A network conference interaction optimization method based on real-time data analysis

Abstract

The application discloses a network conference interaction optimization method based on real-time data analysis and relates to the technical field of network communication, which comprises the following steps: according to a current conference round state set, pre-sorting each candidate participating terminal to form a pre-authorization shadow queue; according to a continuous speaking candidate sequence, detecting whether a current speaking session entity enters a switching period, and selecting a target switching conference from the continuous speaking candidate sequence to form smooth switching control information when it is detected that the current speaking session entity enters the switching period. Through linkage construction of the pre-authorization shadow queue and the continuous speaking candidate sequence, in combination with the real-time data acquisition and processing capacity of the edge node, the application can complete the connection object screening and order stability processing before the current speaking is finished, so that the switching is established on the basis of continuous turn taking, the switching matching degree, the time sequence continuity and the control convergence efficiency are improved, and low-delay response and cloud load sharing are realized.

Description

Technical Field

[0001] This invention relates to the field of network communication technology, and in particular to a method for optimizing network meeting interaction based on real-time data analysis. Background Technology

[0002] With the development of packet-switched networks, real-time audio and video transmission, edge-side session control, and multi-party interaction protocols, web conferencing has expanded from single media transmission to real-time collaborative sessions covering speaking requests, host confirmation, shared connection, and candidate takeover control. It can complete session status collection, basic ordering authorization, and media distribution during the meeting. Session status collection and preliminary processing can be performed at edge nodes to reduce cloud load and achieve low-latency real-time control.

[0003] However, in scenarios involving concurrent requests from multiple candidate participating terminals, delayed host confirmation, and link fluctuations, existing methods often employ a discrete switching approach of reordering and authorizing after the speech ends. This makes it difficult to establish a stable succession candidate chain in advance before the current speech is released, and it lacks the utilization of edge node collection and processing. This results in switching gaps, connection mismatches, and concentrated session control jitter, affecting the temporal consistency and smoothness of continuous speech. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a network conference interaction optimization method based on real-time data analysis to solve the problems of delayed speech handover and insufficient session control continuity in multi-party network conferences by combining edge computing.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a network conferencing interaction optimization method based on real-time data analysis, comprising: collecting the speaking status of the current speaking session entity and the session interaction status of each candidate participating terminal, and merging them into a current meeting round state set; pre-sorting each candidate participating terminal according to the current meeting round state set to form a pre-authorized shadow queue; performing inertial correction on the pre-authorized shadow queue based on the pre-authorized shadow queue, combined with the successive speaking relationship and speaking switching order between the current speaking session entity and each candidate participating terminal, to form a continuous speaking candidate sequence; detecting whether the current speaking session entity has entered a switching period according to the continuous speaking candidate sequence, and selecting a target switching meeting from the continuous speaking candidate sequence when the current speaking session entity has entered a switching period, forming smooth switching control information; sending a pre-authorized status to the target switching meeting according to the smooth switching control information, sending a confirmation prompt and switching announcement to the host, and generating a switching transition state set; and detecting the voice activity of the target switching meeting and the host's release confirmation during the switching period based on the switching transition state set, and switching and recombining the speaking data corresponding to the current speaking session entity and the candidate speaking data corresponding to the target switching meeting according to the detection results to form a new meeting round state set.

[0008] As a preferred embodiment of the network meeting interaction optimization method based on real-time data analysis described in this invention, the specific steps of merging the current meeting round state set are as follows:

[0009] The current speaking session entity's speaking status and the session interaction status of each candidate participating terminal are converted in terms of status identifier and validity are determined, and session status data that meets the round merging requirement is retained.

[0010] The session state data that meet the round merging conditions are merged and reorganized according to the meeting participants, state categories, and time sequence relationships to generate the current meeting round state set.

[0011] As a preferred embodiment of the network meeting interaction optimization method based on real-time data analysis described in this invention, the specific steps for forming the pre-authorized shadow queue are as follows:

[0012] Based on the current meeting round status set, extract the speaking requests, conversation interactions, and current speaking conversation entity association data corresponding to each candidate participating terminal, and organize and align them in time sequence to obtain the candidate participating terminal ranking data;

[0013] Based on the candidate participating terminal ranking data, the validity of requests, interaction activity, link availability and connection participation of each candidate participating terminal are used to determine admission, and those that meet the pre-ranking conditions are used as the candidate participating terminal set.

[0014] Extract the meeting identifier data, candidate order data, and pending speaking data of each candidate terminal from the candidate terminal set, and then arrange and merge them in order to obtain a pre-authorized shadow queue.

[0015] As a preferred embodiment of the network meeting interaction optimization method based on real-time data analysis described in this invention, the generation of candidate order data corresponding to each candidate participating terminal includes extracting request order data, interaction activity data, link availability data, and successive speaking relationship data from the current meeting round status set and the candidate participating terminal historical interaction set, and statistically forming the corresponding median value;

[0016] The candidate order comparison value is formed by performing difference sorting and correlation reorganization on the request order data, the median value of the request order data, the interaction activity data, the median value of the interaction activity data, the link availability data, the median value of the link availability data, the successive speech relationship data, and the median value of the successive speech relationship data.

[0017] Based on the candidate order comparison value, priority candidate determination and follow-up candidate determination are performed for each candidate participating terminal, and the ranking result is used as candidate order data in the pre-authorized shadow queue.

[0018] As a preferred embodiment of the network conference interaction optimization method based on real-time data analysis described in this invention, the specific steps for forming a continuous speaking candidate sequence are as follows:

[0019] Extract the succession speech relationship data and speech switching order data between the current speaking session entity and each candidate participating terminal in the pre-authorized shadow queue, and align and organize them to obtain the candidate participating terminal associated data set;

[0020] Based on the candidate participant terminal sequence data, successive speaking relationship data, and speaking switching order data in the candidate participant terminal association data set, the successive succession order between candidate participant terminals is reorganized, and the reorganized candidate participant terminal arrangement is subjected to retention of candidate order and order extension to form candidate participant terminal inertial correction order data.

[0021] Based on the inertial correction order data of the candidate participating terminals, the order of the candidate participating terminals in the pre-authorized shadow queue is updated and rearranged to form a continuous speaking candidate sequence.

[0022] As a preferred embodiment of the network meeting interaction optimization method based on real-time data analysis described in this invention, the step of forming candidate participating terminal inertial correction order data includes statistically analyzing the candidate order data and the successive speaking relationship data between the current speaking session entity and each candidate participating terminal into corresponding order position values.

[0023] The candidate inertia comparison value is formed by continuously comparing and reorganizing the candidate order data, order position value, speech switching order data, and successive speech relationship data.

[0024] Each candidate participating terminal is determined as either a priority candidate participating terminal or a follow-up candidate participating terminal. When the candidate inertia comparison value is lower than the continuous acceptance requirement, the candidate's ranking is downgraded, and this is used as the inertia correction order data for the candidate participating terminals.

[0025] As a preferred embodiment of the network meeting interaction optimization method based on real-time data analysis described in this invention, the step of detecting whether the current speaking session entity has entered the switching period involves extracting the continuous speaking and voice activity change data of the current speaking session entity from the current meeting round state set, the continuous speaking candidate sequence, and the host's interaction record, as well as the connection waiting data of the candidate participating terminals and the switching prompt data of the host, and performing time-series correlation and state comparison to form switching period determination data.

[0026] Based on the time-switching determination data, the current speaking session entity is determined to be either in a continuous speaking state or in a time-switching state. When the time-switching determination data simultaneously meets the conditions of speaking completion, candidate waiting, and host permission, the time-switching state is used as the trigger condition for selecting the target switching meeting.

[0027] As a preferred embodiment of the network conferencing interaction optimization method based on real-time data analysis described in this invention, the specific steps for forming smooth switching control information are as follows:

[0028] Extract candidate order data, successive speech relationship data, speech switching order data, and conversation interaction state data corresponding to the consecutive speech candidate sequence, and perform correlation and organization to obtain candidate switching correlation data;

[0029] Based on the candidate switching association data, extract the speech tail data corresponding to the current speaking session entity and the control permission data corresponding to the host end, and perform time sequence comparison and switching determination. When the control permission data meets the switching requirements and the candidate switching association data meets the succession requirements, the target switching meeting data is obtained.

[0030] Extract the pre-authorization data, host confirmation prompt data, and handover announcement data corresponding to the target handover meeting from the target handover meeting data, and combine them to form smooth handover control information.

[0031] As a preferred embodiment of the network meeting interaction optimization method based on real-time data analysis described in this invention, the specific steps for generating the set of transition states are as follows:

[0032] Extract the target switching meeting and host terminal identifiers of the smooth switching control information, confirm the prompts and switching announcements, and split and reassemble them to obtain the data of target pre-authorization, host confirmation prompts and switching announcements;

[0033] Based on the data from target pre-authorization, host confirmation prompts, and handover announcements, targeted transmissions and timing alignments are performed on the target handover meeting, the host end, and the participating meetings to obtain target pre-authorization feedback data, host confirmation feedback data, and announcement reception data.

[0034] The target pre-authorization feedback data, host confirmation feedback data, and forecast reception data are correlated and merged, and a handover transition state set is generated when the target pre-authorization feedback data and host confirmation feedback data are consistent.

[0035] As a preferred embodiment of the network meeting interaction optimization method based on real-time data analysis described in this invention, the specific steps for forming a new set of meeting round states are as follows:

[0036] Based on the set of transition states, the voice input data of the target switching meeting during the switching period and the confirmation feedback data of the host during the switching period are obtained, and time sequence alignment and validity filtering are performed to obtain candidate speech continuation data and candidate release confirmation data.

[0037] Extract voice activity information from candidate speech continuation data and confirmation status information from candidate release confirmation data, and perform correlation comparison and consistency verification to generate switchover judgment data;

[0038] When the switching judgment data meets the conditions for establishing a continuation of speech and the conditions for the host to allow the session to proceed, the speech occupancy data corresponding to the current speaking session entity and the candidate speech data corresponding to the target switching meeting are switched and recombined to generate a new meeting round state set; otherwise, the speech occupancy state corresponding to the current speaking session entity is maintained.

[0039] The beneficial effects of this invention are as follows: by linking the pre-authorized shadow queue with the continuous speaking candidate sequence, and combining the real-time data acquisition and processing capabilities of edge nodes, the successor selection and order stabilization processing can be completed before the current speaking ends, so that the handover is based on the continuous call rounds, improving the handover matching degree, timing coherence and control convergence efficiency, and achieving low latency response and cloud load sharing. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart of a method for optimizing web conferencing interaction based on real-time data analysis.

[0042] Figure 2 This is a flowchart illustrating the formation of the current meeting round state set and the pre-authorized shadow queue.

[0043] Figure 3 A flowchart for forming a continuous speaking candidate sequence and smooth switching control information.

[0044] Figure 4 A flowchart for updating the transition state set and the new meeting round state set. Detailed Implementation

[0045] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0046] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0048] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for optimizing network meeting interaction based on real-time data analysis, including the following steps:

[0049] S1: Collect the speaking status of the current speaking session entity and the session interaction status of each candidate participating terminal, and merge them into the current meeting round status set.

[0050] It should be noted that the current speaking session entity refers to the terminal or session object that currently holds the right to speak, the candidate participating terminal refers to the terminal or session object that will obtain the right to speak in the next round, and the host terminal refers to the terminal or session object that has the responsibility of confirming the right to speak.

[0051] The speaking status and session interaction status of the current speaking entities and candidate participating terminals can be collected and preliminarily processed in real time at edge nodes. Edge nodes are responsible for analyzing audio activity in the audio stream, chronologically organizing session interaction statuses, and converting status identifiers, and then synchronizing the processed data to the conference server in real time. By performing data preprocessing at the edge, the load on the cloud is reduced, data transmission latency is lowered, and the current meeting round status set can promptly reflect the real-time status of the speaking entities, providing a data foundation for candidate ranking and pre-authorization shadow queue generation.

[0052] S1.1: Perform state identifier conversion and validity determination on the speaking status of the current speaking session entity and the session interaction status of each candidate participating terminal, and retain the session status data that meets the round merging requirement.

[0053] Furthermore, the speaking status of the current speaking session entity is first divided and its status identifier is transformed according to speaking occupation, continuous speaking, and speaking termination. The input data includes audio stream, and the speaking occupation process, continuous speaking process, and speaking termination process of the current speaking session entity in the current round are formed into distinguishable speaking status data.

[0054] The session interaction status of each candidate participating terminal is then divided and its status identifier is converted according to speaking request, voice activity, link availability, connection waiting and host permission. The input data sources include host control log, link quality indicators and session control signals. The speaking request process, voice activity process, link availability process, connection waiting process and host permission process of each candidate participating terminal in the current round are formed into distinguishable session interaction status data.

[0055] Based on the correspondence and timing relationship between the current speaking session entity and each candidate participating terminal, the speaking status data of the current speaking session entity and the session interaction status data of each candidate participating terminal are interactively associated and matched in rounds. The speaking occupation process of the current speaking session entity is associated with the speaking request process of the candidate participating terminal, the continuous speaking process of the current speaking session entity is associated with the succession waiting process of the candidate participating terminal, and the speaking closing process of the current speaking session entity is associated with the voice activity process and host permission process of the candidate participating terminal to form a switching preparation correspondence.

[0056] The validity of the associated speaking status data and session interaction status data is determined. First, the validity of the speaking request is determined based on whether it is in the current round and whether it forms a continuity relationship with the current speaking session entity. Then, the availability of the link is determined based on whether the link availability data indicates that the candidate participating terminal has continuous transmission capability. Finally, the continuity participation is determined based on whether the continuity waiting status maintains the temporal continuity with the continuous speaking process of the current speaking session entity, whether the voice activity status maintains the activity consistency with the speaking request status, and whether the host permission status maintains the switching correspondence with the speaking closing process.

[0057] When the speaking request is valid, the link availability is valid, and the continuity of participation is valid, the session interaction state data of the corresponding candidate participating terminal is retained, and the speaking state data of the corresponding current speaking session entity and the session interaction state data of the candidate participating terminal are determined as the session state data that satisfies the round merging. When the speaking request is invalid, the link availability is invalid, or the continuity of participation is invalid, the session interaction state data of the corresponding candidate participating terminal is removed, and the merging participation of the corresponding candidate participating terminal in the current round is terminated. By processing the session state data entering the state set of the current meeting round, the speaking acceptance validity, link transmission availability, and interaction participation consistency are simultaneously achieved.

[0058] S1.2: Merge and reorganize the session state data that meet the round merging conditions according to the meeting object, state category and time sequence relationship to generate the current meeting round state set.

[0059] Furthermore, the session state data that meets the round merging conditions are divided into objects according to the current speaking session entity and each candidate participating terminal. The speaking state and session interaction state are classified according to a unified state category, including speaking request, voice activity, link availability, connection waiting, and host permission. The original link feedback corresponds to link availability, the host control corresponds to host permission, and the shared connection corresponds to connection waiting. The state data after the state identifier is converted are matched and merged according to the temporal relationship.

[0060] The associated and merged session state data is reorganized, and the speaking occupation, continuous speaking and speaking end status of the current speaking session entity are correlated with the speaking requests, voice activities, link availability, connection waiting and host permission of each candidate participating terminal, and the current meeting round state set is generated in the same round.

[0061] S2: Based on the current meeting round status set, pre-sort each candidate participating terminal to form a pre-authorized shadow queue.

[0062] S2.1: Based on the current meeting round status set, extract the speaking requests, conversation interactions, and current speaking conversation entity association data corresponding to each candidate participating terminal, and organize and align them in time sequence to obtain the candidate participating terminal ranking data.

[0063] Furthermore, based on the meeting objects and status categories in the current meeting round status set, extract the speaking requests, conversation interactions, and current speaking conversation entity association data corresponding to each candidate participating terminal. Then, organize the speaking requests, conversation interactions, and current speaking conversation entity association data according to the temporal association relationship, and merge the speaking requests, conversation interactions, and current speaking conversation entity association data corresponding to the same candidate participating terminal into the same sorting order.

[0064] Then, according to the chronological order in the current meeting round status set, the merged speaking requests, conversation interactions, and current speaking conversation entity association data are aligned in time sequence. The speaking requests, conversation interactions, and current speaking conversation entity association data corresponding to each candidate participating terminal are arranged in a comparable order at the same time position to obtain the candidate participating terminal ranking data.

[0065] S2.2: Based on the candidate participating terminal ranking data, the validity of requests, interaction activity, link availability and connection participation of each candidate participating terminal are judged for admission, and those that meet the pre-ranking conditions are taken as the candidate participating terminal set.

[0066] Furthermore, the request sequence data, interaction activity data, link availability data, and follow-up participation data corresponding to each candidate participating terminal are extracted from the candidate participating terminal ranking data. Then, according to the candidate participating terminal identifier and the correspondence with the current round, the request sequence data, interaction activity data, link availability data, and follow-up participation data are matched and sorted in time. The request sequence data, interaction activity data, link availability data, and follow-up participation data corresponding to the same candidate participating terminal are combined to form associative candidate participating terminal determination data.

[0067] First, the request validity is determined based on the order of requests. Requests that are in the current round, maintain a connection with the current speaking session entity, and are not expired are retained as valid request data. Next, the interaction activity is determined based on the interaction activity data. Interaction activity data that indicates the candidate participating terminal's continuous participation in the current round's interaction process is retained as valid interaction data. Link availability is determined based on the link availability data. Link availability data that indicates the candidate participating terminal's stable connection transmission capability is retained as valid link data. Finally, the continuity participation data is determined based on the continuity participation data. Continuity participation data that indicates the formation of a continuity relationship between the candidate participating terminal and the current speaking session entity is retained as valid continuity data.

[0068] The system compares and correlates valid request data with valid interaction data, retaining candidate terminals that simultaneously meet the criteria of valid speaking requests and active interaction. Then, it compares and correlates valid link data with valid connection data, retaining candidate terminals that simultaneously meet the criteria of link availability and connection participation. The results of the above two types of retention are cross-merged, and only candidate terminals that simultaneously meet the criteria of valid speaking requests, active interaction, link availability, and connection participation are retained as the set of candidate terminals that meet the pre-sorting conditions. Candidate terminals that do not simultaneously meet the above conditions are removed from the pre-sorting process.

[0069] The formula for determining the candidate order is:

[0070] ;

[0071] in, This indicates the request order of the data; a smaller value indicates an earlier request and higher priority. This represents the median value of the data requested in the order they were received. This indicates the activity order value corresponding to the interactive activity data. The smaller the order, the higher the priority of continuing the conversation with the currently speaking entity. This represents the median value corresponding to the interactive activity data. This value represents the link order corresponding to the available data on the link. The larger the value, the higher the interaction activity and the higher the priority. This represents the median value corresponding to the available data on the link. This value represents the sequence order of consecutive messages; a larger value indicates higher link availability and higher priority. This represents the median value corresponding to the consecutive speech relationship data. Indicates the candidate order determination value. This indicates a state where the criteria for priority candidate determination are not met.

[0072] It should be noted that when When the value is 1, the corresponding candidate terminal is determined as a priority candidate terminal. When the value is 0, the corresponding candidate terminal is determined as a follow-up candidate terminal, based on the candidate order determination value. The determination results will further distinguish the candidate participating terminals that meet the pre-sorting conditions into priority candidate participating terminals and follow-up candidate participating terminals, and use them as the basis for forming the candidate order in the pre-authorized shadow queue.

[0073] The priority order of participating terminals is determined based on the status data of the current round. The determination conditions include that the speaking request is valid, the interaction is active, the link is available, and the continuity of participation with the current speaking session entity all meet the candidate meeting ranking data processing. Candidate participating terminals that meet all four conditions are determined as priority candidates, and those that do not meet all four conditions are determined as deferred candidates.

[0074] S2.3: Extract the meeting identifier data, candidate order data and pending speaking data of each candidate terminal in the candidate terminal set, and sort and merge them in order to obtain the pre-authorized shadow queue.

[0075] Furthermore, for each candidate participant terminal in the candidate participant terminal set, the meeting identifier data, candidate ranking data, and unactivated speaking data are extracted. The meeting identifier data and candidate ranking data are then organized accordingly, and each candidate participant terminal is arranged sequentially according to the order represented by the candidate ranking data. During the sequential arrangement process, the meeting identifier data and unactivated speaking data in the corresponding ranking are merged and recombined according to the candidate participant terminal affiliation relationship. The meeting identifier data, candidate ranking data, and unactivated speaking data form a queue structure with consistent order, thus obtaining the pre-authorized shadow queue.

[0076] S3: Based on the pre-authorized shadow queue, and combined with the successive speaking relationship and speaking switching order between the current speaking session entity and each candidate participating terminal, the pre-authorized shadow queue is inertially corrected to form a continuous speaking candidate sequence.

[0077] S3.1: Extract the succession speech relationship data and speech switching order data between the current speaking session entity and each candidate participating terminal in the pre-authorized shadow queue, and align and organize them to obtain the candidate participating terminal associated data set.

[0078] Furthermore, the meeting identifier data corresponding to the current speaking session entity and the meeting identifier data corresponding to each candidate participating terminal are read from the pre-authorized shadow queue. According to the meeting identifier data, the corresponding successive speaking relationship data and speaking switching order data between the current speaking session entity and each candidate participating terminal are extracted from the pre-authorized shadow queue item by item. According to the correspondence between the current speaking session entity and each candidate participating terminal, the successive speaking relationship data and speaking switching order data are sorted, matched in position, and aligned in time. The successive speaking relationship data and speaking switching order data corresponding to the same candidate participating terminal are associated and corresponded. The successive speaking relationship data and speaking switching order data that have been associated and corresponded are merged and sorted to obtain the candidate participating terminal association data set.

[0079] S3.2: Based on the candidate participant terminal sequence data, successive speaking relationship data, and speaking switching order data in the candidate participant terminal association data set, the successive succession order between candidate participant terminals is reorganized, and the candidate participant terminal arrangement after reorganization is subjected to retention of candidate order and order extension to form candidate participant terminal inertial correction order data.

[0080] Furthermore, the order of candidate participating terminals is first compared according to the order of the speaking switch data to obtain the order basis of the candidate participating terminals in the current round. Then, the successive speaking relationship data is mapped to the order basis. Whether the current speaking session entity and each candidate participating terminal form a question-and-answer succession, supplementary succession, roll call response succession, or shared succession succession is determined by obtaining the text content through speech recognition of the audio stream and combining it with chat history and control logs.

[0081] The succession order of candidate participating terminals is reorganized. Candidate participating terminals whose succession is consistent with the succession order data are identified as consecutively succeeding candidate participating terminals, and their candidate ranking in the candidate participating terminal sequence data is retained. Candidate participating terminals whose succession is inconsistent with the succession order data, whose succession direction is interrupted, or whose succession position is misaligned are identified as sequentially succeeding candidate participating terminals, and their ranking is shifted to the next position in the reorganized arrangement.

[0082] The successive acceptance of candidate participating terminals and the successive acceptance of candidate participating terminals are continuously checked. When the successive acceptance order between consecutively accepting candidate participating terminals is consistent with the speaking switching order data, the corresponding arrangement is retained as a valid acceptance arrangement. When there are still gaps, misalignments, or cross-acceptances between consecutively accepting candidate participating terminals, the order of the corresponding candidate participating terminals is adjusted until the successive acceptance order between candidate participating terminals and the speaking switching order data form a continuous correspondence. The valid acceptance arrangement after the successive acceptance is completed is written into the candidate participating terminal order data to form the candidate participating terminal inertial correction order data.

[0083] It should be noted that the successive acceptance of candidate participating terminals and the sequential acceptance of candidate participating terminals are checked for continuity. When the successive acceptance order between consecutive candidate participating terminals is consistent with the speaking switch order data, the corresponding arrangement is retained as a valid acceptance arrangement. If there are still gaps in acceptance, misaligned order, or overlapping acceptance, the order of the corresponding candidate participating terminals is adjusted until any of the following conditions are met:

[0084] The sequential order of candidate participating terminals and the speaking switching order data form a continuous correspondence;

[0085] To reach the maximum number of iterations and prevent infinite loops;

[0086] Rearranging the order cannot further improve the continuity of the content.

[0087] For situations where breaks, misalignments, or overlapping connections cannot be completely eliminated, candidate participating terminals are marked as successor candidates, and their adjusted relative order is retained to ensure that the overall arrangement is as close as possible to a continuous succession sequence. The valid succession arrangement after continuity verification is written into the candidate participating terminal sequence data, forming the candidate participating terminal inertial correction sequence data.

[0088] Maximum number of iterations: Prevents the sorting process from looping infinitely and ensures executability.

[0089] Convergence condition: The preceding and following sequences correspond continuously or cannot be further improved.

[0090] Handling exceptions that cannot be eliminated: Continue the succession marker and retain the relative order to ensure logical continuity.

[0091] Retain the original logic: prioritize consecutive data, and move subsequent data to the next order to form an inertial correction sequence.

[0092] In a better way, through reorganization, the arrangement of candidate participating terminals is no longer determined solely by the order of candidates, but is simultaneously constrained by the successive speaking relationship and the speaking switching order. Candidate participating terminals that continue to speak are given priority to retain their previous positions, while candidate participating terminals whose speaking is interrupted or misaligned are moved to the back, thereby improving the matching degree of the continuous speaking candidate sequence with the actual turn-taking process.

[0093] S3.3: Based on the candidate participating terminal inertial correction order data, update and rearrange the order of candidate participating terminals in the pre-authorized shadow queue to form a continuous speaking candidate sequence.

[0094] Furthermore, based on the candidate terminal inertial correction order data, the candidate terminal identification data, candidate priority data, and pending speaking data in the pre-authorized shadow queue are correspondingly organized with the candidate terminal inertial correction order data. Then, according to the priority and subsequent priority order of the candidate terminals represented by the candidate terminal inertial correction order data, the candidate terminal order in the pre-authorized shadow queue is updated and rearranged. The candidate priority data corresponding to the priority candidate terminals is moved forward, and the candidate priority data corresponding to the subsequent candidate terminals is moved backward. Finally, the candidate terminal identification data, candidate priority data, and pending speaking data after the update and rearrangement are output in the updated order to form a continuous speaking candidate sequence.

[0095] S4: Based on the continuous speaking candidate sequence, detect whether the current speaking session entity has entered the switching period, and when the current speaking session entity is detected to have entered the switching period, select the target switching session from the continuous speaking candidate sequence to form smooth switching control information.

[0096] S4.1: Detect whether the currently speaking session entity has entered the switching period.

[0097] Furthermore, the detection of whether the current speaking session entity has entered the switching period is jointly completed based on the current meeting round state set, the continuous speaking candidate sequence, and the host's interaction record. Continuous speaking data and voice activity change data corresponding to the current speaking session entity are extracted from the current meeting round state set; connection waiting data corresponding to candidate participating terminals are extracted from the continuous speaking candidate sequence; and switching prompt data corresponding to the host is extracted from the host's interaction record. According to the correspondence between the current speaking session entity, candidate participating terminals, and the host, as well as the sequential relationship within the current round, the continuous speaking data, voice activity change data, connection waiting data, and switching prompt data are time-aligned and correlated, forming a correspondence between the continuous speaking process of the current speaking session entity, the connection waiting process of candidate participating terminals, and the switching prompt process of the host within the same round.

[0098] The system compares the status of continuous speaking data, voice activity change data, connection waiting data, and switching prompt data. If the continuous speaking data indicates that the current speaking session entity has transitioned from continuous speaking to the end of speaking, the voice activity change data indicates that the effective speaking intensity of the current speaking session entity has decreased, the connection waiting data indicates that the candidate participating terminal has entered the connection waiting state, and the switching prompt data indicates that the host has issued a switching permission prompt, then the current speaking session entity is determined to have entered the switching period, and the corresponding switching period determination data is output for use in the subsequent target switching meeting selection and smooth switching control information formation process. If the continuous speaking data still indicates that the current speaking session entity is in a continuous speaking state, the voice activity change data does not indicate the end of speaking, the connection waiting data does not indicate that the candidate participating terminal has entered the connection waiting state, and the switching prompt data does not indicate that the host has issued a switching permission prompt, then the current speaking session entity is determined not to have entered the switching period, and the determination result of not entering the switching period is retained in the current meeting round status set.

[0099] The formula for determining the time period switch is:

[0100] ;

[0101] in, This indicates the status value of the current speaking entity in the previous round of the session. This indicates the closing status value of the current speaking session entity in the current round. This indicates the status value of voice activity changes in the current round. This indicates the connection waiting status value of the candidate participating terminals in the current round. This indicates the status value of the host switching prompt in the current round. Indicates the current round number. This indicates the value used to determine the time period switching. This indicates a state where the conditions for determining the time period are not met.

[0102] when When the value is 1, it indicates that the current speaking session entity has entered a switching period, and the corresponding switching period determination data is output for use in the target switching meeting selection and smooth switching control information formation process. When the value is 0, it is determined that the current speaking session entity has not entered the switching period, and the determination result of not entering the switching period is retained in the current meeting round state set.

[0103] It should be noted that the continuous speaking status comes from the speaking occupation and speaking end status data of the current speaking session entity. If it indicates a transition from continuous speaking to speaking end, then the switching preparation conditions are met.

[0104] Changes in speech activity are recorded from the audio stream. If the intensity of effective speech decreases, it indicates that the current speech is entering its final stage.

[0105] The connection waiting status comes from the connection waiting status data of the candidate participating terminal. If it shows that the candidate participating terminal is ready to continue speaking, then the connection condition is met.

[0106] The switching prompt status comes from the switching permission instruction in the host's control log. If the host has issued a switching permission prompt, then the host permission conditions are met.

[0107] When all four states indicate that the switching conditions are met, the current speaking session entity is determined to have entered the switching period, and the corresponding switching period determination data is output for use in the target switching conference selection and smooth switching control information formation process.

[0108] If any condition is not met, such as the speaker still being in a continuous speaking state, the voice activity not indicating the end of the speech, the candidate participating terminal not entering the continuation waiting state, or the host not issuing a prompt to allow switching, then it is determined that the current speaking session entity has not entered the switching period, and the determination result is retained in the current meeting round state set.

[0109] S4.2: Extract the candidate order data, successive speech relationship data, speech switching order data, and session interaction state data corresponding to the continuous speech candidate sequence, and perform correlation and organization to obtain candidate switching correlation data.

[0110] Furthermore, firstly, object matching is performed on the candidate ranking data, successive speaking relationship data, speaking switching order data, and session interaction status data according to the meeting identifier. This establishes an object correspondence between the candidate ranking data, successive speaking relationship data, speaking switching order data, and session interaction status data corresponding to the same candidate participating terminal. Then, according to the arrangement position and sequential order in the continuous speaking candidate sequence, the data that has completed object matching is aligned in order and time. The ranking position, successive succession position, switching order position, and interaction status position of the same candidate participating terminal in the current switching round are established in time. According to the object correspondence and time correspondence corresponding to each candidate participating terminal, the candidate ranking data, successive speaking relationship data, speaking switching order data, and session interaction status data are merged into corresponding data combinations to obtain candidate switching associated data.

[0111] S4.3: Based on the candidate switching association data, extract the speech tail data corresponding to the current speaking session entity and the control permission data corresponding to the host end, and perform time sequence comparison and switching determination. When the control permission data meets the switching requirements and the candidate switching association data meets the succession requirements, obtain the target switching meeting data.

[0112] Furthermore, candidate ranking data, succession speaking relationship data, speaking switching order data, and session interaction status data are extracted from the candidate switching association data. Then, the speaking end data is extracted from the current speaking session entity, and control permission data is extracted from the host end. According to the correspondence between the meeting objects of the current speaking session entity, candidate participating terminals, and the host end, as well as the sequential relationship within the current round, the candidate ranking data, succession speaking relationship data, speaking switching order data, session interaction status data, speaking end data, and control permission data are time-aligned and associated. The candidate ranking, succession status, switching order position, session interaction status, speaking end process of the current speaking session entity, and permission process of the host end are formed into a correspondence under the same switching round.

[0113] The system performs switching determination based on candidate switching association data, speech end data, and control permission data. First, it determines whether the current speaking session entity is in the switching preparation stage from continuous speaking to speech closing based on the speech end data. Then, it determines whether the host has issued a switching permission result for the current round based on the control permission data. Based on candidate order data, successive speaking relationship data, speaking switching order data, and session interaction status data, it determines whether the candidate participating terminal has formed a continuous succession relationship with the current speaking session entity, whether it maintains the order of succession in the current round, and whether it has the status of succession participation.

[0114] When the control permission data indicates that the host has issued a permission result to switch, and the candidate switch association data indicates that the candidate participating terminal and the current speaking session entity maintain continuous connection, consistent sequence connection and have the ability to continue participating, the corresponding candidate participating terminal is retained and will be determined as the target switch meeting data.

[0115] If any condition is not met, such as the control permission data not indicating that the switch is allowed, the candidate switch association data not indicating continuous acceptance, or the candidate switch association data not indicating the continuous participation status, the corresponding candidate participating terminal will be immediately screened out and will not be included in the target switch meeting data.

[0116] Ensure that the target switching meeting data only includes candidate participating terminals that meet all switching requirements and have the capacity to take over.

[0117] Ideally, by comparing time sequences and processing the switching decision, the target switching meeting data can simultaneously meet the conditions of the current speaking session entity ending its speech, the host granting permission, and the candidate participating terminal taking over the connection. This ensures that the pre-authorization data, host confirmation prompt data, and switching announcement data are all based on an executable switching object.

[0118] S4.4: Extract the pre-authorization data, host confirmation prompt data, and handover announcement data corresponding to the target handover meeting from the target handover meeting data, and combine them to form smooth handover control information.

[0119] Furthermore, the target switching meeting data is associated with meeting objects to obtain the pre-authorization data, host confirmation prompt data, and switching announcement data corresponding to the target switching meeting. Then, according to the candidate order data, successive speaking relationship data, and speaking switching order data, the pre-authorization data, host confirmation prompt data, and switching announcement data are organized and compared in time. The pre-authorization data corresponds to the pending speaking position in the target switching meeting, the host confirmation prompt data corresponds to the confirmation interaction position in the host, and the switching announcement data corresponds to the switching prompt position in the current round, forming a data combination with corresponding content and consistent time. The data combination with corresponding content and consistent time is merged according to the target switching meeting, host, and switching round. The pre-authorization content corresponding to the target switching meeting, the confirmation prompt content corresponding to the host, and the switching announcement content corresponding to the current round are written into the same control data structure to form smooth switching control information.

[0120] S5: Based on the smooth handover control information, send a pre-authorization status to the target handover meeting, send a confirmation prompt and handover announcement to the host, and generate a handover transition status set.

[0121] S5.1: Extract the target switching conference and host terminal identifiers of the smooth switching control information, confirm the prompts and switching announcements, and split and reassemble them to obtain the data of target pre-authorization, host confirmation prompts and switching announcements.

[0122] Furthermore, the target switching meeting identifier, host identifier, confirmation prompt, and switching announcement content in the smooth switching control information are associated with meeting objects and their status identifiers are converted. The authorization pointing content corresponding to the target switching meeting, the confirmation prompt content corresponding to the host, and the switching announcement content are distinguished. The confirmation prompt and switching announcement content are split according to the target switching meeting identifier and the host identifier. The split content is then time-aligned and reassembled. The authorization pointing content corresponding to the target switching meeting identifier is reassembled into target pre-authorization, the confirmation prompt content corresponding to the host identifier is reassembled into host confirmation prompt data, and the switching announcement content is reassembled into switching announcement data.

[0123] S5.2: Based on the target pre-authorization, host confirmation prompt, and handover announcement data, perform targeted sending and timing alignment for the target handover meeting, the host end, and the participating meetings respectively, and obtain target pre-authorization feedback data, host confirmation feedback data, and announcement reception data.

[0124] Furthermore, the target switching meeting identifier, host identifier, and participant meeting identifier are extracted separately. Based on the target switching meeting identifier, host identifier, and participant meeting identifier, the data of target pre-authorization, host confirmation prompt, and switching announcement are matched with the meeting objects and the content is split. The target pre-authorization is associated with the target switching meeting, the host confirmation prompt is associated with the host, and the switching announcement is associated with the participant meeting, forming a data group for sending to different meeting objects.

[0125] The data for target pre-authorization, host confirmation prompts, and handover announcements are organized according to the chronological order within the current handover round. Targeted transmission is then performed based on the temporal correspondence between the target handover meeting, the host, and the participating meetings. The target handover meeting receives the target pre-authorization, the host receives the host confirmation prompt, and the participating meetings receive the handover announcement. During the targeted transmission process, the data for target pre-authorization, host confirmation prompts, and handover announcements are time-aligned to maintain the correspondence between the authorization reception process of the target handover meeting, the confirmation reception process of the host, and the announcement reception process of the participating meetings within the same handover round.

[0126] Extract, match, and merge the authorized reception content returned by the target switching meeting, the confirmation feedback content returned by the host, and the preview reception content returned by the participating meetings. The authorized reception content corresponding to the target switching meeting identifier is retained as target pre-authorization feedback data, the confirmation feedback content corresponding to the host identifier is retained as host confirmation feedback data, and the preview reception content corresponding to the participating meeting identifier is retained as preview reception data. Then, output the target pre-authorization feedback data, host confirmation feedback data, and preview reception data.

[0127] S5.3: Associate and merge the target pre-authorization feedback data, host confirmation feedback data, and pre-announcement reception data, and generate a handover transition state set when the target pre-authorization feedback data and host confirmation feedback data are consistent.

[0128] Furthermore, based on the target switching meeting identifier and the host identifier, the target pre-authorization feedback data, host confirmation feedback data, and pre-receipt reception data are matched with the meeting objects to ensure that the target switching meeting and the corresponding feedback data of the host are in the same switching round. The matched target pre-authorization feedback data and host confirmation feedback data are judged for consistency based on the pre-authorization reception status and the host confirmation status. Data that simultaneously meets the conditions of pre-authorization being received and host confirmation being valid is marked as consistent.

[0129] The data that has achieved consistency is aligned and merged with the corresponding pre-received data according to the target switching meeting identifier and switching round, forming a data combination with corresponding content and consistent round. The merged and time-aligned data combination is written into the switching transition state set, which accurately represents the speech preparation and interaction status of the target switching meeting in the current switching round.

[0130] By associating and merging the target pre-authorization feedback data, host confirmation feedback data, and pre-announcement reception data, a set of handover transition states is generated.

[0131] The determination criteria are: the target pre-authorization feedback data has been received and the host confirms the feedback is valid. The data sources correspond to the directed data and the received feedback data, respectively. Only data that meets the consistency condition will be written into the switching transition state set, representing the speaking preparation and interaction status of the target switching meeting in the current round. If the pre-authorization feedback and the host confirms the feedback are inconsistent, a delayed retry will be performed by requesting feedback again within the time window. The host confirmation or pre-authorization will be selected as the determination criterion based on priority. If the retry and priority determination still cannot reach an agreement, the round switching will be marked as abnormal, and the speaking status of the current speaking session entity will remain unchanged, waiting for the next round to switch again.

[0132] S6: Based on the set of transitional states, detect the voice activity of the target switching conference and the release confirmation of the host during the switching period, and switch and reorganize the speech occupancy data corresponding to the current speaking session entity and the candidate speech data corresponding to the target switching conference according to the detection results to form a new set of conference round states.

[0133] S6.1: Based on the set of transition states, obtain the voice input data of the target switching meeting during the switching period and the confirmation feedback data of the host during the switching period, and perform time sequence alignment and validity filtering to obtain candidate speech continuation data and candidate release confirmation data.

[0134] Furthermore, the meeting objects in the transition state set are matched and distinguished according to the target switching meeting identifier and the host identifier. Based on the start and end positions corresponding to the switching time period, the voice input data corresponding to the target switching meeting and the confirmation feedback data corresponding to the host are segmented into time periods, thus limiting the voice input process of the target switching meeting and the confirmation feedback process of the host to the same switching time period.

[0135] The voice input data and confirmation feedback data are time-aligned. The start position, duration position and change position of voice activity in the voice input data are matched with the start position, hold position and change position of confirmation in the confirmation feedback data. This establishes a correspondence between the voice activity process of the target switching meeting and the confirmation feedback process of the host end within the same switching round.

[0136] The validity of voice input data and confirmation feedback data is screened. Meeting identifier consistency means that the target switching session entity identifier in the voice input data corresponds to the target switching session entity identifier in the switching transition state set, and the host identifier in the confirmation feedback data corresponds to the host identifier in the switching transition state set. Switching time period consistency means that the generation time of both voice input data and confirmation feedback data falls within the start and end time range of the current switching time period. Sequential correspondence means that the generation time of the host's release confirmation data is not later than the effective access time of the target switching session entity's continued voice activity, or they form an associative sequential correspondence within the same switching time period. Continuous and valid confirmation status means that the host's release confirmation status in the current switching time period has not been revoked or overwritten by a new rejection confirmation, and it maintains a correspondence with the current target switching session entity.

[0137] Voice input data with consistent meeting identifiers, consistent switching time periods, and corresponding time sequences are retained to form candidate succession speech data. Confirmation feedback data with consistent meeting identifiers, consistent switching time periods, and continuous and valid confirmation status are retained to form candidate release confirmation data.

[0138] For cases where there are inconsistencies in meeting identifiers, inconsistencies in switching time periods, timing misalignments between voice input data and confirmation feedback data, or situations where the host's confirmation status is revoked, overwritten by rejected confirmation, or does not correspond to the target switching session entity, these are identified as confirmation status interruptions or invalid feedback. The corresponding voice input data and confirmation feedback data are then removed so that the retained candidate speech continuation data and candidate release confirmation data can jointly represent the arrival status of the target switching session entity's speech continuation within the current switching time period and the host's release confirmation status within the current switching time period.

[0139] Ideally, by using time-series alignment and validity filtering, candidate succession speech data and candidate release confirmation data are combined into a corresponding and verifiable state combination within the same switching time period, providing a consistent input basis for the generation of switching judgment data.

[0140] During the handover determination process, the voice input data of the target handover meeting and the confirmation feedback data of the host can be initially aligned in time and screened for validity at the edge node.

[0141] Edge nodes first verify the sequential order of voice activities and confirmation status based on rounds, switching time periods, and correspondence with meeting participants, and mark invalid or interrupted data. Then, the processing results are sent to the meeting server for switching determination. Through edge-side processing, low-latency real-time feedback and switching response are achieved, ensuring that the continuous speaking candidate sequence and the set of switching transition states are highly consistent with the actual conversation in terms of timing. At the same time, it reduces the processing pressure on the cloud and improves the overall response efficiency and controllability.

[0142] S6.2: Extract the voice activity information from the candidate speech continuation data and the confirmation status information from the candidate release confirmation data, and perform correlation comparison and consistency verification to generate handover judgment data.

[0143] Furthermore, based on the correspondence between candidate succession speech data and candidate release confirmation data formed within the same switching time period, voice activity information in candidate succession speech data and confirmation status information in candidate release confirmation data are extracted respectively. Then, according to the target switching meeting identifier, host identifier, and the correspondence of switching time period, object matching and temporal correspondence are performed on the voice activity information and confirmation status information, forming a related data group under the same switching round between the voice activity process of the target switching meeting and the confirmation feedback process of the host.

[0144] The activity start, activity duration, and activity change in the voice activity information are correlated and compared with the confirmation time, confirmation hold, and confirmation change in the confirmation status information. The process of the successive speech arrival represented by the voice activity information is matched sequentially and in terms of status with the process of the host release represented by the confirmation status information. It is then determined whether the voice activity information and the confirmation status information maintain a continuous correspondence within the same switching time period.

[0145] The consistency of the corresponding results after the correlation comparison is checked. When the voice activity information indicates that the target switching meeting has formed a succession speech and the confirmation status information indicates that the host has formed a release confirmation, and there is no time sequence misalignment, status conflict or confirmation missing between the voice activity information and the confirmation status information, the corresponding correlation result is retained. When there is a time sequence misalignment, status conflict or confirmation missing between the voice activity information and the confirmation status information, the corresponding correlation result is removed. Switching judgment data is generated based on the retained correlation results. The switching judgment data jointly represent the succession speech status of the target switching meeting in the current switching period and the release status of the host in the current switching period, providing a judgment basis for switching and reorganization.

[0146] S6.3: When the switching judgment data meets the conditions for the establishment of the continuation of speech and the conditions for the host to allow the session, the speech occupancy data corresponding to the current speaking session entity and the candidate speech data corresponding to the target switching meeting are switched and recombined to generate a new meeting round state set; otherwise, the speech occupancy state corresponding to the current speaking session entity is maintained.

[0147] Furthermore, based on the handover determination data, the voice activity information in the candidate succession speech data and the confirmation status information in the candidate release confirmation data are judged to be valid. According to the correspondence between the target handover conference identifier, the current speech session entity identifier and the current handover time period, the handover correspondence between the speech occupancy data corresponding to the current speech session entity and the candidate speech data corresponding to the target handover conference is determined.

[0148] When the switching judgment data meets the conditions for establishing a continuation of speech and the conditions for the host to allow the session, the status identifier of the speech occupancy data corresponding to the current speaking session entity is changed, and the current speaking session entity is changed from the speaking occupancy state to the speaking exit state. At the same time, the status identifier of the candidate speaking data corresponding to the target switching meeting is changed, and the target switching meeting is changed from the candidate speaking state to the effective speaking state. According to the meeting object, status category and time sequence relationship, the speech occupancy data and the candidate speaking data after the status identifier change are switched and reorganized. The exit process of the current speaking session entity and the effective process of the target switching meeting are connected in the same switching round, and a new meeting round status set is generated.

[0149] If the switching judgment data does not meet the conditions for establishing a continuation of speech or the conditions for the host to allow the speech to proceed, the switching and reorganization of the speech occupancy data corresponding to the current speaking session entity and the candidate speech data corresponding to the target switching meeting will not be performed. Instead, the speech occupancy data corresponding to the current speaking session entity will be kept in the occupancy state in the current meeting round state set, and the speech switching process under the current switching round will be terminated.

[0150] Ideally, by switching and reorganizing, a new set of meeting round states can be generated only when both the succession of a speech and the moderator's permission are met. This ensures that the speech handover between the current speaking session entity and the target switching meeting maintains state continuity and temporal consistency, and prevents candidate speech data that does not meet the switching conditions from entering the new set of meeting round states.

[0151] In summary, this invention, through the coordinated construction of a pre-authorized shadow queue and a continuous speaking candidate sequence, combined with the real-time data acquisition and processing capabilities of edge nodes, can complete the selection of succession objects and the stabilization of their order before the current speaking session ends. This allows the handover to be based on continuous call rounds, improving the handover matching degree, timing coherence, and control convergence efficiency, while simultaneously achieving low-latency response and cloud load sharing.

[0152] 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, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for optimizing online meeting interaction based on real-time data analysis, characterized in that, include: Collect the speaking status of the current speaking session entity and the session interaction status of each candidate participating terminal, and merge them into the current meeting round status set; Based on the current meeting round status set, each candidate participating terminal is pre-sorted to form a pre-authorized shadow queue; Based on the pre-authorized shadow queue, and combined with the successive speaking relationship and speaking switching order between the current speaking session entity and each candidate participating terminal, the pre-authorized shadow queue is inertially corrected to form a continuous speaking candidate sequence; Based on the continuous speaking candidate sequence, it is detected whether the current speaking session entity has entered the switching period, and when the current speaking session entity is detected to have entered the switching period, the target switching session is selected from the continuous speaking candidate sequence to form smooth switching control information; Based on the smooth handover control information, a pre-authorization status is sent to the target handover meeting, a confirmation prompt and handover announcement are sent to the host, and a handover transition status set is generated; Based on the set of transitional states, the voice activity of the target switching conference and the release confirmation of the host are detected during the switching period. Based on the detection results, the speech occupancy data corresponding to the current speaking session entity and the candidate speech data corresponding to the target switching conference are switched and recombined to form a new set of conference round states.

2. The network meeting interaction optimization method based on real-time data analysis as described in claim 1, characterized in that, The process of merging the current meeting round status set involves the following steps: The current speaking session entity's speaking status and the session interaction status of each candidate participating terminal are converted in terms of status identifier and validity are determined, and session status data that meets the round merging requirement is retained. The session state data that meet the round merging conditions are merged and reorganized according to the meeting participants, state categories, and time sequence relationships to generate the current meeting round state set.

3. The network meeting interaction optimization method based on real-time data analysis as described in claim 1, characterized in that, The specific steps for forming the pre-authorized shadow queue are as follows: Based on the current meeting round status set, extract the speaking requests, conversation interactions, and current speaking conversation entity association data corresponding to each candidate participating terminal, and organize and align them in time sequence to obtain the candidate participating terminal ranking data; Based on the candidate participating terminal ranking data, the validity of requests, interaction activity, link availability and connection participation of each candidate participating terminal are used to determine admission, and those that meet the pre-ranking conditions are used as the candidate participating terminal set. Extract the meeting identifier data, candidate order data, and pending speaking data of each candidate terminal from the candidate terminal set, and then arrange and merge them in order to obtain a pre-authorized shadow queue.

4. The network meeting interaction optimization method based on real-time data analysis as described in claim 3, characterized in that, The process of generating candidate order data for each candidate participating terminal includes extracting request sequence data, interaction activity data, link availability data, and successive speaking relationship data from the current meeting round status set and the candidate participating terminal historical interaction set, and statistically forming the corresponding median value. The candidate order comparison value is formed by performing difference sorting and correlation reorganization on the request order data, the median value of the request order data, the interaction activity data, the median value of the interaction activity data, the link availability data, the median value of the link availability data, the successive speech relationship data, and the median value of the successive speech relationship data. Based on the candidate order comparison value, priority candidate determination and follow-up candidate determination are performed for each candidate participating terminal, and the ranking result is used as candidate order data in the pre-authorized shadow queue.

5. The network meeting interaction optimization method based on real-time data analysis as described in claim 1, characterized in that, The specific steps for forming a continuous candidate speaking sequence are as follows: Extract the succession speech relationship data and speech switching order data between the current speaking session entity and each candidate participating terminal in the pre-authorized shadow queue, and align and organize them to obtain the candidate participating terminal associated data set; Based on the candidate participant terminal sequence data, successive speaking relationship data, and speaking switching order data in the candidate participant terminal association data set, the successive succession order between candidate participant terminals is reorganized, and the reorganized candidate participant terminal arrangement is subjected to retention of candidate order and order extension to form candidate participant terminal inertial correction order data. Based on the inertial correction order data of the candidate participating terminals, the order of the candidate participating terminals in the pre-authorized shadow queue is updated and rearranged to form a continuous speaking candidate sequence.

6. The network meeting interaction optimization method based on real-time data analysis as described in claim 5, characterized in that, The formation of candidate participating terminal inertial correction order data includes statistically analyzing the candidate order data and the succession speaking relationship data between the current speaking session entity and each candidate participating terminal into corresponding order position values. The candidate inertia comparison value is formed by continuously comparing and reorganizing the candidate order data, order position value, speech switching order data, and successive speech relationship data. Each candidate participating terminal is determined as either a priority candidate participating terminal or a follow-up candidate participating terminal. When the candidate inertia comparison value is lower than the continuous acceptance requirement, the candidate's ranking is downgraded, and this is used as the inertia correction order data for the candidate participating terminals.

7. The network meeting interaction optimization method based on real-time data analysis as described in claim 1, characterized in that, The process of detecting whether the current speaking session entity has entered a switching period involves extracting the continuous speaking and voice activity change data of the current speaking session entity from the current meeting round status set, the continuous speaking candidate sequence, and the host's interaction records, as well as the connection waiting data of the candidate participating terminals and the switching prompt data of the host, and performing time-series correlation and status comparison to form switching period determination data. Based on the time-switching determination data, the current speaking session entity is determined to be either in a continuous speaking state or in a time-switching state. When the time-switching determination data simultaneously meets the conditions of speaking completion, candidate waiting, and host permission, the time-switching state is used as the trigger condition for selecting the target switching meeting.

8. The network meeting interaction optimization method based on real-time data analysis as described in claim 1, characterized in that, The specific steps for generating smooth switching control information are as follows: Extract candidate order data, successive speech relationship data, speech switching order data, and conversation interaction state data corresponding to the consecutive speech candidate sequence, and perform correlation and organization to obtain candidate switching correlation data; Based on the candidate switching association data, extract the speech tail data corresponding to the current speaking session entity and the control permission data corresponding to the host end, and perform time sequence comparison and switching determination. When the control permission data meets the switching requirements and the candidate switching association data meets the succession requirements, the target switching meeting data is obtained. Extract the pre-authorization data, host confirmation prompt data, and handover announcement data corresponding to the target handover meeting from the target handover meeting data, and combine them to form smooth handover control information.

9. The network meeting interaction optimization method based on real-time data analysis as described in claim 1, characterized in that, The specific steps for generating the set of transition states are as follows: Extract the target switching meeting and host terminal identifiers of the smooth switching control information, confirm the prompts and switching announcements, and split and reassemble them to obtain the data of target pre-authorization, host confirmation prompts and switching announcements; Based on the data from target pre-authorization, host confirmation prompts, and handover announcements, targeted transmissions and timing alignments are performed on the target handover meeting, the host end, and the participating meetings to obtain target pre-authorization feedback data, host confirmation feedback data, and announcement reception data. The target pre-authorization feedback data, host confirmation feedback data, and forecast reception data are correlated and merged, and a handover transition state set is generated when the target pre-authorization feedback data and host confirmation feedback data are consistent.

10. The network meeting interaction optimization method based on real-time data analysis as described in claim 1, characterized in that, The specific steps for forming a new set of meeting round states are as follows: Based on the set of transition states, the voice input data of the target switching meeting during the switching period and the confirmation feedback data of the host during the switching period are obtained, and time sequence alignment and validity filtering are performed to obtain candidate speech continuation data and candidate release confirmation data. Extract voice activity information from candidate speech continuation data and confirmation status information from candidate release confirmation data, and perform correlation comparison and consistency verification to generate switchover judgment data; When the switching judgment data meets the conditions for establishing a continuation of speech and the conditions for the host to allow the session to proceed, the speech occupancy data corresponding to the current speaking session entity and the candidate speech data corresponding to the target switching meeting are switched and recombined to generate a new meeting round state set; otherwise, the speech occupancy state corresponding to the current speaking session entity is maintained.

Images