overload of rendering servers, etc., which is densely clustered on the time axis, and the event occurrence intensity value of multiple consecutive window slices exceeds the threshold, forming a concentrated signal disturbance segment after merging.

[0046] Step S4: Determine the time base correlation: For any two signal links with concentrated signal disturbances, if there is a segment crossing on the time base reference source, then a time base correlation relationship is identified between any two signal links. Based on the time base correlation relationship, the crossing duration of the concentrated signal disturbance segment is calculated, and a crossing duration distribution record is generated.

[0047] The specific execution process for this step is as follows: For any two signal links (denoted as link α and link β), obtain the set of concentrated signal disturbance segments P(α) for link α and the set of concentrated signal disturbance segments P(β) for link β; traverse and compare the concentrated signal disturbance segments in the two sets. If any pair of concentrated signal disturbance segments exists... , If a segment intersection occurs on the time axis, it is determined that there is a time base correlation between the two signal links. The criterion for determining a segment intersection is that the intersection of the two segments is not empty. For example, suppose the signal disturbance concentration segment of the motion capture signal link is P(α) = [10:05:00.000, 10:05:17.000], and the signal disturbance concentration segment of the rendering synchronization signal link is P(β) = [10:05:03.000, 10:05:20.000]; the intersection of P(α) and P(β) is [10:05:03.000, 10:05:17.000], the intersection duration is 14 seconds, which is not empty, therefore it is determined that there is a time base correlation between the motion capture signal link and the rendering synchronization signal link.

[0048] For signal link pairs identified as having a time base correlation, all interactive display operation cycles with segment crossovers are counted to generate a crossover cycle record set for the signal link pair. Specifically, taking multiple consecutive interactive display operation cycles as units, the link pair is checked cycle by cycle to see if there is a segment crossover. Assuming that in 10 consecutive interactive display operation cycles, link α and link β have a crossover in the concentrated signal disturbance segment in 6 of the cycles, the crossover cycle record set contains the identification information (such as cycle number) of these 6 cycles.

[0049] Simultaneously, the cross duration length of the concentrated signal disturbance sections of the two signal links in each cross cycle is statistically analyzed to generate a cross duration distribution record for the signal link pair. The cross duration length refers to the intersection length of the concentrated signal disturbance sections of the two signal links on the time axis. Taking P(α)=[10:05:00.000, 10:05:17.000] and P(β)=[10:05:03.000, 10:05:20.000] as an example, the intersection is [10:05:03.000, 10:05:17.000], and the cross duration length is 14 seconds. The cross duration lengths in all 6 cross cycles are statistically analyzed to form a cross duration distribution record, for example {12, 8, 15, 6, 10, 9}, in seconds, with an average of (12+8+15+6+10+9) / 6=10 seconds.

[0050] Step S5: Distinguish between strongly correlated link pairs and weakly correlated link pairs: For signal link pairs identified as having a time base correlation, for each signal link in the signal link pair, the instantaneous time base deviation value change trajectory in each signal disturbance concentration segment is extracted from the time base offset event log of the signal link, forming the first and second amplitude trajectory sequences respectively; based on the first and second amplitude trajectory sequences, the waveform similarity value of the signal link pair is analyzed to distinguish between strongly correlated link pairs and weakly correlated link pairs.

[0051] The specific execution process for this step is as follows: First, for signal link pairs (link α and link β) with time base correlation, retrieve the cross duration distribution record of the link pair, and calculate the mean of the cross duration within the range of the cross duration distribution record. The mean is (12+8+15+6+10+9) / 6 = 10 seconds.

[0052] Then, the signal disturbance concentration segments corresponding to the cross duration length being greater than or equal to the mean cross duration length are selected and used as valid cross periods; signal disturbance concentration segments with cross duration length less than the mean are filtered as invalid cross periods.

[0053] For link α, the instantaneous time base deviation value change trajectory within the effective cross period is extracted from the time base offset event log of link α. Specifically, the instantaneous time base deviation value recorded for each time base offset event within the time period is extracted in chronological order. Assuming that a total of 5 time base offset events occur within the effective cross period, and their instantaneous time base deviation values ​​are +6ms, +9ms, +12ms, +8ms, and +5ms respectively, the first amplitude trajectory sequence is formed.

[0054] Similarly, for link β, the instantaneous time base offset value change trajectory within the same effective cross period is extracted from the time base offset event log of link β. Assuming that link β has a total of 4 time base offset events within the effective cross period, and their instantaneous time base offset values ​​are +7ms, +10ms, +11ms, and +6ms respectively, a second amplitude trajectory sequence is formed.

[0055] It should be noted that, due to the different transmission paths and processing logic of the signals carried by the two signal links, the frequency and timing of their time base offset events may not be completely consistent.

[0056] Using the first amplitude trajectory sequence [6, 9, 12, 8, 5] as a reference, a time base reference source scaling match is performed on the second amplitude trajectory sequence [7, 10, 11, 6] to find the first amplitude trajectory sequence. With the second amplitude trajectory sequence The optimal matching path between them.

[0057] The specific calculation process for scaling matching is as follows: First, construct an m x n distance matrix D (m=5, n=4), where the matrix element D(i, j) represents... The i-th point and The absolute value of the amplitude difference between the j-th points, i.e.: D(i,j)=| - |; =[6, 9, 12, 8, 5], =[7, 10, 11, 6]; Distance matrix D (row correspondence) The 5 points, corresponding to the column (4 points) D(1,1)=|6-7|=1, D(1,2)=|6-10|=4, D(1,3)=|6-11|=5, D(1,4)=|6-6|=0; D(2,1)=|9-7|=2, D(2,2)=|9-10|=1, D(2,3)=|9-11|=2, D(2,4)=|9-6|=3; D(3,1)=|12-7|=5, D(3,2)=|12-10|=2, D(3,3)=|12-11|=1, D(3,4)=|12-6|=6; D(4,1)=|8-7|=1, D(4,2)=|8-10|=2, D(4,3)=|8-11|=3, D(4,4)=|8-6|=2; D(5,1)=|5-7|=2, D(5,2)=|5-10|=5, D(5,3)=|5-11|=6, D(5,4)=|5-6|=1; Then, the cumulative cost matrix C is calculated, and its recursive formula is as follows: C(i,j)=D(i,j)+min{C(i-1,j-1),C(i-1,j),C(i,j-1)}; Calculate the cumulative cost matrix C item by item; the cumulative deformation cost is C(m,n)=C(5,4)=6.

[0058] The inverse normalized value of the cumulative deformation cost is used as the waveform similarity value of the signal link pair, that is: Sim(α,β)=1 / (1+C(m,n))=1 / (1+6)≈0.143; A preset threshold for waveform similarity (typically set to 0.5) is set. If the waveform similarity value is not lower than the threshold, the signal link pair is classified as a strongly correlated link pair. If the waveform similarity value is lower than the threshold, it is classified as a weakly correlated link pair. Since 0.143 < 0.5, the signal link pair (link α and link β) is classified as a weakly correlated link pair.

[0059] If the waveform similarity value is higher than the threshold, it is classified as a strongly correlated link pair and proceeds to the disturbance pattern recognition and correction process in step S6.

[0060] Step S6: Determine the disturbance mode and perform adjustments: For strongly correlated link pairs, statistical feature parameters of the two signal links in each amplitude trajectory sequence are extracted and combined to construct the perturbation feature vector of the strongly correlated link pairs to determine the perturbation mode and guide the signal data adjustment method; for weakly correlated link pairs, no signal data adjustment is performed.

[0061] Suppose that after processing in step S5, the waveform similarity value Sim(γ, δ) = 0.85 > 0.5 for a strongly correlated link pair (link γ and link δ), and it is classified as a strongly correlated link pair; its first amplitude trajectory sequence =[3, 8, 15, 22, 18, 10, 5], Second amplitude trajectory sequence =[4, 9, 16, 23, 19, 11, 6], all with 7 sampling points.

[0062] For signal link pairs (link γ and link δ) that are classified as strongly correlated link pairs, the statistical feature parameters of the two signal links in each amplitude trajectory sequence are extracted; the statistical feature parameters include the mean, dispersion and extreme values ​​of the instantaneous time base deviation values. The statistical characteristic parameters of the two signal links are combined and encoded into the disturbance characteristic vector of the signal link pair; The perturbation feature vector is input into a preset perturbation pattern classifier and matched with the perturbation feature vectors corresponding to different perturbation patterns stored in the perturbation pattern classifier. The perturbation pattern with the best matching result is output.

[0063] The perturbation pattern classifier pre-stores multiple perturbation pattern feature templates, and the perturbation pattern categories corresponding to each template include: Signal source time base drift mode: The feature template shows that the offset amplitude of each link shows a monotonically increasing trend, the mean continues to increase positively, and the typical feature vector features are that the mean of each link deviates significantly from 0, and the extreme value appears at the end of the sequence. Transmission link jitter mode: The feature template shows that the offset amplitude of each link exhibits high-frequency random fluctuations, with a significant increase in dispersion. The typical feature vector features are that the dispersion of each link is large, and the mean is close to 0. Rendering computation overload mode: The feature template shows a significant increase in the offset amplitude of the rendering synchronization signal link while other links are normal. The typical feature vector feature is that only the mean of the rendering synchronization signal link is >10, while the mean of other links is close to 0. Servo action lag mode: The feature template is that the offset amplitude of the servo control signal link of the model is significantly increased while other links are normal. The typical feature vector feature is that only the mean of the servo control signal link is >10, while the mean of other links is close to 0.

[0064] The perturbation pattern classifier calculates the feature space distance (usually Euclidean distance) between the input feature vector and each pattern feature template, and selects the perturbation pattern corresponding to the template with the smallest distance as the optimal matching result output.

[0065] Based on the identified disturbance mode, the corresponding drive parameter correction scheme is retrieved from the preset compensation strategy library; the compensation strategy library stores mapping entries between each disturbance mode and the correction operation. Signal source time base drift mode → Correction operation: Send a time base resynchronization command to the signal source device to reset the sampling clock phase of the signal source device and make it relock the system time base reference source; Transmission link jitter mode → Correction operation: Send parameter adjustment instructions to the buffer nodes in the transmission link to increase the buffer queue depth by 50% from the default value in order to smooth transmission delay jitter; Rendering computation overload mode → Correction operation: Send a frame rate adjustment command to the graphics rendering server to reduce the target rendering frame rate from 60fps to 45fps, and at the same time reduce the rendering detail level parameter to reduce the computational load. Servo motion lag mode → Correction operation: Send control parameter adjustment command to the model servo driver to increase the acceleration parameter by 20% and shorten the control command cycle from 10ms to 5ms.

[0066] The retrieved drive parameter correction scheme is executed to adjust the signal data generated by the signal link when driving the signal source device; specifically, according to the instructions of the above correction operation, a parameter adjustment command is sent to the corresponding device, and the device performs the corresponding parameter change after receiving the command.

[0067] For signal link pairs classified as weakly correlated link pairs, no drive parameter correction operation is triggered.

[0068] It's worth noting that the essential purpose of "distinguishing between strongly correlated and weakly correlated link pairs" is to determine whether the misalignment of two signal links exhibiting simultaneous time base shifts is driven by a common source or by independent factors. When the waveform similarity value of two signal links exceeds a threshold, it means that the instantaneous time base deviation trajectories of the two links are highly consistent. They simultaneously exhibit similar amplitude shifts at similar time points, and the trends of increase and decrease in the shifts are synchronized. This likely reflects that the two links are affected by a common disturbance source. Therefore, intervention should be implemented against this common disturbance source. If the strongly correlated link pair does not trigger correction, the common disturbance source will continue to act, and the time base deviation will continue to accumulate along the time axis, causing the synchronization error between the affected signal links to continuously increase. In interactive display scenarios, this cumulative effect will manifest as a continuously widening lag difference between the virtual digital human's actions and the physical model's actions, and an increased degree of asynchrony between voice and lip movements, until it exceeds the user's acceptable experience threshold, causing a substantial deterioration in the display effect. By correcting this common disturbance factor, the synchronization state of the affected links can be restored simultaneously.

[0069] In this second embodiment: a physical model interactive display system based on digital human-driven system is provided. The system includes one or more processors, a digital human-driven platform and a computer-readable storage medium.

[0070] The digital human-driven platform is used to schedule one or more computer programs, which, when executed by one or more processors, enable the digital human-driven platform to implement the steps of the aforementioned method for interactive display of physical models based on digital human-driven technology; a computer-readable storage medium is used to carry the computer programs, which can be executed by a processor to implement the steps of the aforementioned method.

[0071] Specifically, the digital human-driven platform includes the following functional modules: The multimodal signal link construction module is used to construct multimodal signal links for interactive display of digital human-driven physical models. It configures a unified time base reference source and synchronization deviation allowable range for each signal link, adds timestamp markers to the signal data of the signal source device along each signal link, and establishes the signal timestamp sequence of each signal link.

[0072] The time base offset event log generation module is used to continuously measure the instantaneous time base offset value of the actual timestamp mark transmitted in each signal link based on the time base reference source during the interactive display operation, and mark the time base offset event based on the instantaneous time base offset value to form the time base offset event log of each signal link.

[0073] The signal disturbance concentration section definition module is used to divide the time base offset event log into continuous window slices of equal time periods, count the event occurrence intensity value of each window slice, and mark the window slice as an active slice; and merge and define the active slices arranged continuously in the same signal link as the signal disturbance concentration section.

[0074] The time base correlation determination module is used to determine that if there is a segment crossing on the time base reference source for any two signal links with concentrated signal disturbances, then a time base correlation relationship exists between any two signal links. Based on the time base correlation relationship, the module calculates the crossing duration length of the concentrated signal disturbance segment and generates a crossing duration distribution record.

[0075] The link correlation differentiation module is used to identify signal link pairs with time base correlation. For each signal link in the signal link pair, it extracts the instantaneous time base deviation value change trajectory in each signal disturbance concentration segment from the time base offset event log of the signal link, forming first and second amplitude trajectory sequences respectively. Based on the first and second amplitude trajectory sequences, it analyzes the waveform similarity value of the signal link pair to distinguish between strongly correlated link pairs and weakly correlated link pairs.

[0076] The disturbance mode determination and adjustment module is used to extract the statistical feature parameters of the two signal links in each amplitude trajectory sequence for strongly correlated link pairs, and combine them to construct the disturbance feature vector of the strongly correlated link pairs to determine the disturbance mode and guide the signal data adjustment method; for weakly correlated link pairs, no signal data adjustment is performed.

[0077] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0078] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for interactive display of physical models based on digital human-driven technology, characterized in that, The method includes the following steps: Step S1: Construct a multimodal signal link for interactive display of digital human-driven physical model, configure a unified time base reference source and synchronization deviation allowable range for each signal link, add timestamp marks to the signal data of the signal source device along each signal link, and establish the signal timestamp sequence of each signal link; Step S2: During the interactive display operation, the instantaneous time base deviation value of the timestamp mark actually transmitted in each signal link is continuously calculated based on the time base reference source, and the time base offset event is marked based on the instantaneous time base deviation value to form a time base offset event log for each signal link. Step S3: Perform continuous window slicing of the time base offset event log for equal time periods, count the event occurrence intensity value of each window slice, and mark the window slice as an active slice; merge and define the active slices arranged continuously in the same signal link as the signal disturbance concentration section; Step S4: For any two signal links with concentrated signal disturbances, if there is a segment crossing on the time base reference source, it is determined that there is a time base correlation between the two signal links; based on the time base correlation, the crossing duration of the concentrated signal disturbances is counted, and a crossing duration distribution record is generated. Step S5: For signal link pairs identified as having a time base correlation, extract the instantaneous time base deviation value change trajectory in each signal disturbance concentration segment from the time base offset event log of each signal link in the signal link pair, and form the first and second amplitude trajectory sequences respectively; based on the first and second amplitude trajectory sequences, analyze the waveform similarity value of the signal link pair to distinguish between strongly correlated link pairs and weakly correlated link pairs; Step S6: For strongly correlated link pairs, extract the statistical feature parameters of the two signal links in each amplitude trajectory sequence, and combine them to construct the perturbation feature vector of the strongly correlated link pairs to determine the perturbation mode and guide the signal data adjustment method; for weakly correlated link pairs, no signal data adjustment is performed.

2. The interactive display method for physical models based on digital human-driven technology according to claim 1, characterized in that, The specific execution steps for constructing the multimodal signal link for interactive display of the digital human-driven physical model in step S1 include: A unified addressing is implemented for the signal source devices in the interactive display process of digital human-driven physical models. The signal source devices include motion capture sensor arrays, voice pickup terminals, graphics rendering servers, and model servo drivers. Each signal source device is assigned a dedicated signal link, and all signal links share the same time base reference source. Specifically, motion capture sensor arrays are assigned motion capture signal links, voice pickup terminals are assigned voice drive signal links, graphics rendering servers are assigned rendering synchronization signal links, and model servo drivers are assigned model servo control signal links. A timestamp marker is deployed at the output port of each signal link. The timestamp marker is used to add a timestamp marker to each frame of signal data based on the time base reference source when the signal data of the driving signal source device is injected into the distribution network. A synchronization deviation allowable range is set for each signal link, and the width of the synchronization deviation allowable range is determined according to the quality of service requirements of the signal data transmitted by the corresponding signal link.

3. The interactive display method for physical models based on digital human-driven technology according to claim 1, characterized in that, The specific execution steps for labeling the time base offset event based on the instantaneous time base deviation value in step S2 include: At the start of the interactive display cycle, the expected timestamp sequence of each signal link is constructed using the integer time of the time base reference source as the anchor point. For each signal link, the actual driving signal frame and its attached signal timestamp sequence transmitted by each signal link are received in real time. The difference between the timestamp marker in the signal timestamp sequence and the anchor point in the expected timestamp sequence of the corresponding sequence number of the driving signal frame is calculated to obtain the instantaneous time base deviation value of the driving signal frame. If the instantaneous time base deviation value of any signal link exceeds the synchronization deviation allowable range of that signal link, the drive signal frame corresponding to the instantaneous time base deviation value is marked as a time base offset event, and the instantaneous time base deviation value and occurrence time of the time base offset event are recorded; if the instantaneous time base deviation value falls back to within the synchronization deviation allowable range, the time base offset event is marked as terminated, and the duration after the termination of the time base offset event is recorded.

4. The interactive display method for physical models based on digital human-driven technology according to claim 1, characterized in that, The specific execution steps for defining the concentrated signal disturbance section in step S3 include: Set the window slice width and window step length, and advance the sliding window on the time base offset event log by the window step length to divide the time base offset event log into several continuous window slices of equal width. For each window slice of each signal link, the count of time base offset events falling within the window slice is accumulated, and the ratio of the count to the width of the window slice is used as the event occurrence intensity value of each window slice; if the event occurrence intensity value exceeds a preset intensity threshold, then each window slice is marked as an active slice. For active slices that are continuously arranged in the same signal link, merge and consolidate them, and define the consolidated continuous time period as the concentrated signal disturbance section of the same signal link.

5. The interactive display method for physical models based on digital human-driven technology according to claim 1, characterized in that, The specific execution steps for determining the existence of a time base correlation in step S4 include: For any two signal links, obtain the set of signal disturbance concentration segments for each of the two signal links; traverse and compare the signal disturbance concentration segments in the two sets of signal disturbance concentration segments. If any pair of signal disturbance concentration segments forms a segment intersection on the time base reference source, it is determined that there is a time base correlation relationship between the two signal links. For signal link pairs that are identified as having a time base correlation, the interactive display operation cycles of all segments with crossover are counted to generate a crossover cycle record set for the signal link pairs; the crossover duration length of the concentrated signal disturbance segment of any two signal links in each crossover cycle is counted to generate a crossover duration distribution record for the signal link pairs.

6. The interactive display method for physical models based on digital human-driven technology according to claim 1, characterized in that, The specific execution steps for distinguishing between strongly correlated link pairs and weakly correlated link pairs in step S5 include: For signal link pairs with time-base correlation, retrieve the cross-time distribution records of the signal link pairs. Within the range of the cross duration distribution record, the mean of the cross duration is calculated, and the signal disturbance concentration segment corresponding to the cross duration length being greater than or equal to the mean of the cross duration length is selected. The selected signal disturbance concentration segment is taken as the valid cross period, and the unselected signal disturbance concentration segment is taken as the invalid cross period. For one signal link in the signal link pair, the instantaneous time base deviation value change trajectory within each effective cross period is extracted from the time base offset event log of that signal link to form a first amplitude trajectory sequence; for the other signal link in the signal link pair, a corresponding second amplitude trajectory sequence is formed. Using the first amplitude trajectory sequence as a reference, the time base reference source is scaled and matched on the second amplitude trajectory sequence to find the optimal matching path between the first amplitude trajectory sequence and the second amplitude trajectory sequence, and the cumulative deformation cost along the optimal matching path is calculated. The inverse normalized value of the cumulative deformation cost is used as the waveform similarity value of the signal link pair; A preset threshold for waveform similarity is set. If the waveform similarity value is not lower than the threshold, the signal link pair is classified as a strongly correlated link pair. If the waveform similarity value is lower than the threshold, the signal link pair is classified as a weakly correlated link pair.

7. The interactive display method for physical models based on digital human-driven technology according to claim 1, characterized in that, The specific steps for determining the disturbance mode in step S6 include: For signal link pairs classified as strongly correlated link pairs, statistical feature parameters of the two signal links in each amplitude trajectory sequence are extracted. The statistical feature parameters include the mean, dispersion, and extreme values ​​of the instantaneous time base deviation. The statistical feature parameters of the two signal links are combined and encoded into the perturbation feature vector of the signal link pair. The perturbation feature vector is input into a preset perturbation pattern classifier and matched with the perturbation feature vectors corresponding to different perturbation patterns stored in the perturbation pattern classifier, and the perturbation pattern with the best matching result is output. Based on the determined disturbance mode, the corresponding drive parameter correction scheme is retrieved from the preset compensation strategy library, which stores mapping entries between each disturbance mode and the correction operation; the retrieved drive parameter correction scheme is executed to adjust the signal data generated by the signal link pair when driving the signal source device; For signal link pairs classified as weakly correlated link pairs, no drive parameter correction operation is triggered.

8. A physical model interactive display system based on digital human-driven technology, characterized in that, The physical model interactive display system includes a computer-readable storage medium for carrying a computer program that can be executed by a processor to implement the steps of the digital human-driven physical model interactive display method as described in any one of claims 1 to 7.

9. The interactive display system for physical models based on digital human-driven technology according to claim 8, characterized in that, The physical model interactive display system also includes: One or more processors; A digital human-driven platform is used to schedule one or more computer programs, which, when executed by one or more processors, enable the digital human-driven platform to implement the steps of the digital human-driven physical model interactive display method.

10. A physical model interactive display system based on digital human-driven technology according to claim 9, characterized in that, The digital human-driven platform includes: The multimodal signal link construction module is used to construct multimodal signal links for interactive display of digital human-driven physical models. It configures a unified time base reference source and synchronization deviation allowable range for each signal link, adds timestamp markers to the signal data of the signal source device along each signal link, and establishes the signal timestamp sequence of each signal link. The time base offset event log generation module is used to continuously calculate the instantaneous time base offset value of the timestamp mark actually transmitted in each signal link based on the time base reference source during the interactive display operation, and mark the time base offset event based on the instantaneous time base offset value to form the time base offset event log of each signal link. The signal disturbance concentration section definition module is used to divide the time base offset event log into continuous window slices of equal time periods, count the event occurrence intensity value of each window slice, and mark the window slice as an active slice; and merge and define the active slices arranged continuously in the same signal link as a signal disturbance concentration section. The time base correlation determination module is used to determine that if there is a segment intersection on the time base reference source for any two signal links with concentrated signal disturbances, then the two signal links are identified as having a time base correlation relationship; based on the time base correlation relationship, the intersection duration of the concentrated signal disturbance segment is calculated and an intersection duration distribution record is generated; The link correlation differentiation module is used to extract the instantaneous time base deviation value change trajectory in each signal disturbance concentration segment from the time base offset event log of each signal link in the signal link pair that is identified as having a time base correlation relationship, and form a first and a second amplitude trajectory sequence respectively; based on the first and second amplitude trajectory sequences, the waveform similarity value of the signal link pair is analyzed to distinguish between strongly correlated link pairs and weakly correlated link pairs; The disturbance mode determination and adjustment module is used to extract the statistical feature parameters of the two signal links in each amplitude trajectory sequence for strongly correlated link pairs, and combine them to construct the disturbance feature vector of the strongly correlated link pairs to determine the disturbance mode and guide the signal data adjustment method; for weakly correlated link pairs, no signal data adjustment is performed.

Images