CN118509423B - A method for transmitting audio and video in fusion communication - Google Patents

Abstract

A method for transmitting audio and video in converged communication, the steps of which include: a server collects device information of a terminal, and selects a coding standard with the smallest bandwidth usage as an adaptation coding standard corresponding to the terminal. The server dynamically allocates the bandwidth of the terminal based on congestion control technology, and pushes the corresponding stream of the adaptation coding standard according to the request information of the terminal. The server also collects user tags and performs a weighted evaluation on the terminal; when the server detects network congestion, the media service of the terminal is suspended based on the result of the weighted evaluation; when the correlation between the content of the media service and the user tag exceeds a threshold, the server performs a secondary weighted evaluation on the terminal, and suspends or resumes the media service of the terminal. The present invention maximizes the use of network bandwidth through the server's same-source multi-stream push method, combined with congestion control technology, to provide users of each participating terminal with a high-quality audio and video communication experience, reduce hardware investment in a specific communication system, and avoid waste of resources.

Description

A method for transmitting audio and video in fusion communication

Technical Field

The present invention belongs to the field of communication technology, and in particular relates to a method for transmitting integrated communication audio and video.

Background Art

At present, converged communication technology is widely used in various fields. Its main form is video conferencing, which ensures that participants can participate in online meetings in real time through their respective terminal devices through synchronous audio and video transmission. In particular, when an emergency occurs, the converged communication command and dispatch platform can be used to achieve front-end and rear-end coordination and coordinated business deployment, and quickly start a hybrid audio and video conference to achieve sharing and communication of event information among all personnel.

Since converged communications have a strong timeliness and the network environment cannot be well guaranteed in emergency situations, transmission optimization of its specific implementation methods can play a crucial role in scheduling effects and costs.

At present, the most common audio and video transmission optimization solutions on the market are based on the optimization solutions provided by the Web Video Voice Real-Time Communication Protocol (hereinafter referred to as: WebRTC), which has the following disadvantages:

(1) Failure to effectively utilize network bandwidth and solve packet loss and delay problems in weak networks;

(2) It cannot effectively adapt to the decoding requirements of each terminal, thereby increasing infrastructure investment and wasting resources;

(3) When network congestion occurs, the network bandwidth cannot be effectively adjusted, resulting in a waste of network resources and affecting the quality of the meeting.

Therefore, it is necessary to propose new solutions for converged communication audio and video transmission technology in order to maximize the use of network bandwidth and provide high-quality audio and video communication experience.

Summary of the invention

The present invention provides a fusion communication audio and video transmission method to solve the above problems.

The present invention adopts the following technical solutions:

A method for transmitting audio and video in converged communication, the main steps of which include:

The server establishes connections with multiple terminals, and the server collects device information of the multiple terminals;

The server lists the supported coding standards corresponding to each of the multiple terminals based on the device information, and selects the coding standard with the smallest bandwidth share from the corresponding supported coding standards as the adapted coding standard corresponding to each of the multiple terminals;

The server dynamically allocates bandwidth of the multiple terminals based on congestion control technology;

The server pushes the corresponding streams of the adapted coding standard to the multiple terminals according to the request information of the multiple terminals and the bandwidth of the multiple terminals;

The server collects a user tag of each user of each terminal among the multiple terminals, and extracts feature information from the user tag; the server calculates a weight of the feature information;

The server performs a weighted evaluation on each of the multiple terminals based on the weight;

When the server detects network congestion, the server suspends a media service of at least one of the plurality of terminals based on a result of the weighted evaluation;

A dynamic weight is also provided in the weighted evaluation; when the server detects network congestion, when the correlation between the content of the media service pushed by the server and the user tag exceeds a threshold, the server performs a secondary weighted evaluation on each of the multiple terminals based on the dynamic weight, and the server suspends or resumes the media service of at least one of the multiple terminals based on the result of the secondary weighted evaluation.

Optionally, the weight is obtained based on a hierarchical analysis method, and the step of the server calculating the weight includes:

Determining a hierarchical model for the weight calculation;

Establishing an importance judgment matrix of the feature information;

The importance of the feature information is calculated by a feature vector method, and the obtained value is normalized to obtain the weight.

Optionally, the step of calculating the correlation between the content of the media service and the user tag includes:

The server extracts information from the content of the media service in real time based on the large language model, and the server calculates correlation between the information extraction result and the user tag of each user of each of the multiple terminals to obtain a calculation result;

The server compares the calculation result with the threshold value.

Optionally, the server and the multiple terminals are connected via a public network or a private network;

The data transmission between the server and the plurality of terminals is based on a web video and voice real-time communication protocol;

The congestion control technology is a congestion control algorithm based on the delay gradient and packet loss rate of the web video and voice real-time communication protocol.

Optionally, the step of the server collecting device information of the multiple terminals includes:

The webpage video and voice real-time communication protocol is configured with a session description protocol, and the text of the session description protocol includes the device information;

The server collects the text of the session description protocol, and obtains the device information by the server;

The device information includes type information, operating system information, and firmware information of the multiple terminals.

Optionally, the coding standard includes VP8, VP9, H.264, and H.265;

When at least one of the multiple terminals supports AV1 decoding, the encoding standard also includes AV1.

Optionally, the result of the weighted evaluation is a weighted value of each of the multiple terminals, and the multiple terminals are arranged in ascending order according to the weighted values;

When network congestion is detected, media services of corresponding terminals are suspended in descending order of weighted values until the network congestion ends;

When it is detected that the network congestion ends, the server reversely resumes the media service of the suspended terminal.

Optionally, the characteristic information of the user tag includes multiple titles; and the user of each of the multiple terminals selects or marks the title.

Optionally, the characteristic information of the user tag further includes a plurality of preset options, and at least one of the preset options is correspondingly provided in at least part of the titles;

The preset options are each provided with a score, and the score is used for the weighted evaluation.

Optionally, each of the titles is configured with an importance coefficient, and the coefficient is used to calculate the weight.

Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

The present invention uses a server same-source multi-stream push method combined with congestion control technology to maximize network bandwidth, provide high-quality audio and video communication experience for users of each participating terminal, reduce hardware investment in a specific communication system, and avoid resource waste.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

Fig. 1 is a schematic diagram of the system of the present invention;

FIG2 is a flow chart of the streaming method of the present invention;

FIG3 is a flow chart of weighted evaluation of the present invention.

DETAILED DESCRIPTION

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not require further definition and explanation in the subsequent drawings.

It should also be noted that the methods used in the present invention are all conventional methods unless otherwise specified; the raw materials and devices used are all conventional commercially available products unless otherwise specified.

The present invention provides a method for transmitting audio and video in converged communication, the main steps of which include:

First, the server and the terminal of this embodiment communicate based on the WebRTC protocol (the abbreviation of Web Real-Time Communication, which is a technology that supports web browsers to conduct real-time voice calls or video chats). Although it can realize point-to-point communication, considering the particularity of the application scenario, in this embodiment, corresponding servers are still deployed in the communication network, including signaling servers and media servers (signaling services and media services can also be integrated in a single server, as shown in Figure 1, and other servers are used as redundant backups), and corresponding signaling services and media services are provided accordingly, and then the above two services are associated through event notifications. Among them, the media service is mainly responsible for processing and transmitting actual media content, such as audio, video, images, etc. In a video call, the media service is responsible for encoding, transmitting and decoding the video and audio data of both parties, so that users can see and hear each other clearly and smoothly. The event notification service is used to transmit the occurrence information of specific events to the signaling service party. These events can be changes in system status, results of user operations, the occurrence of abnormal situations, etc. The signaling service plays a control and coordination role in communication, and is responsible for establishing, managing and terminating communication sessions, and transmitting control information during the session. In each terminal, such as terminal 1, terminal 2, terminal 3, and terminal 4, a display device and corresponding terminal devices are respectively provided. The terminal devices are specifically data acquisition devices such as cameras and microphones, so as to realize real-time online audio and video mixed communication and two-way communication.

The server collects the device information of the terminal; optionally, the server collects the Session Description Protocol (SDP Unified Plan) in WebRTC during the media negotiation phase, and its text includes device information. Specifically, the developer adds device information requirements in the Unified Plan, and specifies which device information needs to be included in the Unified Plan text; the device information is obtained using the API provided by WebRTC. For example, you can use navigator.media Devices to obtain a list of media devices, from which relevant information such as graphics cards can be obtained, and the corresponding Unified Plan text can be constructed. The server collects the text of the Unified Plan and obtains the above-mentioned device information; optionally, the device information specifically includes the type information, operating system information, and firmware information of the terminal; for example, the type information is divided into tablets, mobile phones, laptops, desktops, dedicated devices, etc.; the operating system information is divided into Android, iOS, Mac OS, windows, linux, etc.; the firmware information is the corresponding version number of each system.

The server lists the encoding standards supported by each terminal based on the device information, and selects the encoding standard with the smallest bandwidth share in the list for each terminal as the adaptation encoding standard corresponding to the terminal; optionally, the encoding standards include VP8, VP9, H.264, and H.265; and when the server finds through the Unified Plan text that some terminals have graphics cards and support AV1 decoding, the above encoding standards also include AV1. Among them, H.265 is High Efficiency Video Coding (HEVC), MPEG-H part 2, specifically a video compression standard, which provides about twice the data compression ratio at the same video quality level compared to AVC, or significantly improves the video quality at the same bit rate; in the same resolution mode, H.265 occupies about 39%-44% less bandwidth than H.264, and AV1 occupies about 50% less bandwidth than H.264.

For example, the server detects that both terminal 1 and terminal 4 are dedicated Android tablets, but the version number is less than 10 (Android 8), and the supported encoding standards are H.264 and H.265. According to the usage scenario and conditions (video resolution, frame rate, bit rate, etc.), the bandwidth share is estimated. The encoding standards in the list of terminal 1 and terminal 4 are sorted from small to large: [H.265], [H.264]. Therefore, the server selects H.265 as the adaptation encoding standard for terminal 1 and terminal 4 respectively, and uses it for streaming. Similarly, if the server detects that terminal 2 and terminal 3 are web terminals of dedicated devices, they only support H.264 and other encoding standards, and H.264 is the encoding format with the smallest bandwidth share. Therefore, the H.264 coding standard is selected as the adaptive coding standard for terminal 2 and terminal 3, and the streaming is pushed based on it; similarly, if the server detects that terminal z (not shown in the figure) is a Windows desktop computer, and its browser and hardware support AV1 decoding, the bandwidth share is estimated according to the usage scenario and conditions (video resolution, frame rate, bit rate, etc.), and the list is (from small to large): [AV1], [H.265], [VP9], [VP8], [H.264]. Therefore, the server selects AV1 as the adaptive coding standard for terminal z, and pushes the stream based on it. Further, the above operation can also be completed by inquiry, that is, as shown in Figure 2, when the terminal subscribes to the stream, the server accesses the terminal or collects the Unified Plan text, and compares the coding standards from the smallest bandwidth share to the largest (such as H.265, H.264) in turn, and selects the adaptive coding standard to push the stream, and the terminal performs the corresponding streaming operation. In this way, each terminal does not need to perform transcoding operations, which not only saves network bandwidth resources, but also reduces the pressure on the terminal.

Among them, before the specific streaming, the server also needs to confirm whether the terminal is allocated enough bandwidth to provide the corresponding service. Therefore, in the initial stage or during the communication period, the server dynamically allocates the bandwidth of the terminal based on the congestion control technology in real time, and adjusts the data transmission rate according to the degree of congestion by monitoring the network congestion to avoid the occurrence of network congestion. Specifically, the congestion control technology in this embodiment adopts the congestion control algorithm of delay gradient and packet loss rate, namely the GCC (Google Congestion Control) algorithm. Further, in this embodiment, as shown in Figure 2, when the packet loss rate is greater than 25%, the server determines that the currently established converged communication audio and video transmission network has network congestion, and the server rejects the media service of the terminal that requests to join the new terminal, or suspends the media service of at least some terminals. Among them, media services include streaming reception service and streaming push service, which can also be referred to as streaming reception and streaming push. Since network congestion specifically exists in uplink congestion or downlink congestion, the suspension of terminal media services in actual work can be determined according to the congestion situation, such as suspending streaming reception or streaming push of some terminals, or suspending both streaming reception and streaming push, so as to adjust the bandwidth usage of the overall meeting to meet actual needs.

When the server detects network congestion (as mentioned above, the communication data packet loss rate is greater than 25%), the media services of the terminals ranked higher in order are suspended until the network congestion ends;

When the network congestion is detected to be over, the server reversely resumes the media services of the suspended terminals. Reverse should be understood as the reverse order of suspending the terminals, and resumes the services from the last suspended terminal in chronological order until the network congestion is detected or all terminals resume the media services.

The above process should be understood as a dynamic process, that is, when the bandwidth does not match the video conferencing requirements, the server will make real-time adjustments based on the detected network congestion to avoid the problem of idle bandwidth other than the reserved bandwidth when network resources improve or when a terminal exits voluntarily; or the problem of terminals that need to be paused still occupying bandwidth when the bandwidth is insufficient.

A corresponding strategy is required to suspend the media service of a terminal. Suspending the terminal that enters the meeting later or has a poor network environment based on conventional ideas is only applicable to general media services, and is not applicable to special situations or emergency situations. Especially in some special meetings, the worse the network environment, the more complicated the scenario is, and the corresponding meeting support is required, and some participants must stay online. Therefore, the method of this embodiment also includes a step in which the server performs a weighted evaluation on each terminal, and suspends the terminal service accordingly according to the result of the weighted evaluation.

Specifically, the server collects the user tag of each user of each terminal, and extracts feature information from the user tag; the server calculates the weight of the feature information; wherein the weight is obtained based on the hierarchical analysis method, and the server calculation step includes:

Determine the hierarchical model for weight calculation;

Establish the importance judgment matrix of feature information;

The importance of feature information is calculated by the eigenvector method, and the obtained value is normalized to obtain the weight:

The server performs a weighted evaluation on each terminal based on the weight;

When the server detects network congestion, the server suspends a media service of at least one of the plurality of terminals based on a result of the weighted evaluation.

Among them, the methods of saving system computing resources are as follows:

S1. When the terminal initially connects to the server, the server collects the user tags of each terminal. Among them, the characteristic information of the user tag in this embodiment is the title of the user tag and the corresponding options. At least part of the titles and corresponding options of the user tags are given by the conference initiator in the conference creation stage according to the system preset, suggestion or custom method, such as [Position], [Willingness to Participate], [Emergency], [Special Mark], etc. Further, there are multiple options under each title (different titles have different options, such as: "Yes", "No", "Senior", "Intermediate", "Assistant", "Supervisor", etc.), and the conference initiator can score the importance of each title; the user selects the options of the aforementioned user tags at the terminal (each option has a corresponding score built in), and the terminal reports it when connecting to the server.

S2. The server allocates some computing resources to perform weight calculation and weighted evaluation. The weights are obtained based on the hierarchical analysis method, as shown below:

S21, determine the weight calculation hierarchical structure model, which can be conventionally divided into a target layer, a quasi-side layer, and a solution layer;

S22, establishing a feature information importance judgment matrix (pairwise comparison matrix);

Assuming that for the criterion of importance, there are [Position] and [Special Tag] in the user tag, and a pairwise comparison is performed, the following judgment matrix is obtained (Table 1):

Importance Position Special markings Position 1 3 Special markings 1/3 1

Table 1

S23, calculating weight;

The importance of the feature information is calculated by the feature vector method, where the coefficient of the importance is the above-mentioned artificially set value, [Position] is 1, and [Special Mark] is 3;

Compute the row product of a matrix:

[Position] Row product = 1×3 = 3;

[Special mark] Row product = (1/3) × 1 = 1/3;

Calculate the nth root of the product of each row (n is the matrix order, which is 2 in this embodiment);

The nth root of [position] = 3^(1/2)≈1.732;

[Special mark] nth root = (1/3)^(1/2)≈0.577;

The importance values of each title calculated are normalized to obtain the weight of each title:

The weight of [position] = 1.732/(1.732+0.577)≈0.75;

The weight of [special mark] = 0.577/(1.732+0.577)≈0.25;

S24, weighted assessment;

Assume that there are terminal z ₁ and terminal z ₂ among the terminals, and the option score of user A of terminal z ₁ under the title [Position] is 0.8, and the option score of user A under the title [Special Mark] is 0.6; the option score of user B of terminal z ₂ under the title [Position] is 0.6, and the option score of user B under the title [Special Mark] is 0.8. Since the terminal corresponds to the user, it can be understood that the result of the terminal weighted evaluation is the result of the user weighted evaluation, that is, the weighted value.

The result of the weighted evaluation of user A (the weighted value of terminal z ₁ ) = 0.8 × 0.75 + 0.6 × 0.25 = 0.75;

The result of the weighted evaluation of user B (the weighted value of terminal z ₂ ) = 0.6 × 0.75 + 0.8 × 0.25 = 0.65;

Optionally, dynamic weighting can be introduced in the above weighted evaluation, and dynamic weights are also provided in the weighted evaluation. As shown in FIG3, under the premise that the server detects network congestion, the server detects in real time whether the correlation between the content of the current media service and the user tag exceeds the threshold. When the correlation exceeds the threshold, the server performs a secondary weighted evaluation on the terminal based on the dynamic weight, and based on the result of the secondary weighted evaluation, the server adjusts the bandwidth of each terminal to suspend or resume the media service of some terminals. The dynamic weight is specifically to preset a weight item (such as [Conference Relevance]) in the media service, and calculate the corresponding weight together with other titles according to the above steps. At the same time, a large language model is deployed on the server or an existing large language model is called. Further, the server extracts information from the content of the media service in real time based on the large language model, and the server calculates the correlation between the result of the information extraction and the user tags covered by each terminal to obtain the calculation result; the server compares the calculation result with the threshold. Preferably, in the above scheme, a method that is conducive to saving system computing resources is to prepare the conference process materials in advance. Specifically, the process materials are converted into text and input into the large language model during the media service creation stage (i.e., the conference creation stage); and the content of the media service (conference audio converted into text) is continuously input into the large language model in real time after the communication starts through the attention mechanism. During this period, the server system periodically inquires the large language model about the process of the current media service through the question-and-answer mode of the large language model. If the correlation between the process (which can be understood as the process title) and a user tag of a certain terminal exceeds the threshold after the inquiry, the server dynamically weights the terminal or all terminals, and forms a secondary weighted evaluation based on the original weighted evaluation. The specific operation in this embodiment is to assign a score to the [Conference Relevance] title of the terminal (previously, each terminal was assigned 0 for this item, so it was not included in the initial weighted evaluation), so as to increase the result of the secondary weighted evaluation of the terminal. Furthermore, the secondary weighted launch mechanism can be confirmed according to actual conditions. On the basis of the above embodiments, a simpler way is that when the server detects that the current meeting process has ended or a new meeting process has started, it re-assigns the score of the [Meeting Relevance] title to 0, thereby exiting the secondary weighted evaluation and returning to the initial weighted evaluation result.

S3. When the server detects network congestion (packet loss rate is greater than 25%), it suspends media services for some terminals with low weighted evaluation results;

For example: Based on the above weighted evaluation results:

The initial weighted evaluation result of user A = x _a ×m ₁ + y _a ×m ₂ + 0 ×m ₃ = a _start ;

Where a _start is the initial weighted evaluation result of user A; m ₁ is the weight of [Position], x _{a is} the score corresponding to the option selected by user A under the title [Position]; m ₂ is the weight of [Special Tag], y _a is the score corresponding to the option selected by user A under the title [Special Tag]; m ₃ is the weight of [Conference Relevance];

The initial weighted evaluation result of user B = x _b ×m ₁ + y _b ×m ₂ + 0 ×m ₃ = b _start ;

Where b _start is the initial weighted evaluation result of user B; x _b is the score corresponding to the option selected by user B under the title [Position]; y _b is the score corresponding to the option selected by user B under the title [Special Mark];

Wherein, b _start < a _start , and a _start is smaller than the weighted evaluation results of other terminals; therefore, the server preferentially suspends the media service of terminal z ₂ based on the result;

Furthermore, when there is a process called "Construction Site Guidance" in the conference process, and a user tag of terminal z ₂ is the customized "Construction Party", when the large language model feedbacks that the conference process is "Construction Site Guidance", the server calculates through a model or algorithm (such as a bag-of-words model, a latent semantic analysis model, etc.) that the relevance between terminal z ₂ and the current conference process exceeds the manually set threshold, and thus assigns a value λ to terminal z ₂ [conference relevance]. The result of the adjusted secondary weighted evaluation is:

The secondary weighted evaluation result of user B = x _b ×m ₁ + y _b ×m ₂ + λ ×m ₃ = b ^' ;

Where b ^' is the result of the second weighted evaluation after dynamic weighting of user B, and it is assumed that b ^' ＞a _start ;

Thus, dynamic weighting is implemented for terminal z _2. When the conference process is strongly correlated with the user tag, the media service of terminal z ₂ is restored through dynamic weighting, and the media service of terminal z ₁ is suspended.

It should be noted that in the above embodiments, suspending a terminal by default relieves bandwidth pressure. When suspending a terminal cannot end network congestion, the terminals are continuously suspended in ascending order of weighted values until network congestion ends.

S4. When the network congestion is detected to be over, the server reversely resumes the media service of the suspended terminal. For example: when both terminal z ₁ and terminal z ₂ are suspended and there is no dynamic weighting, the push/receive streaming of terminal z ₁ is resumed first, and if no network congestion is detected, the push/receive streaming of terminal z ₂ is resumed. Similarly, the server resumes the terminal service operation until the network congestion is detected, or when all terminals resume the media service.

While dynamically allocating bandwidth to terminals, the server also reserves bandwidth, that is, it keeps a portion of bandwidth free and reserves bandwidth resources for unsubscribed streams so that subsequent terminals can conduct meetings normally. The specific proportion of reserved bandwidth is determined based on actual conditions or past situations.

The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the protection scope of the present invention.

Claims (10)

1. A method for transmitting audio and video in converged communication, characterized in that the steps include: The server establishes connections with multiple terminals, and the server collects device information of the multiple terminals; The server lists the supported coding standards corresponding to each of the multiple terminals based on the device information, and selects the coding standard with the smallest bandwidth share from the corresponding supported coding standards as the adapted coding standard corresponding to each of the multiple terminals; The server dynamically allocates bandwidth of the multiple terminals based on congestion control technology; The server pushes the corresponding streams of the adapted coding standard to the multiple terminals according to the request information of the multiple terminals and the bandwidth of the multiple terminals; The server collects a user tag of each user of each terminal among the multiple terminals, and extracts feature information from the user tag; the server calculates a weight of the feature information; The server performs a weighted evaluation on each of the multiple terminals based on the weight; When the server detects network congestion, the server suspends a media service of at least one of the plurality of terminals based on a result of the weighted evaluation; A dynamic weight is also provided in the weighted evaluation; when the server detects network congestion, when the correlation between the content of the media service pushed by the server and the user tag exceeds a threshold, the server performs a secondary weighted evaluation on each of the multiple terminals based on the dynamic weight, and the server suspends or resumes the media service of at least one of the multiple terminals based on the result of the secondary weighted evaluation. 2. The method for transmitting audio and video in converged communication according to claim 1, wherein the weight is obtained based on a hierarchical analysis method, and the step of the server calculating the weight comprises: Determining a hierarchical model for said weight calculation; Establishing an importance judgment matrix of the feature information; The importance of the feature information is calculated by a feature vector method, and the obtained value is normalized to obtain the weight. 3. The method for transmitting audio and video in converged communication according to claim 2, wherein the step of calculating the correlation between the content of the media service and the user tag comprises: The server extracts information from the content of the media service in real time based on the large language model, and the server calculates correlation between the information extraction result and the user tag of each user of each of the multiple terminals to obtain a calculation result; The server compares the calculation result with the threshold value. 4. The method for transmitting audio and video in converged communication according to claim 1, characterized in that the server and the plurality of terminals are connected via a public network or a private network; The data transmission between the server and the plurality of terminals is based on a web video and voice real-time communication protocol; The congestion control technology is a congestion control algorithm based on the delay gradient and packet loss rate of the web video and voice real-time communication protocol. 5. The method for transmitting audio and video in converged communication according to claim 4, wherein the step of collecting device information of the plurality of terminals by the server comprises: The webpage video and voice real-time communication protocol is configured with a session description protocol, and the text of the session description protocol includes the device information; The server collects the text of the session description protocol, and obtains the device information by the server; The device information includes type information, operating system information, and firmware information of the multiple terminals. 6. The method for transmitting audio and video in converged communication according to claim 1, wherein the coding standard comprises VP8, VP9, H.264, and H.265; When at least one of the multiple terminals supports AV1 decoding, the encoding standard also includes AV1. 7. The method for transmitting audio and video in converged communication according to claim 1, characterized in that the result of the weighted evaluation is the weighted value of each terminal in the plurality of terminals, and the plurality of terminals are arranged from small to large according to the weighted value; When network congestion is detected, media services of corresponding terminals are suspended in descending order of weighted values until the network congestion ends; When it is detected that the network congestion ends, the server reversely resumes the media service of the suspended terminal. 8. The converged communication audio and video transmission method according to claim 1 is characterized in that the characteristic information of the user tag includes multiple titles; and the user of each of the multiple terminals selects or marks the title. 9. The method for transmitting audio and video in converged communication according to claim 8, characterized in that the characteristic information of the user tag also includes a plurality of preset options, and at least one of the preset options is correspondingly provided in at least part of the titles; The preset options are each provided with a score, and the score is used for the weighted evaluation. 10. The converged communication audio and video transmission method according to claim 8 or 9, characterized in that each of the titles is configured with an importance coefficient, and the coefficient is used to calculate the weight.