A multi-terminal live broadcast interaction data transmission method and system based on frequency modulation relay

Abstract

The application relates to the technical field of digital communication, and discloses a multi-terminal live broadcast interaction data transmission method and system based on frequency modulation relay, which comprises the following steps: acquiring multi-terminal live broadcast interaction data under frequency modulation relay, extracting delay tolerance characteristics of different data types and constructing a transmission resource allocation matrix; performing repeated data detection on the multi-terminal live broadcast interaction data, combining network delay compensation to establish a time sequence distribution mode of repeated data, generating a check reference sequence; embedding the check reference sequence into a data stream coded according to the transmission resource allocation matrix and the multi-terminal live broadcast interaction data, forming a composite transmission data stream, and performing frequency modulation on the composite transmission data stream to generate a frequency modulation transmission signal; extracting the check reference sequence from the frequency modulation transmission signal at a receiving end, comparing and analyzing the actual distribution of received repeated data, and adjusting the transmission resource allocation matrix according to the comparison result. The application provides data transmission guarantee for the conversion of network data into a frequency modulation broadcast signal.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital communication, in particular to a multi-end live interactive data transmission method and system based on frequency modulation relay. BACKGROUND

[0002] Traditional network data transmission mainly relies on TCP / IP protocol stack for point-to-point or point-to-multipoint data distribution. However, in large-scale concurrent live scenarios, network congestion and bandwidth limitation problems are increasingly prominent. Frequency modulation broadcasting has the advantages of wide coverage, stable transmission and multi-user simultaneous reception, and is suitable for real-time data distribution scenarios of large-scale users. In recent years, researchers have begun to explore the technical path of converting network interactive data into frequency modulation broadcast signals. Through frequency modulation relay technology, efficient conversion of network data to broadcast signals is realized, combining the interactivity of the Internet and the wide-area coverage of broadcasting, providing a new technical idea for solving the data transmission bottleneck in multi-end live interaction. At the same time, with the progress of coding technology and signal processing technology, data modulation and demodulation technology based on frequency modulation carrier is also maturing, laying a technical foundation for the broadcast transmission of network data.

[0003] However, the existing network data to frequency modulation broadcast signal conversion technology has many shortcomings. First, in terms of transmission resource allocation, the existing technology lacks in-depth analysis of the delay tolerance characteristics of different data types, resulting in unreasonable transmission resource allocation. High-priority real-time interactive data is mixed with low-priority auxiliary data, causing transmission delay of critical data. Second, there is a problem of repeated data transmission in multi-end live interaction scenarios. Users' same operations or comments will generate repeated interactive data on multiple terminals. The existing technology directly models based on the timestamp of data arriving at the server, without stripping the timestamp pollution caused by different network transmission delays of different terminals and different paths, resulting in distortion of the timing model established, which cannot be used as a reliable benchmark to evaluate the quality of subsequent one-way broadcast transmission. In addition, the existing frequency modulation technology lacks a precise transmission quality optimization mechanism when processing multi-end interactive data, and cannot adjust the resource allocation strategy in real time according to the actual transmission effect. When the data synchronization decreases, it is difficult to determine whether it is caused by network jitter at the front end or transmission damage of the frequency modulation broadcast channel, resulting in poor adjustment strategy effect and low transmission efficiency. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] To solve the above technical problems, the present application provides the following technical scheme: a multi-end live interactive data transmission method based on frequency modulation relay, comprising:

[0006] Obtain multi-end live interactive data under frequency modulation relay, extract delay tolerance characteristics of different data types and construct a transmission resource allocation matrix;

[0007] Perform repeated data detection on the multi-end live interactive data, and establish a time sequence distribution mode of repeated data in combination with network delay compensation, generate a check reference sequence based on the time sequence distribution mode;

[0008] Embed the check reference sequence into a data stream coded according to the transmission resource allocation matrix and the multi-end live interactive data, form a composite transmission data stream, and perform frequency modulation on the composite transmission data stream according to the transmission resource allocation matrix to generate a frequency modulation transmission signal;

[0009] Extract the check reference sequence from the frequency modulation transmission signal at the receiving end, and compare and analyze the actual distribution of the received repeated data, and adjust the transmission resource allocation matrix according to the comparison result.

[0010] As a preferred scheme of the multi-end live interactive data transmission method based on frequency modulation relay, wherein: the delay tolerance characteristics of different data types in the multi-end live interactive data are extracted by using a data type identifier to identify the data categories of the multi-end live interactive data;

[0011] A user behavior response statistical model is constructed for the data types, and the time interval from data sending to the observable interactive response of the user is recorded;

[0012] Based on the time interval, the change trend of the user response rate with the increase of the time interval is analyzed, the delay time value corresponding to the time when the response rate decreases to the reference response rate threshold is determined, and the delay time value is normalized to obtain the delay tolerance coefficient.

[0013] As a preferred scheme of the multi-end live interactive data transmission method based on frequency modulation relay, wherein: the construction of the transmission resource allocation matrix includes converting the delay tolerance coefficient into an encoding priority value through a linear inverse proportional function;

[0014] The encoding priority value is used as a dominant factor of encoding resource allocation, a mapping relationship between the encoding priority value and the encoding bit number and the error correction strength is established, and the corresponding encoding priority value, encoding bit number and error correction strength are allocated for each data type;

[0015] The transmission resource allocation matrix is constructed based on the encoding priority value, the encoding bit number and the error correction strength of each data type.

[0016] As a preferred scheme of the multi-terminal live broadcast interactive data transmission method based on frequency modulation relay, wherein: the mapping relationship between the encoding priority value and the encoding bit number and the error correction strength comprises load evaluation of available frequency bands based on the spectrum resource capacity of frequency modulation broadcast, calculation of a frequency band load coefficient, and division of the available frequency bands into low-load frequency bands, medium-load frequency bands, and high-load frequency bands according to the frequency band load coefficient.

[0017] The data types to be allocated are arranged in descending order of the encoding priority value, and when there are multiple data types with the same encoding priority value, the ratio of the delay tolerance coefficient of each data type to the encoding priority value is calculated as a secondary sorting basis, and the ratio of the delay tolerance coefficient to the encoding priority value is arranged in ascending order;

[0018] The frequency band resources are allocated to the data types in sequence according to the sorting order, and the data types with high sorting order are preferentially allocated to the low-load frequency bands, and when the low-load frequency band capacity is insufficient, the data types are sequentially allocated to the medium-load frequency bands and the high-load frequency bands.

[0019] For the data types allocated to the high-load frequency bands, the adjusted encoding bit number is obtained by combining the basic encoding bit number of the corresponding data type and the load compensation coefficient; the redundancy code increase amount is calculated according to the reduction amount of the encoding bit number, and the adjusted error correction strength is obtained by combining the redundancy code increase amount and the basic error correction strength.

[0020] As a preferred scheme of the multi-terminal live broadcast interactive data transmission method based on frequency modulation relay, wherein: the repeated data detection is performed on the multi-terminal live broadcast interactive data, and the time sequence distribution pattern of the repeated data is established by combining the network delay compensation comprises content similarity detection of the multi-terminal live broadcast interactive data, and identification of repeated data generated by multiple terminals;

[0021] The source terminal and the historical average network delay of the repeated data are obtained, and the network delay compensation of the repeated data is performed to obtain user response time data;

[0022] The event trigger time point causing the repeated data is identified, the difference between the user response time of each terminal and the event trigger time point is calculated to obtain the relative response delay data of each terminal;

[0023] The relative response delay data is analyzed by using a time window, the reference occurrence frequency and the standard time interval of the repeated data are calculated, and the time sequence distribution pattern of the repeated data is formed.

[0024] As a preferred scheme of the multi-terminal live broadcast interactive data transmission method based on frequency modulation relay, wherein: the check reference sequence is generated based on the time sequence distribution pattern comprises obtaining the reference occurrence frequency and the standard time interval in the time sequence distribution pattern;

[0025] In combination with the multi-terminal synchrony repetition mode, the same interaction data generated by the multiple terminals within the same time window is modeled by a synchrony distribution model to obtain a synchrony distribution feature, wherein the synchrony distribution model adopts a probability model considering measurement uncertainty, and a delay uncertainty parameter is constructed based on the standard deviation of the network delay of each terminal;

[0026] The reference appearance frequency, the standard time interval, and the synchrony distribution feature are digitally encoded, and the encoding results are combined to generate a verification data unit containing the reference appearance frequency, the standard time interval, a synchrony coefficient, and the delay uncertainty parameter;

[0027] All verification data units are arranged in chronological order to form a verification reference sequence.

[0028] As a preferred scheme of the multi-terminal live interaction data transmission method based on frequency modulation relay according to the application, wherein: the forming of the composite transmission data stream includes differentiating the encoding of the multi-terminal live interaction data according to the encoding priority value of each data type in the transmission resource allocation matrix;

[0029] The verification reference sequence is dispersedly embedded into the encoded data stream according to a preset insertion rule to form a composite data stream.

[0030] As a preferred scheme of the multi-terminal live interaction data transmission method based on frequency modulation relay according to the application, wherein: the frequency modulation modulation of the composite transmission data stream according to the transmission resource allocation matrix to generate a frequency modulation transmission signal includes serial-parallel conversion of the composite data stream to convert the serial data stream into a parallel data format suitable for frequency modulation modulation;

[0031] Different bit positions of the parallel data are modulated onto different carrier frequencies by using frequency modulation modulation technology;

[0032] According to the encoding priority value in the transmission resource allocation matrix, different frequency band carriers are allocated to the composite data stream to generate a frequency modulation transmission signal containing the multi-terminal live interaction data and the verification reference sequence.

[0033] As a preferred scheme of the multi-terminal live interaction data transmission method based on frequency modulation relay according to the application, wherein: the adjustment of the transmission resource allocation matrix according to the comparison result includes demodulation and decoding of the frequency modulation transmission signal at the receiving end to separate the original multi-terminal live interaction data and the verification reference sequence;

[0034] The repeated data is extracted from the original multi-terminal live interaction data, and the appearance frequency, time interval, and synchrony distribution feature of the repeated data are counted to form actual distribution data;

[0035] The actual distribution data is compared with the verification reference sequence, a joint probability model is established based on the delay uncertainty parameter in the verification reference sequence by using Bayesian inference method, the joint influence of network layer disturbance and broadcast layer disturbance is comprehensively considered, a synchronization deviation caused by the transmission quality reduction of the frequency modulation broadcast is obtained, and the transmission resource allocation matrix is adjusted according to the synchronization deviation.

[0036] A multi-end live interactive data transmission system based on frequency modulation relay, wherein:

[0037] A live data acquisition module acquires multi-end live interactive data under frequency modulation relay, extracts delay tolerance features of different data types, and constructs a transmission resource allocation matrix.

[0038] A verification reference sequence module detects repeated data of the multi-end live interactive data, establishes a time sequence distribution mode of the repeated data in combination with network delay compensation, generates a verification reference sequence based on the time sequence distribution mode, and adjusts the transmission resource allocation matrix according to a synchronization deviation caused by the transmission quality reduction of the frequency modulation broadcast.

[0039] A frequency modulation signal conversion module embeds the verification reference sequence into a data stream coded according to the transmission resource allocation matrix and the multi-end live interactive data, forms a composite transmission data stream, and generates a frequency modulation transmission signal by frequency modulation and modulation of the composite transmission data stream according to the transmission resource allocation matrix.

[0040] A sequence comparison and analysis module extracts the verification reference sequence from the frequency modulation transmission signal at a receiving end, compares and analyzes the actual distribution of the repeated data received, and adjusts the transmission resource allocation matrix according to the comparison result.

[0041] The beneficial effects of the present application: the multi-end live interactive data transmission method based on frequency modulation relay provided by the present application realizes differentiated processing of different data types by constructing a transmission resource allocation matrix based on delay tolerance characteristics, effectively solving the key data delay problem caused by unreasonable resource allocation in traditional frequency modulation transmission. By introducing a repeated data detection and timing distribution modeling mechanism combined with network delay compensation, repeated interactive data in a multi-end environment can be identified, the reference occurrence frequency and standard time interval are calculated and combined with the synchronization distribution characteristics and delay uncertainty parameters to establish a more realistic timing law, effectively avoiding the bandwidth waste caused by redundant data transmission. By embedding the verification reference sequence into the composite transmission data stream, real-time monitoring and comparison of the transmission process are realized, and the receiving end can establish a joint probability model based on the Bayesian inference method combined with the delay uncertainty parameter to distinguish the influence of network layer disturbance and broadcast layer disturbance, and adjust the transmission resource allocation strategy according to the synchronization deviation, thereby improving the adaptability and robustness of the frequency modulation broadcast system to network environment changes. Not only the integrity and synchronization of the data are maintained, but also the wide coverage advantage of frequency modulation broadcast is fully utilized, providing stable and reliable data transmission guarantee for the conversion of network data to frequency modulation broadcast signals in multi-end live interactive scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0043] Figure 1 The overall flowchart of a multi-end live interactive data transmission method based on frequency modulation relay provided for the first embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings in the specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0045] Embodiment 1, refer to Figure 1 For an embodiment of the present application, a multi-end live interactive data transmission method based on frequency modulation relay is provided, comprising:

[0046] S1: Obtain multi-terminal live interactive data under frequency modulation relay, extract delay tolerance characteristics of different data types, and construct a transmission resource allocation matrix.

[0047] In this embodiment, the multi-terminal live interactive data is derived from the interactive information generated by multiple user terminals in a live scene, including but not limited to text comments, like operations, gift rewards, bullet screen messages, and voice information. Real-time acquisition of multi-terminal live interactive data is achieved by setting up a data collection device, which is in communication connection with a live platform server. Through the configuration of a data collection interface, the interactive data stream uploaded by each terminal is received in real time.

[0048] It should be noted that in the traditional network data transmission scene, multi-terminal live interactive data is usually transmitted through a two-way network connection in a point-to-point or point-to-multipoint manner, which is prone to problems such as delay, packet loss, and excessive server pressure in a high-concurrency scenario, and it is difficult to guarantee the real-time and reliability of different terminal user interactive data. Therefore, this embodiment converts multi-terminal interactive data into a traditional FM broadcast channel for transmission, where FM broadcast has the characteristics of wide coverage, stable transmission, and no need to establish point-to-point connection. Once sent, it can be received by all user terminals in the covered area, avoiding the linear growth of server burden with the increase of user number in traditional network transmission. However, in FM broadcast, interactive data based on network protocols is converted into FM broadcast signals for one-way transmission, which faces many problems: first, the frequency spectrum resources of FM broadcast are limited and need to be shared, unlike network transmission which can allocate bandwidth according to demand; second, FM broadcast is a one-way transmission and cannot adjust the transmission quality through an instant feedback mechanism like network transmission; third, different types of interactive data have different sensitivity to transmission delay and packet loss. If a uniform transmission strategy is adopted, it will lead to a decline in the transmission quality of key data or waste of frequency spectrum resources.

[0049] Therefore, this embodiment proposes a differentiated resource allocation strategy based on delay tolerance characteristics. Delay tolerance refers to the maximum time delay range that a user can accept between the sending of a specific type of interactive data and its perception by the user. Different types of interactive data have different delay tolerance characteristics. For example, gift reward data usually needs to be displayed within 1 second to ensure the user's instant feedback experience, and has a low delay tolerance. Historical bullet screen data allows a delay of 2-3 seconds, and has a high delay tolerance. By accurately extracting and quantifying the delay tolerance characteristics of each data type, a scientific basis can be provided for subsequent transmission resource optimization and allocation, achieving a balance between the priority protection of key data and the overall transmission efficiency under the constraint of limited frequency spectrum resources.

[0050] Further, the delay tolerance feature of different data types in the multi-terminal live interaction data is extracted by using a data type identifier to identify the data categories of the multi-terminal live interaction data. A user behavior response statistical model is constructed for the data types, and the time interval from data sending to the user producing observable interaction response is recorded. Based on the time interval, the change trend of the user response rate with the increase of the time interval is analyzed, the delay time value corresponding to the response rate falling below the baseline response rate threshold is determined, and the delay time value is normalized to obtain the delay tolerance coefficient.

[0051] The data type identifier is used to identify the data categories of the multi-terminal live interaction data, which means that the data type identifier can use a deep learning-based classification model to achieve preliminary determination of the data categories, such as text, operation or notification, by analyzing the protocol field and type identification field of the data packet. Then, the feature vector of the data content is extracted, including data length, sending frequency and user triggering mode, and the pre-trained classification model is used for fine identification, and finally the multi-terminal live interaction data is divided into specific data types.

[0052] For each identified data type, a user behavior response statistical model is constructed to quantify the user's perception sensitivity to different data delays. The construction of the user behavior response statistical model includes recording the time interval from the generation of data at the sending end to the user's observable interaction response for a specific data type. The observable interaction response refers to the subsequent operation behavior of the user after receiving a certain type of data, such as the time the user replies after seeing a comment, the time the user likes after seeing a gift reward, the time the user enters a product page after seeing a product notification, or the time the user sends related comments after seeing a scroll. By statistically analyzing the behavior data of a large number of user samples under different delay conditions, the distribution characteristics of the time interval corresponding to each data type are obtained.

[0053] The change trend of the user response rate with the increase of the time interval is analyzed based on the time interval, a delay time value corresponding to when the response rate drops to a benchmark response rate threshold is determined, and the delay time value is normalized to obtain a delay tolerance coefficient. For each data type, the response rate of the user in different time delay intervals is calculated based on the distribution characteristics of the time interval, and a response rate-delay time curve is drawn. The delay time value corresponding to when the response rate of the response rate-delay time curve drops to the benchmark response rate threshold is calculated, and the benchmark response rate threshold is usually set to 50% to 70% of the initial response rate in this embodiment. The delay tolerance coefficient is obtained by normalizing the delay time value, so that the coefficient of the data type with the lowest delay tolerance is close to zero, and the coefficient of the data type with the highest delay tolerance is close to one, thereby intuitively reflecting the difference in time sensitivity of different data types in the transmission process and providing a quantitative basis for subsequent resource allocation.

[0054] Exemplarily, assuming that there are three main data types in the live broadcast system that need to be analyzed, for the gift reward data, it is found through statistics that when the delay is within 0.5 seconds, the subsequent like response rate of the user is 90%; when the delay is 1 second, the response rate drops to 50%; and when the delay is 1.5 seconds, the response rate drops to 20%; the benchmark response rate threshold is set to 55% of the initial response rate (i.e., 49.5%), and the delay time value of the gift reward data can be determined as 1 second through the response rate curve. For the ordinary comment data, it is found through statistics that when the delay is within 1 second, the subsequent interaction response rate of the user is 80%; when the delay is 2 seconds, the response rate drops to 50%; and when the delay is 3 seconds, the response rate drops to 30%, and the same benchmark response rate threshold proportion is set to determine that the delay time value of the ordinary comment data is 2 seconds. For the historical barrage data, it is found through statistics that when the delay is within 3 seconds, the user viewing response rate is 70%; when the delay is 5 seconds, the response rate drops to 40%; and when the delay is 8 seconds, the response rate drops to 20%, and the delay time value of the historical barrage data is determined as 5 seconds. After normalizing the delay time value, it can be seen that the delay tolerance coefficient of the gift reward data is 0, indicating that it is most sensitive to delay and needs the highest priority guarantee; the delay tolerance coefficient of the historical barrage data is 1, indicating that it is least sensitive to delay and can accept a lower transmission priority; the coefficient of the ordinary comment data is 0.25, which is at an intermediate level, thereby providing a clear basis for subsequent resource allocation.

[0055] Further, the constructing the transmission resource allocation matrix comprises converting the delay tolerance coefficient into an encoding priority value through a linear inverse proportion function. The encoding priority value is taken as a dominant factor of encoding resource allocation, a mapping relationship between the encoding priority value and the encoding bit number and the error correction strength is established, and the corresponding encoding priority value, encoding bit number and error correction strength are allocated to each data type based on the encoding priority value, the encoding bit number and the error correction strength of each data type. The transmission resource allocation matrix is constructed based on the encoding priority value, the encoding bit number and the error correction strength of each data type.

[0056] In the frequency modulation broadcast transmission scenario, limited spectrum resources need to be reasonably allocated among various data types, and data with low delay tolerance needs to obtain better transmission quality guarantee, therefore, the delay tolerance coefficient is converted into an encoding priority value through a linear inverse proportion function, so that the data type with smaller delay tolerance coefficient corresponds to larger encoding priority value, thereby obtaining more transmission resource guarantee in subsequent resource allocation; the calculation formula of the encoding priority value is as follows:

[0057]

[0058] wherein, represents the encoding priority value of the i th data type; represents the preset maximum encoding priority value; represents a conversion coefficient, which is used to control the influence strength of the delay tolerance on the priority, and is adjusted according to the number of data types and the delay tolerance distribution range in the specific application scenario, and is usually valued in the range of fifty to one hundred, so as to ensure that the converted encoding priority value has reasonable distinguishability; represents the delay tolerance coefficient of the i th data type. Secondly, in the traditional frequency modulation broadcast system, a unified encoding scheme and error correction mechanism are usually adopted for all data, which is simple to implement, but in the case of limited spectrum resources, the characteristics of different data types cannot be optimized, important data cannot be fully guaranteed, and data insensitive to delay occupies too many resources; the present application establishes a mapping relationship between the encoding priority value, the encoding bit number and the error correction strength, determines appropriate encoding bit number and error correction strength for each data type, and maximizes the utilization efficiency of the overall spectrum resources on the premise of guaranteeing the transmission quality of high-priority data.

[0059] Specifically, the establishing the mapping relationship between the encoding priority value, the encoding bit number and the error correction strength, and allocating the corresponding encoding priority value, encoding bit number and error correction strength to each data type comprises,

[0060] Specifically, the establishing the mapping relationship between the encoding priority value, the encoding bit number and the error correction strength, and allocating the corresponding encoding priority value, encoding bit number and error correction strength to each data type comprises,

[0061] ​In practical applications, FM broadcast bands typically already carry audio signals and other auxiliary information. The spectrum resources available for interactive data transmission are the remaining spectrum space obtained through spectrum reuse or time-division multiplexing techniques without affecting the primary audio quality. Therefore, based on the FM broadcast spectrum resource capacity, a load assessment is performed on the available frequency bands, calculating the band load factor. Based on this load factor, the available frequency bands are divided into low-load, medium-load, and high-load bands. The data types to be allocated are sorted in descending order of their coding priority values. When multiple data types have the same coding priority value, the ratio of the delay tolerance coefficient to the coding priority value is calculated as a secondary sorting criterion, and the data types are then sorted in ascending order based on this ratio. Frequency band resources are allocated to each data type in the sorting order, prioritizing the data types ranked higher in the sorting to the low-load band. When the low-load band capacity is insufficient, the data types are then allocated to the medium-load and high-load bands in that order.

[0062] The formula for calculating the frequency band load factor is as follows:

[0063]

[0064] in, Indicates the first Load factor of each frequency band; Indicates the first Current channel occupancy rate of each frequency band; The maximum reference value for channel occupancy; Indicates the first Interference intensity in each frequency band; This represents the maximum reference value for interference intensity; and These represent the weighting coefficients of channel occupancy and interference intensity in load assessment, respectively. In this embodiment, The value is 0.6. The value is set to 0.4; by utilizing the load factor, the degree of congestion and transmission quality potential of the frequency band are comprehensively reflected, providing a quantitative basis for subsequent frequency band allocation.

[0065] Furthermore, dividing the available frequency bands into low-load, medium-load, and high-load bands based on the frequency band load factor refers to setting a load threshold. and ,in When the frequency band load factor When, the corresponding frequency band is divided into a low-load frequency band; when When the frequency band is divided into medium load bands; when the frequency band load factor is... At that time, it is divided into high-load frequency bands; in this embodiment, the load threshold is... The value is 0.3, which is the load threshold. The value is 0.7. The low-load frequency band has better transmission conditions and lower interference levels, making it suitable for carrying high-priority data with strict transmission quality requirements, such as gift rewards. The medium-load frequency band has moderate transmission conditions and can carry general-priority data, such as text comments and likes. The high-load frequency band has poor transmission conditions and is usually only used to carry low-priority data with low transmission quality requirements, such as bullet screen messages, or as a backup frequency band when spectrum resources are scarce.

[0066] After allocating frequency band resources to various data types, for data types allocated to high-load frequency bands, the transmission conditions in high-load frequency bands are poor. If standard coding parameters are still used, the transmission error rate may increase. Therefore, the adjusted coding bit count is obtained by combining the basic coding bit count of the corresponding data type with the load compensation coefficient. The formula for calculating the adjusted coding bit count is as follows:

[0067]

[0068] in, Indicates the first The number of encoded bits for each data type is adjusted. The number of encoded bits will be reduced according to the load, but the reduction is controlled to avoid data transmission failure due to excessive compression. Indicates the first The basic number of encoding bits for each data type; This represents the load compensation adjustment factor, which typically ranges from 0.1 to 0.3. Indicates the first The load factor of the frequency band to which each data type is assigned; This indicates the load threshold.

[0069] Secondly, while reducing the number of coding bits saves spectrum resources, it also reduces data redundancy, making the data more susceptible to channel noise and interference. To compensate for the effects of channel noise and interference, the redundancy of error correction coding is increased to improve the data's resistance to interference.

[0070] Specifically, the increase in redundancy is calculated based on the reduction in the number of encoded bits. This increase in redundancy is then combined with the base error correction strength to obtain the adjusted error correction strength. The formula for calculating the adjusted error correction strength is as follows:

[0071]

[0072] in, Indicates the first Error correction strength after adjustment for different data types; Indicates the first The basic error correction strength for each data type; Indicates the first The basic number of encoding bits for each data type; Indicates the first The number of encoded bits is adjusted for different data types. The number of encoded bits will be reduced according to the load, but the reduction will be controlled to avoid data transmission failure due to excessive compression.

[0073] It should be noted that when calculating the adjusted error correction strength, for every unit decrease in the number of coded bits, the error correction strength increases by one unit. This achieves equal compensation between data compression and enhanced error correction capability, thereby ensuring that even in high-load frequency bands, although data information bits are compressed, the reliability of data transmission can still be maintained by synchronously enhancing the redundancy of error correction coding, achieving a balance between spectral efficiency and transmission quality. For example, if the basic number of coded bits is 64 bits and the basic error correction strength is 16 bits, when the adjusted number of coded bits is reduced to 51.2 bits, the error correction strength will increase to 28.8 bits. The total transmission overhead remains relatively stable, but the error correction capability is enhanced, effectively solving the problem of transmission quality assurance under the condition of limited FM broadcast spectrum resources.

[0074] Furthermore, constructing the transmission resource allocation matrix based on the encoding priority value, number of encoded bits, and error correction strength for each data type refers to assigning a corresponding encoding priority value, number of encoded bits, and error correction strength to each data type within the transmission resource allocation matrix. The transmission resource allocation matrix is ​​a multi-dimensional data structure, with each row corresponding to a data type and each column corresponding to a resource parameter. The matrix has the following dimensions: ,in The total number of data types is represented by the three columns, which represent the encoding priority value, the number of encoded bits, and the error correction strength, respectively. The transmission resource allocation matrix allows for quick lookup of the transmission parameters that should be used for each data type, providing a configuration basis for the subsequent encoding and modulation process.

[0075] In view of the problem that the traditional frequency modulation broadcast adopts a unified transmission strategy for all data types, resulting in the key interactive data not being prioritized, a transmission resource allocation matrix based on delay tolerance characteristics is constructed, a user behavior response statistical model is established to quantify the time sensitivity difference of different data types; by recording the time interval from data sending to the observable interactive response of the user, the delay tolerance coefficient of each data type is calculated, the delay tolerance coefficient is converted into the encoding priority value through a linear inverse proportional function, so that the data with strict time requirements can obtain higher transmission priority and stronger error correction protection; at the same time, the available frequency spectrum is divided into different quality levels through frequency band load evaluation, and the high priority data is allocated to the low load frequency band with the best transmission condition according to the encoding priority, for the data type allocated to the high load frequency band, a compensation mechanism of encoding bit number compression and error correction strength increase is adopted, which saves the spectrum resources while maintaining the transmission reliability, and realizes the effective guarantee of the transmission quality of the key data under the constraint of limited spectrum.

[0076] S2: Repeated data detection is performed on the multi-end live interactive data, and a time sequence distribution mode of repeated data is established in combination with network delay compensation, and a check reference sequence is generated based on the time sequence distribution mode.

[0077] After completing the initial transmission resource allocation, the problem of transmission quality verification in the one-way transmission scenario of frequency modulation broadcast needs to be solved; unlike traditional network data transmission, frequency modulation broadcast is a one-way transmission mode, the sending end cannot receive the acknowledgement signal from the receiving end, and cannot compensate for the loss of packet retransmission mechanism, this one-way transmission mode solves the problem of server maintaining a large number of connections in high concurrency scenarios, but it completely cuts off the perception channel of transmission quality, so that the sending end cannot optimize and adjust according to the actual transmission effect.

[0078] In the embodiment, by deeply analyzing the inherent regularity of multi-terminal live interactive data, it is found that the interactive data generated by different terminal users in the live scene has repeatability and timing characteristics. Specifically, in the live room, multiple users may send the same or highly similar comment content, such as popular phrases such as 666, anchor is great, etc. at the same time period. Multiple users may almost simultaneously react to the same product listing event. The like operation of multiple terminals presents a clear clustering pattern in time distribution. The frequency of occurrence, time interval and multi-terminal synchronization of these repeated data should follow certain statistical rules under normal transmission conditions. However, the time stamp of the original data arriving at the server contains variable network delay, and direct modeling based on this will introduce errors. Therefore, the application introduces a network delay compensation mechanism to strip the network layer disturbance, more accurately establishes the expected timing distribution pattern of user behavior, and embeds it as a verification benchmark in the transmission signal. The receiving end can more accurately assess the transmission quality changes introduced by the frequency modulation broadcast channel by comparing the deviation of the actual received repeated data distribution from the expected benchmark, thereby realizing the monitoring and optimization of the frequency modulation broadcast transmission effect without relying on bidirectional feedback.

[0079] Specifically, the multi-terminal live interactive data is subjected to repeated data detection, and the timing distribution pattern of repeated data is established in combination with network delay compensation, which includes content similarity detection of the multi-terminal live interactive data to identify repeated data generated by multiple terminals. The source terminal and historical average network delay of the repeated data are obtained, and the repeated data are subjected to network delay compensation to obtain user response time data. The event trigger time point causing the repeated data is identified, the difference between the user response time of each terminal and the event trigger time point is calculated to obtain the relative response delay data of each terminal. The relative response delay data is analyzed by using a time window, the benchmark occurrence frequency and standard time interval of the repeated data are calculated, and the timing distribution pattern of the repeated data is formed.

[0080] The content similarity detection adopts different comparison strategies for different data types. For text data such as text comments and bullet screen messages, a combination of string exact matching and fuzzy matching is adopted. First, the same text content is quickly filtered out by using a hash function, and then the editing distance of similar but not exactly the same text is calculated. When the editing distance is less than the editing distance preset threshold, it is determined as repeated data. The editing distance threshold is adjusted according to the text length. For operation data such as like operation and gift reward, the consistency of operation type and target object is mainly compared. For notification data such as product listing and delisting notification, the consistency of event type and associated entity is mainly compared, so as to accurately identify the repeated interactive data generated by multiple terminal users.

[0081] After identifying the repeated data, the embodiment further acquires source terminal information and historical average network delay of each repeated data, the historical average network delay is obtained by continuously monitoring the round-trip time (RTT) between the terminal and the sending end and taking the average, specifically using the sliding window average method, the RTT measurement values in the recent time period (such as the last 10 minutes) are counted, and the arithmetic mean value is calculated after removing the outliers, to obtain the historical average network delay of the terminal; for each repeated data, record the timestamp of arriving at the sending end, and obtain the estimated time of the user actually producing the corresponding interactive behavior on the terminal side, that is, the user response time data, by subtracting the historical average network delay of the corresponding terminal from the arrival timestamp.

[0082] Among them, in the live scene, the large number of repeated data is usually triggered by specific events, such as the anchor announcing the draw activity, the time-limited listing of goods, and the wonderful performance clip; these events have a clear time marker in the live stream, and the event trigger time point that causes the repeated data is identified by analyzing the metadata of the live content or through an event detection algorithm; after determining the event trigger time point, the difference between the user response time of each terminal and the event trigger time point is calculated to obtain the relative response delay data of each terminal.

[0083] In the embodiment, the relative response delay data is analyzed by using a time window, the reference occurrence frequency and the standard time interval of the repeated data are calculated, and the time sequence distribution pattern of the repeated data is formed, including: for each repeated data content, the following reference parameters are extracted from the statistical results of all time windows: the reference occurrence frequency is obtained by calculating the average number of occurrences of repeated data in a unit time, and the calculation formula is as follows:

[0084]

[0085] Among them, represents the reference occurrence frequency; represents the total number of statistical time windows; represents the number of occurrences of repeated data in the th time window; represents the size of the time window.

[0086] The standard time interval is obtained by calculating the average time interval between adjacent occurrences of repeated data, and the calculation formula is as follows:

[0087]

[0088] Among them, represents the standard time interval; represents the total number of time intervals counted in all time windows; represents the th time interval value.

[0089] By extracting and modeling the reference occurrence frequency and the standard time interval, in combination with network delay compensation and relative response delay analysis, a timing distribution pattern of each repeated data is formed, wherein the timing distribution pattern not only contains statistical characteristics of the data in the time dimension, but also truly reflects the collective law of user behavior, and strips the delay disturbance of the network transmission layer, thereby providing a more reliable and accurate reference benchmark for subsequent transmission quality verification.

[0090] Further, generating a verification benchmark sequence based on the timing distribution pattern comprises obtaining the reference occurrence frequency and the standard time interval in the timing distribution pattern. In combination with the multi-terminal synchrony repetition pattern, the same interactive data generated by multiple terminals in the same time window is subjected to synchrony distribution modeling to obtain synchrony distribution characteristics, wherein the synchrony distribution modeling adopts a probability model considering measurement uncertainty, and a delay uncertainty parameter is constructed based on the standard deviation of the network delay of each terminal. The reference occurrence frequency, the standard time interval and the synchrony distribution characteristics are digitally encoded, and the encoding results are combined to generate a verification data unit containing the reference occurrence frequency, the standard time interval, a synchrony coefficient and the delay uncertainty parameter. All verification data units are arranged in chronological order to form a verification benchmark sequence.

[0091] The multi-terminal synchrony repetition pattern refers to the cooperative behavior characteristics of multiple user terminals in a live interactive scene; when the host announces important activities or shows exciting content, a large number of audiences will send similar comments or like operations in a short time, forming obvious time aggregation characteristics; however, since the network delay compensation is an estimation based on the historical average network delay, the actual network delay has volatility, which introduces measurement uncertainty; if the measurement uncertainty is not considered, it will lead to evaluation deviation of the real user behavior synchrony, therefore, the embodiment adopts a probability model considering measurement uncertainty for synchrony distribution modeling.

[0092] Specifically, in combination with the multi-terminal synchrony repetition pattern, the same interactive data generated by multiple terminals in the same time window is subjected to synchrony distribution modeling to obtain synchrony distribution characteristics, which means that for each terminal, in addition to recording the historical average network delay, the standard deviation of the network delay is also counted; the network delay standard deviation is obtained by taking the square root of the average of the sum of squares of deviations of historical multiple measurement values and the average value, and the standard deviation reflects the fluctuation degree of the terminal network delay.

[0093] Based on the standard deviation of the network delay of each terminal, a delay uncertainty parameter is constructed; the delay uncertainty parameter is obtained by taking the arithmetic mean of the network delay standard deviations of all terminals that generate repeated data, and this parameter reflects the uncertainty level of the overall network delay measurement.

[0094] In the synchronization distribution modeling, the influence of measurement uncertainty is considered by using a probability model. For repeated data in each time window, the standard deviation of the relative response delay data of each terminal is first calculated as the observed dispersion, which contains the combined influence of the real user behavior dispersion and the measurement uncertainty.

[0095] Based on the probability model, it is assumed that the real user behavior dispersion and the measurement uncertainty are independent of each other, and there is a relationship between the two and the observed dispersion as follows:

[0096]

[0097] wherein, represents the observed relative response delay standard deviation; represents the real user behavior dispersion; represents the delay uncertainty parameter.

[0098] The estimated value of the real user behavior dispersion can be derived from the above relationship, that is, by taking the square root of the square of the observed dispersion minus the square of the delay uncertainty parameter; when the observed dispersion is less than the delay uncertainty parameter, the real dispersion is zero, indicating that the observed dispersion can be completely explained by the measurement uncertainty. By performing an arithmetic average of the real user behavior dispersion of multiple time windows, the synchronization distribution characteristics of the repeated data are obtained.

[0099] It should be noted that, in order to facilitate coding and transmission, the synchronization distribution characteristics are normalized to obtain a synchronization coefficient; specifically, the synchronization distribution characteristics are subtracted from the standard time interval and divided by the standard time interval, and then the difference is subtracted from the ratio to obtain the synchronization coefficient; the synchronization coefficient has a value range of zero to one, and the closer the synchronization coefficient is to one, the stronger the multi-terminal time aggregation characteristics of the corresponding repeated data.

[0100] Further, the specific structure of the verification data unit includes a data type identifier for identifying the specific data type corresponding to the verification data unit; a reference occurrence frequency for recording the reference occurrence frequency ; a standard time interval for recording the standard time interval ; a synchronization coefficient for recording the synchronization coefficient; a delay uncertainty parameter for recording the delay uncertainty parameter; and a time stamp field for recording the time reference point corresponding to the verification data unit.

[0101] In the embodiment, arranging all the check data units in time sequence to form a check reference sequence refers to sorting the check data units of different time periods and different data types according to the timestamp field to form a time-ordered check reference sequence; in the check reference sequence, an 8-bit synchronization identification field (such as “11110000”) is added before each check data unit, facilitating identification and extraction by the receiving end. The check reference sequence not only contains the timing statistical characteristics of repeated data, but also embeds the uncertainty information of network delay, so that the receiving end can distinguish between network layer disturbance and broadcast layer disturbance when performing transmission quality evaluation, and accurately locate the quality problem in the frequency modulation transmission process.

[0102] To solve the fundamental obstacle that the sending end cannot obtain feedback information from the receiving end for transmission quality evaluation in the one-way transmission mode of frequency modulation broadcast, and the influence of network delay disturbance on the establishment of timing distribution mode, a network delay compensation mechanism is introduced to strip the delay disturbance introduced by the network transmission layer before establishing the timing distribution mode of repeated data, so as to avoid misjudging the network layer delay as user behavior difference. At the same time, a probability model considering measurement uncertainty is used to model the synchronization distribution, a delay uncertainty parameter is constructed based on the standard deviation of the network delay of each terminal, and is embedded into the check reference sequence, so that the receiving end can decompose the observed timing deviation into network layer disturbance and broadcast layer disturbance when performing transmission quality evaluation, accurately identify the quality problem introduced by the frequency modulation transmission process, and avoid misjudging the inherent fluctuation of the network layer as a decrease in frequency modulation transmission quality. By using a sliding time window to statistically analyze the compensated relative response delay data, the reference occurrence frequency and standard time interval are extracted, and combined with the synchronization distribution characteristics considering measurement uncertainty, a timing distribution mode that can truly reflect the collective behavior of users under normal transmission conditions is formed; the timing rule is encoded into a check reference sequence containing the reference occurrence frequency, standard time interval, synchronization coefficient, and delay uncertainty parameter, providing a more accurate and reliable self-checking monitoring capability for frequency modulation broadcast transmission, and improving the accuracy of transmission quality evaluation and the effectiveness of resource allocation optimization.

[0103] S3: embedding the check reference sequence into the data stream encoded according to the transmission resource allocation matrix and the multi-end live interaction data, forming a composite transmission data stream, and frequency modulating the composite transmission data stream according to the transmission resource allocation matrix to generate a frequency modulation transmission signal.

[0104] Further, after completing the generation of the verification reference sequence, it is necessary to integrate the multi-end live interaction data with the verification reference sequence and convert it into a frequency modulation broadcast signal. The traditional frequency modulation broadcast mainly transmits a single type of audio signal, and a unified coding modulation scheme can meet the demand. However, the present application needs to transmit various types and priorities of interaction data and the verification reference sequence in the same frequency channel at the same time. By differentiating the coding of different types of interaction data according to the transmission resource allocation matrix, the high-priority data can obtain stronger anti-interference ability. Then, the verification reference sequence is dispersedly embedded as control information in the coded data stream to ensure the reliability in the transmission process. Finally, in the frequency modulation stage, the coding priority is mapped to the frequency band allocation priority, so that the key data occupies the frequency band with better transmission quality.

[0105] The forming of the composite transmission data stream includes differentiating the coding of the multi-end live interaction data according to the coding priority value of each data type in the transmission resource allocation matrix. The verification reference sequence is dispersedly embedded into the coded data stream according to the preset insertion rule to form a composite data stream.

[0106] The differentiating coding is reflected in two dimensions of coding bit number and error correction strength. The data type with high coding priority value, such as gift rewards, is allocated more bits for information expression and uses strong error correction coding to increase redundant check bits and ensure transmission reliability. The data type with low coding priority value, such as historical barrage, uses compression coding to reduce bit overhead and uses lightweight error correction coding to reduce coding overhead on the premise of ensuring basic information transmission.

[0107] The verification reference sequence as metadata for transmission quality evaluation directly affects the accuracy of subsequent quality monitoring. The present embodiment adopts a dispersed embedding strategy to divide the verification reference sequence into multiple verification data units, embed them into the coded interaction data stream according to the preset insertion rule, and realize risk dispersion through spatial dispersion. Specifically, the preset insertion rule includes: time alignment principle, the position of each verification data unit corresponds to the data region of the time period indicated by its timestamp; uniform distribution principle, the verification data units are uniformly distributed in the entire data stream to avoid local concentration; and identification distinguishing principle, a specific synchronization identification field is added before each verification data unit to enable the receiving end to accurately identify and extract, thereby forming a composite data stream containing multi-end live interaction data and verification reference sequence.

[0108] For example, a check data unit is inserted after encoding one hundred interaction data units; meanwhile, the check data unit is inserted after the interaction data corresponding to the time stamp according to the time alignment principle; a fixed synchronization identification field (such as an eight-bit sequence "11110000") is added before each check data unit to facilitate the receiving end to identify and locate; for example, when one hundred interaction data such as comments and likes are encoded, the corresponding time stamp is checked, and the check data unit closest to the time stamp is selected and inserted after adding the synchronization identification field; the subsequent interaction data is continuously encoded, and the above insertion process is repeated every one hundred data units to form a composite data stream in which the multi-end live interaction data and the check reference sequence are alternately distributed.

[0109] Further, frequency modulation of the composite transmission data stream according to the transmission resource allocation matrix to generate a frequency modulation transmission signal includes serial-parallel conversion of the composite data stream to convert the serial data stream into a parallel data format suitable for frequency modulation, wherein the purpose of the serial-parallel conversion is to decompose the high-speed serial data stream into multiple low-speed parallel data streams to realize multi-carrier parallel transmission. The frequency modulation technology is used to modulate different bit positions of the parallel data onto different carrier frequencies. In this embodiment, the frequency modulation technology uses a multi-carrier modulation method to assign independent sub-carrier frequencies to each parallel data. According to the encoding priority values in the transmission resource allocation matrix, different frequency band carriers are assigned to the composite data stream to generate a frequency modulation transmission signal containing multi-end live interaction data and check reference sequences.

[0110] Specifically, according to the encoding priority values in the transmission resource allocation matrix, different frequency band carriers are assigned to the composite data stream, which means that during carrier allocation, according to the data type and encoding priority value mainly carried in each parallel data stream, high-priority data is preferentially assigned to low-load sub-carriers. For example, parallel data streams carrying gift reward data and commodity on-off notifications are assigned to low-load sub-carriers with the best channel quality; parallel data streams carrying ordinary comments and like operations are assigned to medium-load sub-carriers; and parallel data streams carrying historical bullet screen and user access notifications are assigned to high-load sub-carriers.

[0111] In view of the problems of lack of differentiated processing mechanism and optimization spectrum utilization strategy in the process of converting network interactive data into FM broadcast signal, the efficient broadcast conversion of network data is realized through forming composite transmission data stream and multi-carrier FM modulation processing; different types of interactive data are differentiated coded according to the transmission resource allocation matrix, more bits are allocated to high-priority data for information expression and strong error correction coding is used to increase redundant check bits, and compression coding is used for low-priority data to reduce bit overhead and use lightweight error correction, so as to ensure that all types of data can obtain transmission guarantee matched with importance; the check reference sequence adopts a dispersed embedding strategy, and is embedded into the coded data stream according to the insertion rules of time alignment, uniform distribution and identification distinction, and the influence risk of local transmission failure on overall check function is reduced through spatial dispersion; not only the transmission quality of key interactive data is ensured, but also the overall utilization efficiency of FM broadcast spectrum is maximized.

[0112] S4: extracting the check reference sequence from the FM transmission signal at the receiving end, and comparing and analyzing the actual distribution of the repeated data received, and adjusting the transmission resource allocation matrix according to the comparison result.

[0113] In the conventional network data transmission, the sending end can adjust the transmission parameters through the real-time feedback of the receiving end, such as reducing the sending rate, enhancing the coding strength or switching the transmission path, forming a closed-loop transmission quality control; however, the one-way transmission characteristic of FM broadcast cuts off this feedback link, and the sending end cannot know the actual receiving condition of the receiving end. In the prior art, the FM broadcast system usually adopts fixed transmission parameter configuration, which cannot be optimized according to the actual channel condition, resulting in the decrease of transmission quality when the channel condition changes.

[0114] Therefore, the embodiment indirectly evaluates the transmission quality by comparing and analyzing the check reference sequence with the actual received data distribution at the receiving end; however, directly comparing the actual distribution with the check reference may misjudge the inherent delay disturbance of the network layer as the FM transmission quality problem, resulting in the wrong adjustment of the resource allocation strategy; in order to solve this problem, the embodiment introduces the Bayesian inference method, establishes a joint probability model based on the delay uncertainty parameters embedded in the check reference sequence, comprehensively considers the joint influence of network layer disturbance and broadcast layer disturbance, accurately identifies the synchronization deviation caused by the decrease of FM broadcast transmission quality, and feeds back the evaluation result to the sending end for accurate adjustment of the transmission resource allocation matrix, so as to realize the closed-loop optimization capability similar to network transmission on the premise of maintaining the advantage of one-way broadcast transmission.

[0115] Specifically, the adjusting the transmission resource allocation matrix according to the comparison result comprises demodulating and decoding the frequency modulation transmission signal at the receiving end, separating out the original multi-terminal live interactive data and the check reference sequence. Repeated data is extracted from the original multi-terminal live interactive data, and the occurrence frequency, time interval and synchronization distribution characteristics of the multi-terminal of the repeated data are counted to form actual distribution data. The actual distribution data is compared with the check reference sequence, a joint probability model is established based on the delay uncertainty parameter in the check reference sequence by using Bayesian inference method, the joint influence of network layer disturbance and broadcast layer disturbance is comprehensively considered, the synchronization deviation caused by the decline of frequency modulation broadcast transmission quality is obtained, and the transmission resource allocation matrix is adjusted according to the synchronization deviation.

[0116] It should be noted that the demodulation and decoding uses inverse processing corresponding to the frequency modulation at the sending end, converts the analog frequency signal into a digital bit stream, and then separates out the multi-terminal live interactive data and the check reference sequence according to the preset data format and synchronization identification field.

[0117] The repeated data is extracted from the original multi-terminal live interactive data, and the occurrence frequency, time interval and synchronization distribution characteristics of the multi-terminal of the repeated data are counted to form actual distribution data, which means that the receiving end performs content similarity detection on the received multi-terminal live interactive data, identifies repeated data, then counts the actual occurrence frequency and adjacent occurrence time interval of each repeated data, and calculates the time distribution concentration degree of the multi-terminal generating the same repeated data to obtain the actual occurrence frequency, time interval and synchronization characteristics.

[0118] Further, when comparing the actual distribution data with the check reference sequence, directly calculating the deviation may misjudge the inherent delay disturbance of the network layer as a frequency modulation transmission quality problem; for example, even if the frequency modulation broadcast transmission quality is good, due to the fluctuation of the user terminal network delay, the synchronization of the repeated data observed by the receiving end will also have a certain dispersion degree, and if this factor is not considered, it will lead to incorrect evaluation of the transmission quality; therefore, the Bayesian inference method is used in this embodiment, a joint probability model is established based on the delay uncertainty parameter embedded in the check reference sequence, and the joint influence of network layer disturbance and broadcast layer disturbance is comprehensively considered.

[0119] The purpose of the Bayesian inference method is to infer the synchronization deviation caused by the decline of frequency modulation broadcast transmission quality under the condition that the prior information (delay uncertainty parameter) of the network layer delay uncertainty is known, by observing the actual synchronization dispersion degree.

[0120] Specifically, the joint probability model assumes that the observed actual synchronization dispersion is the result of the joint action of network layer disturbance and broadcast layer disturbance, which are independent of each other; the network layer disturbance is characterized by a delay uncertainty parameter, reflecting the inherent fluctuations of the network delay of the user terminal; the broadcast layer disturbance is characterized by the additional dispersion introduced by the frequency modulation transmission, reflecting the problems such as data loss, delay jitter and selective fading in the frequency modulation broadcast transmission process.

[0121] Based on the independence assumption, the square of the observed actual synchronization dispersion is equal to the sum of the square of the network layer disturbance and the square of the broadcast layer disturbance, and the calculation formula is as follows:

[0122]

[0123] wherein, represents the actual synchronization dispersion obtained by the receiving end statistics; represents the delay uncertainty parameter; represents the synchronization deviation caused by the decline of the quality of frequency modulation broadcast transmission.

[0124] The synchronization deviation caused by the decline of the quality of frequency modulation broadcast transmission, that is, the square root of the square of the actual synchronization dispersion minus the square of the delay uncertainty parameter; when the actual synchronization dispersion is less than or equal to the delay uncertainty parameter, the broadcast layer disturbance is zero, indicating that the observed dispersion can be completely explained by the inherent fluctuations of the network layer, and the quality of frequency modulation transmission is good; when the actual synchronization dispersion is greater than the delay uncertainty parameter, the broadcast layer disturbance is positive, indicating that the frequency modulation transmission process introduces additional synchronization destruction, and the transmission resource allocation matrix needs to be adjusted.

[0125] Further, the Bayesian inference method is used to optimize the above estimation, wherein in the Bayesian framework, the broadcast layer disturbance is regarded as a parameter to be estimated, the delay uncertainty parameter provides a prior distribution, and the actually observed synchronization dispersion provides a likelihood function, and the posterior distribution is calculated by the Bayesian formula to obtain the optimal estimation value and confidence interval of the broadcast layer disturbance.

[0126] The Bayesian inference comprises, first, constructing a prior distribution based on the delay uncertainty parameter; assuming that the broadcast layer disturbance should be close to zero under normal transmission conditions, a normal distribution centered on zero with a variance related to the delay uncertainty parameter is adopted as the prior distribution. Second, constructing a likelihood function based on the actually observed synchronization dispersion; assuming that the observation noise obeys a normal distribution, the likelihood function describes the probability of observing the current actual synchronization dispersion under a given broadcast layer disturbance parameter. Then, calculating the posterior distribution according to the Bayesian formula; the posterior distribution integrates the prior information and the observation information, and gives the probability distribution of the broadcast layer disturbance parameter after observing the actual synchronization dispersion. Finally, extracting the optimal estimation value from the posterior distribution, usually adopting the maximum posterior estimation or the posterior expectation as the estimation value of the broadcast layer disturbance, i.e. the synchronization deviation caused by the decline of the frequency modulation broadcast transmission quality.

[0127] After obtaining the synchronization deviation caused by the decline of the frequency modulation broadcast transmission quality, it is also necessary to compare the actual occurrence frequency with the reference occurrence frequency, the actual average time interval with the standard time interval, and calculate the deviation rate of the occurrence frequency and the correlation coefficient of the time interval. Among them, the deviation rate of the occurrence frequency is calculated by the difference between the actual occurrence frequency and the reference occurrence frequency, and the proportion of the reference value; the greater the deviation rate, the more serious the data loss. The correlation coefficient of the time interval is evaluated by the deviation degree of the actual average time interval and the standard time interval; specifically, the absolute difference between the two is calculated, and the proportion of the standard time interval is obtained, and the correlation coefficient is obtained by subtracting the proportion from one; the lower the correlation coefficient, the more serious the transmission delay jitter.

[0128] Further, adjusting the transmission resource allocation matrix according to the synchronization deviation comprises, when the synchronization deviation caused by the decline of the frequency modulation broadcast transmission quality of a certain data type exceeds a preset threshold, it indicates that the multi-end synchronization of the corresponding data type in the frequency modulation transmission process is seriously damaged, which may be due to the selective fading or interference enhancement of the frequency band to which the data type is allocated; at this time, the frequency band allocation strategy needs to be adjusted, and the data type is re-allocated to a frequency band with better transmission conditions, or the error correction strength is enhanced to improve the anti-interference ability.

[0129] When the deviation rate of the occurrence frequency exceeds a preset threshold, it indicates that the corresponding data type is seriously lost in the transmission process, and the error correction strength needs to be enhanced; specifically, the error correction strength parameter value of the data type in the transmission resource allocation matrix is increased, the proportion of redundant check bits is increased, and the data recovery ability is improved.

[0130] When the correlation coefficient of the time interval is lower than the preset threshold, it indicates that the corresponding data type has transmission delay jitter, and the encoding priority value needs to be improved; specifically, the encoding priority value of the data type in the transmission resource allocation matrix is improved, so that it can obtain better frequency band and more transmission guarantee in resource allocation, and the specific adjustment information is sent back to the sending end through an independent low-bandwidth Internet backhaul channel.

[0131] In view of the problems that the fixed configuration of the transmission parameters of the traditional FM broadcast cannot be optimized and adjusted according to the actual channel conditions and transmission effects, and the direct comparison may misjudge the network layer delay disturbance as the FM transmission quality problem, the embodiment constructs a closed-loop optimization control system of FM broadcast transmission through the receiving end verification analysis and the adjustment mechanism of the transmission resource allocation matrix; a joint probability model is established based on the delay uncertainty parameters embedded in the verification reference sequence by using the Bayesian inference method, the joint influence of the network layer disturbance and the broadcast layer disturbance is considered comprehensively, the synchronization deviation caused by the decrease of the FM broadcast transmission quality is accurately identified, and the inherent fluctuation of the network layer is avoided from being misjudged as the FM transmission quality problem; by comparing and analyzing the actual distribution data with the verification reference sequence, the deviation rate of the occurrence frequency, the correlation coefficient of the time interval and the synchronization deviation are calculated, the specific problem type encountered by the specific data type in the transmission process can be accurately identified, including data loss, delay jitter and multi-end synchronization destruction; the sending end can adjust the related parameters in the transmission resource allocation matrix, including enhancing the error correction strength of the specific data type, improving the encoding priority value or re-optimizing the frequency band resource allocation strategy, so that the FM broadcast can continuously optimize the transmission strategy according to the channel environment change, realize the continuous improvement of the transmission quality, and improve the adaptive ability and reliability of the FM broadcast transmission in the multi-end live interactive scene.

[0132] On the other hand, the embodiment also provides a multi-end live interactive data transmission system based on FM relay, which comprises:

[0133] A live data acquisition module acquires multi-end live interactive data under FM relay, extracts delay tolerance characteristics of different data types, and constructs a transmission resource allocation matrix.

[0134] A verification reference sequence module detects repeated data of the multi-end live interactive data, establishes a time sequence distribution mode of the repeated data in combination with network delay compensation, and generates a verification reference sequence based on the time sequence distribution mode.

[0135] An FM signal conversion module embeds the verification reference sequence into a data stream coded according to the transmission resource allocation matrix and the multi-end live interactive data, forms a composite transmission data stream, and modulates the composite transmission data stream according to the transmission resource allocation matrix to generate an FM transmission signal.

[0136] A sequence comparison module extracts the check reference sequence from the frequency modulation transmission signal at the receiving end, and compares and analyzes the received repeated data actual distribution, and adjusts the transmission resource allocation matrix according to the comparison result.

[0137] If the above functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or parts of the present application that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0138] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered a list of executable instructions for implementing logical functions, and can be specifically embodied in any computer-readable medium for use by an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, apparatus or device and execute the instructions, or in conjunction with these instructions. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by an instruction execution system, apparatus or device, or in conjunction with these instruction execution systems, apparatus or devices.

[0139] More specific examples (non-exhaustive list) of computer-readable medium include the following: electrical connections having one or more wires (electronic devices), portable computer disks (magnetic devices), random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memories), fiber optic devices, and portable compact disk read-only memories (CD ROMs). In addition, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, because the program can be electronically obtained, for example, by optical scanning of the paper or other medium, followed by editing, interpreting or otherwise processing, if necessary, in other suitable ways, to be electronically obtained and then stored in the computer memory.

[0140] It should be understood that portions of the present application can be implemented in hardware, software, firmware, or combinations thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, implementation can be with any or a combination of the following technologies, which are all well-known in the art: a discrete logic circuit having logic gates for implementing logic functions upon an application of data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0141] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A multi-terminal live broadcast interaction data transmission method based on frequency modulation relay, characterized in that, The method comprises the following steps: Obtain multi-end live interactive data under frequency modulation relay, extract delay tolerance characteristics of different data types, and construct a transmission resource allocation matrix; Detect repeated data of the multi-end live interactive data, and establish a time sequence distribution pattern of repeated data in combination with network delay compensation, generate a check reference sequence based on the time sequence distribution pattern; Embed the check reference sequence into a data stream coded according to the transmission resource allocation matrix and the multi-end live interactive data, form a composite transmission data stream, and modulate the composite transmission data stream according to the transmission resource allocation matrix to generate a frequency modulation transmission signal; Extract the check reference sequence from the frequency modulation transmission signal at the receiving end, and compare and analyze the actual distribution of received repeated data, and adjust the transmission resource allocation matrix according to the comparison result; Extracting delay tolerance characteristics of different data types in the multi-end live interactive data comprises using a data type identifier to identify the data categories of the multi-end live interactive data; Construct a user behavior response statistical model for the data types, and record the time interval from data sending to the observable interactive response of the user; Analyze the change trend of the user response rate with the increase of the time interval based on the time interval, determine the delay time value corresponding to the time when the response rate decreases to the baseline response rate threshold, and normalize the delay time value to obtain a delay tolerance coefficient; The construction of the transmission resource allocation matrix comprises converting the delay tolerance coefficient into an encoding priority value through a linear inverse proportional function; Use the encoding priority value as the dominant factor of encoding resource allocation, establish a mapping relationship between the encoding priority value, the encoding bit number and the error correction strength, and allocate corresponding encoding priority value, encoding bit number and error correction strength to each data type; Construct a transmission resource allocation matrix based on the encoding priority value, the encoding bit number and the error correction strength of each data type.

2. The method of claim 1, wherein the method is characterized by: The mapping relationship between the encoding priority value, the encoding bit number and the error correction strength comprises load evaluation of available frequency bands based on the spectrum resource capacity of frequency modulation broadcast, calculation of frequency band load coefficient, and division of available frequency bands into low load frequency bands, medium load frequency bands and high load frequency bands according to the frequency band load coefficient; Arrange the data types to be allocated in descending order of encoding priority value, and when there are multiple data types with the same encoding priority value, calculate the ratio of delay tolerance coefficient to encoding priority value as the secondary sorting basis, and arrange the data types in ascending order of the ratio of delay tolerance coefficient to encoding priority value; Allocate frequency band resources to each data type in order, preferentially allocate the data types at the top of the order to the low load frequency bands, and when the capacity of the low load frequency bands is insufficient, allocate the data types to the medium load frequency bands and the high load frequency bands in turn. For the data type allocated to the high load frequency band, the adjusted encoding bit number is obtained by combining the basic encoding bit number of the corresponding data type and the load compensation coefficient; the redundancy code increase amount is calculated according to the reduction amount of the encoding bit number, and the adjusted error correction strength is obtained by combining the redundancy code increase amount and the basic error correction strength.

3. The multi-terminal live interactive data transmission method based on FM repeater as described in claim 2, characterized in that: The repeated data detection is performed on the multi-terminal live interactive data, and the time sequence distribution mode of the repeated data is established in combination with the network delay compensation, including: performing content similarity detection on the multi-terminal live interactive data to identify the repeated data generated by the multi-terminal; The source terminal and the historical average network delay of the repeated data are obtained, and the network delay compensation is performed on the repeated data to obtain the user response time data; The event trigger time point causing the repeated data is identified, the difference between the user response time of each terminal and the event trigger time point is calculated to obtain the relative response delay data of each terminal; The relative response delay data is analyzed by using a time window, the reference occurrence frequency and the standard time interval of the repeated data are calculated, and the time sequence distribution mode of the repeated data is formed.

4. The method of claim 3, wherein the method further comprises: transmitting the interactive data to the first and second end users via the first and second frequency channels, respectively. The generation of the check reference sequence based on the time sequence distribution mode includes: obtaining the reference occurrence frequency and the standard time interval in the time sequence distribution mode; In combination with the multi-terminal synchronization repetition mode, the same interactive data generated by the multiple terminals in the same time window is subjected to synchronization distribution modeling to obtain synchronization distribution characteristics, wherein the synchronization distribution modeling adopts a probability model considering measurement uncertainty, and a delay uncertainty parameter is constructed based on the standard deviation of the network delay of each terminal; The reference occurrence frequency, the standard time interval and the synchronization distribution characteristics are digitally encoded, and the encoding results are combined to generate a check data unit containing the reference occurrence frequency, the standard time interval, the synchronization coefficient and the delay uncertainty parameter; All check data units are arranged in time sequence to form a check reference sequence.

5. The method of claim 4, wherein the method further comprises: transmitting the interactive data to the first and second end users via the first and second frequency channels, respectively. The forming of the composite transmission data stream includes: differentiating the encoding of the multi-terminal live interactive data according to the encoding priority value of each data type in the transmission resource allocation matrix; The check reference sequence is dispersedly embedded into the encoded data stream according to a preset insertion rule to form a composite data stream.

6. The method of claim 5, wherein the method further comprises: The frequency modulation and modulation of the composite transmission data stream according to the transmission resource allocation matrix includes, The serial-parallel conversion of the composite data stream converts the serial data stream into a parallel data format suitable for frequency modulation and modulation; The frequency modulation and modulation technology modulates different bit positions of the parallel data to different carrier frequencies; According to the encoding priority value in the transmission resource allocation matrix, the composite data stream is allocated with carriers of different frequency bands to generate a frequency modulation transmission signal containing multi-terminal live interactive data and a check reference sequence.

7. The method of claim 6, wherein the method further comprises: transmitting the interactive data from the first and second end-points to the intermediate server; and transmitting the interactive data from the intermediate server to the third end-point. The adjustment of the transmission resource allocation matrix according to the comparison result includes: demodulating and decoding the frequency modulation transmission signal at the receiving end to separate the original multi-terminal live interactive data and the check reference sequence; The repeated data is extracted from the original multi-terminal live interactive data, the occurrence frequency, the time interval and the synchronization distribution characteristics of the multi-terminal are counted, and actual distribution data is formed; The repeated data is extracted from the original multi-terminal live interactive data, the occurrence frequency, the time interval and the synchronization distribution characteristics of the multi-terminal are counted, and actual distribution data is formed; The actual distribution data is compared with the verification reference sequence, a joint probability model is established based on the delay uncertainty parameter in the verification reference sequence by using Bayesian inference method, the joint influence of network layer disturbance and broadcast layer disturbance is comprehensively considered, a synchronization deviation caused by the transmission quality decrease of the frequency modulation broadcast is obtained, and the transmission resource allocation matrix is adjusted according to the synchronization deviation.

8. A multi-point live interactive data transmission system based on frequency modulation relay, which adopts the method according to any one of claims 1-7, characterized in that: a live data acquisition module acquires multi-point live interactive data under frequency modulation relay, extracts delay tolerance characteristics of different data types, and constructs a transmission resource allocation matrix; a verification reference sequence module detects repeated data of the multi-point live interactive data, establishes a time sequence distribution mode of the repeated data in combination with network delay compensation, generates a verification reference sequence based on the time sequence distribution mode; a frequency modulation signal conversion module embeds the verification reference sequence into a data stream coded according to the transmission resource allocation matrix and the multi-point live interactive data, forms a composite transmission data stream, and generates a frequency modulation transmission signal by frequency modulation modulation of the composite transmission data stream according to the transmission resource allocation matrix; a sequence comparison and analysis module extracts the verification reference sequence from the frequency modulation transmission signal at a receiving end, compares and analyzes the actual distribution of the received repeated data, and adjusts the transmission resource allocation matrix according to the comparison result.

Images