Video buffer control method and system, electronic device and medium

Abstract

The present disclosure provides a video cache control method, system, electronic device and storage medium to solve the problems of frequent dragging of the progress bar, video re-caching leading to waste of traffic, and unstable download speed. The method comprises: using a multi-level cache area for the intelligent terminal, which is divided into a basic cache area and a prediction cache area, corresponding to the cache of the currently played video segment and the cache of the segment to be played when the video is dragged, respectively; dynamically calculating the cache size of the basic cache area of the intelligent cache area according to the video code rate, signal condition and cache time; determining a dynamic adjustment factor according to the viewing habit parameters and network condition parameters, to determine the cache size of the prediction cache area; and calculating the cache size of the prediction cache area of the intelligent cache area according to the video code rate, the dynamic adjustment factor and the future viewing time when the frequent dragging of the video occurs. The present disclosure not only reduces unnecessary cache traffic, but also guarantees a smooth viewing experience and improves user satisfaction.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of communication and data caching, in particular to a video caching control method, a video caching control system, an electronic device and a computer readable storage medium. BACKGROUND

[0002] With the rapid development of mobile communication technology, especially the commercialization of 5G technology, the bandwidth and data transmission speed of mobile Internet services have been greatly improved, and users' demand for high-definition videos and streaming media content has also grown. Mainstream video websites cache videos. There are mainly two caching strategies: complete caching method: before playing the video, the entire video file is cached first, and the playback does not start until the caching is completed. This way can ensure that there is no lag during playback. Fixed-length caching method: set a fixed cache area, and start playing when the amount of cached data reaches a certain time length; during playback, if the amount of cached data is less than the set value, continue caching until the threshold is reached.

[0003] In the existing video streaming caching scheme, there are the following main problems: Frequent dragging of the progress bar leads to waste of traffic: Many users will drag the progress bar when watching videos due to curiosity or to understand the interestingness of the video, which will cause the video to be re-cached, resulting in unnecessary waste of traffic, especially in the 5G network environment, the bandwidth is high, and the caching speed is extremely fast, but if the user only watches a small part of the video, the rest of the data is wasted. This will lead to user complaints, complaints, and a negative attitude towards operators, and even unwillingness to use 5G. In addition, the load and channel quality of the 5G network is an unstable factor, and network fluctuations can cause unstable download speed, thereby affecting the execution of the video caching strategy. SUMMARY

[0004] In order to at least solve the problems of the existing video streaming caching scheme in the prior art, such as frequent dragging of the progress bar, video re-caching leading to waste of traffic, and network fluctuations leading to unstable download speed. The present disclosure provides a video caching control method, a video caching control system, an electronic device and a computer readable storage medium; not only can reduce unnecessary caching traffic, but also can ensure smooth viewing experience, thereby fully utilizing the high bandwidth advantage of 5G network, optimizing video playback performance, and improving user satisfaction.

[0005] In a first aspect, the present disclosure provides a video caching control method, the method comprising:

[0006] A multi-level cache area is used for the intelligent terminal, which is divided into a basic cache area and a prediction cache area, corresponding to the current playing video segment cache and the dragged video segment cache to be played respectively;

[0007] According to the video code rate, signal condition and buffer time, dynamically calculate the size of the basic cache of the intelligent cache area;

[0008] According to the viewing habit parameters and network condition parameters, determine a dynamic adjustment factor for determining the predicted cache size of the predicted cache area;

[0009] According to the video code rate, dynamic adjustment factor and predicted future viewing time when frequent dragging of the video occurs, calculate the size of the predicted cache of the intelligent cache area.

[0010] Further, the dynamically calculating the size of the basic cache of the intelligent cache area according to the video code rate, signal condition and buffer time comprises:

[0011] The size of the basic cache B of the intelligent cache area is calculated by the following formula init :

[0012]

[0013] And f(RSRP, SINR) is calculated by the following formula

[0014]

[0015] Wherein, M is the video code rate, the signal condition is determined by f(RSRP, SINR), f(RSRP, SINR) is a function based on signal quality RSRP (Reference Signal Receiving Power, Reference Signal Receiving Power) and SINR (Signal to Interference plus Noise Ratio, Signal to Interference plus Noise Ratio) calculation, which describes the influence of signal quality on cache size, RSRP thresh And SINR thresh RSRP and SINR are threshold values, β1 and β2 are adjustment coefficients for adjusting the sensitivity of the function to signal quality, and α is a constant coefficient for further adjusting the change of cache size under different signal quality.

[0016] Further, the calculating the size of the predicted cache of the intelligent cache area according to the video code rate, dynamic adjustment factor and predicted future viewing time when frequent dragging of the video occurs, comprises:

[0017] The size of the predicted cache B is calculated by the following formula pred :

[0018]

[0019] Wherein, T predTo predict the future watching time when the frequent dragging video situation occurs, α(Hu, N) is a dynamic adjustment factor, which depends on the user interest factor H u and the network condition parameter N, for determining the predicted buffer cache amount.

[0020] Further, the method further comprises:

[0021] The dynamic adjustment factor α(Hu, N) is calculated by the following formula:

[0022]

[0023] wherein, H u is the user interest factor, indicating the user's interest in the current video, ranging from 0 to 1, and the coefficient δ represents the influence degree of H u on the buffer amount, CQI (Channel Quality Indicator) represents the network quality, CQI max represents the maximum CQI value, and β represents the CQI adjustment coefficient.

[0024] Further, the method further comprises:

[0025] The historical behavior characteristics of the user are collected, and for each type of video, the following user behavior data is collected: dragging frequency F drag , dragging interval T drag , dragging jump distance L drag , and watching completion degree C complete .

[0026] Based on the collected historical behavior characteristic data of the user, the user interest factor H u of the user for the video is predicted by a preset model considering the characteristics of different video types and user behaviors.

[0027] H u = σ(γ1·f type (T)+γ2·f drag (F drag , T drag , L drag )+γ3·f completion (C complete ))

[0028] wherein, σ(x) is a Sigmoid function, which maps the score to the interval [0, 1], γ1, γ2, γ3 are adjustment coefficients, respectively controlling the influence degree of video type, dragging behavior and watching completion degree on the interest score, and the video type factor f type (T) represents the influence of the video type on the user interest.

[0029] dragging behavior factor f drag (F drag , T drag , L drag ): represents the influence of dragging frequency, dragging interval and dragging jump distance on interest, calculated by the following formula:

[0030]

[0031] α1, α2, α3 are adjustment coefficients, controlling the influence of each factor on interest score;

[0032] watching completion degree factor f completion (C complete ): represents the influence of the completeness of user watching video on interest, calculated by the following formula:

[0033]

[0034] C complete is the completion degree of user watching video (between 0 and 1), if the user watches the video completely, C complete = 1.

[0035] Further, the method further comprises:

[0036] adjusting the video download rate V adjust according to the difference between the cache amount and the playing demand; comprising:

[0037] when the cache amount C current reaches the threshold B thresh , the adjusted video download rate V adjust is calculated by the following formula:

[0038]

[0039] wherein, M is the playing code rate of the video, C remaining is the remaining cache amount, C total is the total cache capacity of the cache area, V init is the initial download rate of the terminal given by the base station according to the CQI of the user terminal and the load of the base station.

[0040] Further, the method further comprises:

[0041] when it is detected that the user does not have dragging behavior within a certain time and the cache area starts to cache new video, switching to the basic cache area, sending the event to the base station, so that the base station automatically restores the terminal to the normal download rate.

[0042] In a second aspect, the disclosure provides a video cache control system, the system comprising:

[0043] a cache setting module configured to use a multi-level cache area for the intelligent terminal, and divide the multi-level cache area into a basic cache area and a prediction cache area, which correspond to a currently played video segment cache and a segment cache to be played by dragging the video, respectively;

[0044] a basic cache area cache module configured to dynamically calculate a cache size of the basic cache area of the intelligent cache area according to a video code rate, a signal condition and a cache time;

[0045] a prediction cache area cache module configured to determine a dynamic adjustment factor according to a viewing habit parameter and a network condition parameter, so as to determine a cache size of the prediction cache area; and

[0046] a prediction cache area cache module configured to dynamically calculate a cache size of the prediction cache area of the intelligent cache area according to the video code rate, the dynamic adjustment factor and a future viewing time when the video is frequently dragged.

[0047] In a third aspect, the present disclosure provides an electronic device including a memory and a processor, the memory storing a computer program, and when the processor executes the computer program stored in the memory, the processor executes the video cache control method according to any one of the first aspect.

[0048] In a fourth aspect, the present disclosure provides a computer readable storage medium, the computer readable storage medium storing a computer program, and when the computer program is executed by a processor, the video cache control method according to any one of the first aspect is implemented.

[0049] Advantages:

[0050] The video cache control method, the video cache control system, the electronic device and the storage medium provided by the present disclosure; by setting the basic cache area and the prediction cache area, which correspond to the currently played video segment cache and the segment cache to be played by dragging the video, respectively, using the dynamic adjustment cache strategy, through real-time analysis of the user's viewing behavior, network status and video content characteristics, accurate cache control is realized, the waste of traffic caused by dragging and frequent jumping of the video progress bar and over-caching of the video is reduced, and at the same time, through the feedback of the real-time network condition, the cache rate is optimized to ensure a smooth viewing experience. Not only can unnecessary cache traffic be reduced, but also the high bandwidth advantage of the 5G network can be fully utilized, the video playback performance can be optimized, and user satisfaction can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 A flowchart of a video cache control method provided by the first embodiment of the present disclosure;

[0052] Figure 2 An architecture diagram of a video cache control system provided by the third embodiment of the present disclosure;

[0053] Figure 3 An architecture diagram of an electronic device is provided for Embodiment Four of the present disclosure. DETAILED DESCRIPTION

[0054] For those skilled in the art to better understand the technical solutions of the present disclosure, the present disclosure will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments and drawings described herein are only for the purpose of explaining the present disclosure, but not limiting the present disclosure.

[0055] It should be noted that the terms "first", "second" and the like in the specification and claims of the present disclosure and the above-described drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence; and, without conflict, the embodiments in the present disclosure and the features in the embodiments can be combined with each other at will.

[0056] Wherein, the terms used in the embodiments of the present disclosure are only for the purpose of describing specific embodiments, and are not intended to limit the present disclosure. The singular forms "a", "said" and "the" used in the embodiments of the present disclosure and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0057] In the subsequent description, the suffix such as "module", "component" or "unit" used to represent elements is only for the purpose of facilitating the description of the present disclosure, and has no specific meaning in itself. Therefore, "module", "component" or "unit" can be used mixedly.

[0058] The technical solutions of the present disclosure and how the technical solutions of the present disclosure solve the technical problems existing in the prior art will be described in detail below with specific embodiments. It can be understood that in the embodiments of the present application, the execution subject can execute part or all of the steps in the embodiments of the present application, and these steps or operations are only examples, and the embodiments of the present application can also execute other operations or variations of various operations. In addition, each step can be executed in a different order as presented in the embodiments of the present application, and it is possible that not all operations in the embodiments of the present application are executed. In addition, the following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described again in some embodiments.

[0059] Figure 1 A flowchart of a video cache control method is provided for Embodiment One of the present disclosure, as shown in Figure 1 The method comprises:

[0060] Step S101: A multi-level cache area is used for the intelligent terminal, which is divided into a basic cache area and a prediction cache area, corresponding to the cache of the currently played video segment and the cache of the segment to be played by dragging the video, respectively;

[0061] Step S102: dynamically calculating the cache size of the basic cache area of the intelligent cache area according to the video code rate, signal condition and cache time;

[0062] Step S103: determining a dynamic adjustment factor according to the viewing habit parameter and the network condition parameter, for determining the predicted cache size of the predicted cache area;

[0063] Step S104: calculating the cache size of the predicted cache area of the intelligent cache area according to the video code rate, the dynamic adjustment factor and the future viewing time when the frequent dragging of the video occurs.

[0064] The purpose of the embodiments of the present disclosure is to realize accurate cache rate control by analyzing the viewing behavior, network status and video content characteristics of the user in real time, reduce unnecessary traffic waste, improve the resource utilization rate of the 5G network, and ensure the smooth viewing experience of the user.

[0065] The intelligent terminal mentioned in the present disclosure can be a smart phone (such as an Android phone, an IOS phone, etc.), a tablet computer, a notebook computer and a smart device, etc. Specifically, in the cache process, the intelligent terminal adopts a multi-level cache area, which is divided into a "basic cache area" and a "predicted cache area". They respectively correspond to the currently played video segment and the segment to be played when the video is dragged, so as to realize accurate control, and both are intelligent mixed memory access.

[0066] Basic cache area: in order to ensure that the video plays without lag, it is used to store the video that the user stably watches, and to ensure that the video plays smoothly. Stable watching can be defined as no dragging event occurring within a continuous x time. The size of the basic cache area is dynamically calculated in combination with the real-time network environment:

[0067] Predicted cache area: used to cache the future content that may be needed when the video is frequently dragged. Frequent dragging is defined as continuous dragging n times or more within a short time (such as 15 seconds), for example, n is 2 or 3, etc. It is considered that the user is curious about the video or wants to know whether the video is interesting.

[0068] For the basic cache area, when the user plays an online video through an App on the intelligent terminal, the cache size is determined by acquiring the basic information of the video played on the intelligent terminal, including the video code rate M, the cache time length T and other parameters.

[0069] Video code rate M (unit: Mbps): obtained by adding the audio code rate and the video code rate of the video content, used to calculate the data size corresponding to the video per second.

[0070] Total video duration (T total ): total duration of the video (unit: seconds).

[0071] Buffer time T init (unit: seconds): In order to ensure smooth video playback, the system will pre-cache data according to the set buffer time, usually set to 1-5 minutes. The length of the buffer time T is set according to the strategy of the video service provider.

[0072] Network signal condition (Q): The intelligent terminal can obtain the network signal strength RSRP and quality SINR of the current base station in real time. RSRP is used to measure the received strength of the wireless signal, with a unit of dBm, and the smaller the value, the worse the signal. SINR is used to measure the ratio of signal to noise, with a unit of dB, and the lower the value, the worse the signal. These two parameters can also determine the adjustment of the initial buffer amount of video playback.

[0073] For the prediction buffer area, first, the dynamic adjustment factor is determined according to the viewing habit parameters and network condition parameters, which is used to determine the prediction buffer area buffer amount; then the prediction buffer area buffer amount size of the intelligent buffer area is calculated according to the video code rate, the dynamic adjustment factor and the future viewing time when the video is frequently dragged. Thus, the user's interest score for the video being watched is predicted according to the user's historical behavior habits, and the buffer size is dynamically adjusted according to the interest score, so that the video playback is more smooth, and the waste of traffic is reduced.

[0074] The basic buffer area is used to cache the video segments that are stably watched. The prediction buffer area is used to determine the video to be cached when the user frequently drags the progress bar. When the video is dragged, the video in the basic buffer area can not be emptied. In this way, the multi-level cache strategy (basic buffer area and prediction buffer area) reduces the frequent emptying of the buffer area and allows personalized adjustment according to the user's viewing habits. By setting different buffer areas to meet different content needs, resource utilization efficiency can be improved.

[0075] The embodiments of the present disclosure set the basic buffer area and the prediction buffer area, respectively corresponding to the current played video segment cache and the segment to be played when the video is dragged, use a dynamic adjustment cache strategy, and realize accurate cache control by analyzing the user's viewing behavior, network status, and video content characteristics in real time. Reduce the waste of traffic caused by video progress bar dragging and frequent jumping, and over-caching of video, while optimizing the cache rate through real-time network status feedback to ensure a smooth viewing experience. Not only can unnecessary cache traffic be reduced, but also the high bandwidth advantage of 5G network can be fully utilized to optimize video playback performance and improve user satisfaction.

[0076] Further, the dynamic calculation of the buffer amount size of the basic buffer area of the intelligent buffer area according to the video code rate, the signal condition and the buffer time comprises:

[0077] The buffer amount B of the basic buffer area is calculated by the following formulainit :

[0078]

[0079] f(RSRP, SINR) is calculated by the following formula:

[0080]

[0081] Where M is the video bitrate, the signal condition is determined by f(RSRP, SINR), f(RSRP, SINR) is a function based on signal quality RSRP and SINR calculation, describes the influence of signal quality on the cache size, RSRP thresh and SINR thresh are the thresholds of RSRP and SINR respectively, β1 and β2 are the adjustment coefficients, used to adjust the sensitivity of the function to signal quality, α is a constant coefficient, used to further adjust the change of cache size under different signal quality.

[0082] According to the video bitrate M, the signal condition Q, dynamically calculate the cache size B of the basic cache area of intelligent cache area init . Set a cache time T init , then the cache size B of the basic cache area init The calculation formula is:

[0083]

[0084] Where:

[0085] 8 means 1MBps = 8Mbps,

[0086] f(RSRP, SINR) is a function based on signal quality (RSRP and SINR) calculation, describes the influence of signal quality on the cache size. The worse the RSRP and SINR, the slower the download speed, which may cause the cache speed to be slower than the play speed, the cache data is not enough, resulting in lag, so it is necessary to ensure that the cache size of the basic cache area is increased when the signal quality is poor, and the cache size of the basic cache area is reduced when the signal quality is good. The cache size of the basic cache area can be intelligently adjusted according to different network quality, realizing intelligent cache. The function can be designed as:

[0087]

[0088] Where: RSRP thresh and SINR threshthresholds of RSRP and SINR, when the signal quality is lower than the threshold, the system needs to increase the buffer amount, the threshold can be set according to the actual situation, such as setting according to the video playing requirements. β1 and β2 are adjustment coefficients, used to adjust the sensitivity of the function to the signal quality. α is a constant coefficient, used to further adjust the change of the buffer amount under different signal quality.

[0089] When the network signal quality is good (RSRP and SINR are high), the value of f(RSRP, SINR) is close to 0, and the basic buffer amount is close to the standard value M*Tinit / 8. When the signal quality is poor (RSRP and SINR are low), the value of f(RSRP, SINR) increases, and the basic buffer amount increases accordingly, ensuring that the video can be played smoothly, even in poor network environment.

[0090] For example, in one embodiment:

[0091] Suppose the code rate of a video is M=4Mbps, the video duration is 120 minutes, and the buffer time Tinit=60 seconds (1 minute). Set the RSRP of the network signal to -110dBm, and the SINR to 12dB, and set the threshold and adjustment coefficient as follows:

[0092] RSRP thresh = -100, SINR thresh = 15dB, β1 = 0.05, β2 = 0.1, α = 0.2

[0093] First, calculate f(RSRP, SINR):

[0094]

[0095] f(RSRP, SINR) ≈ 0.377*0.574 = 0.216

[0096] Then, calculate the buffer amount B of the basic buffer area init :

[0097] B init = 4*60 / 8*(1+0.2*0.216) = 30*(1+0.0432) = 30*1.0432 = 31.296MB,

[0098] In this case, the buffer amount of the basic buffer area is about 187.776MB, considering the poor signal quality (RSRP=-110dBm and SINR=12dB), the buffer amount of the basic buffer area has been significantly increased, ensuring that even if the download speed is slow, the video playing will not be stuck.

[0099] Further, the predicted cache size of the intelligent cache area is calculated according to the video code rate, the dynamic adjustment factor and the predicted future viewing time when the frequent dragging event occurs, and the calculation method comprises the following steps:

[0100] The predicted cache size B is calculated by the following formula pred :

[0101]

[0102] Wherein, T pred is the predicted future viewing time when the frequent dragging event occurs, and α(Hu, N) is the dynamic adjustment factor, which depends on the user interest factor H u and the network condition parameter N, and is used for determining the predicted cache size.

[0103] The predicted cache size is determined by the predicted future viewing time when the frequent dragging event occurs and the user interest factor, when the user interest is high, the predicted cache size is larger, and the user can watch the video for a longer time.

[0104] Further, the dynamic adjustment factor is determined according to the viewing habit parameter and the network condition parameter, and the calculation method comprises the following steps:

[0105] The dynamic adjustment factor α(Hu, N) is calculated by the following formula:

[0106]

[0107] Wherein, H u is the user interest factor, which represents the user interest in the current video, and the range is 0 to 1, the coefficient δ represents the influence degree of H u on the cache size, the channel quality indication CQI represents the network quality, CQI max represents the maximum CQI value, and β represents the CQI adjustment coefficient.

[0108] The predicted cache area is based on the viewing habit parameter H u and the network state, and caches the potential required data. When H u is high, the predicted cache size is increased correspondingly, so that the user can watch the content for a longer time smoothly; when H u is low, the cache size is reduced, so that the flow resources are saved. The specific method is as follows:

[0109]

[0110] Wherein, T pred represents how long the user watches after each dragging after the predicted frequent dragging event occurs.

[0111] The expression of α(Hu, N) is as follows

[0112]

[0113] where H u : user interest factor, range from 0 to 1. The higher the value, the higher the user interest, and the buffer size should be increased appropriately. For example, when the user interest is high (H u > 0.7), the intelligent terminal will increase the predicted buffer size B pred to cache more video content in advance, ensuring the smoothness of continuous viewing. When the user interest is low (H u < 0.3), the intelligent terminal will decrease the predicted buffer size to cache only a small amount of data to save traffic resources.

[0114] coefficient δ: controls the degree of influence of H u on the buffer size. The larger the value, the more obvious the influence of user interest on the buffer.

[0115] CQI (Channel Quality Indicator): channel quality indicator, the higher the value, the better the network quality.

[0116] maximum CQI value CQI max : can be set to a certain standard value (e.g., 30), and the lower the CQI, the poorer the channel quality, and the larger the buffer size.

[0117] CQI adjustment coefficient β: controls the degree of influence of channel quality on the buffer size. The larger the value, the more obvious the influence of channel quality on the buffer.

[0118] When the user interest is high (H u approaches 1) or the channel quality is poor (CQI is much lower than CQI max ), α(H u , N) increases, indicating that the buffer size should be increased to cope with potential network fluctuations and high viewing demand. When the user interest is very low and the channel quality is very good, α(H u , N) approaches 1, indicating that there is no need for excessive caching.

[0119] Assumptions:

[0120] user interest H u = 0.7

[0121] current CQI = 15, CQI max = 30

[0122] set δ = 0.3, β = 0.5

[0123] α(H u , N) = 1 + 0.3 * 0.7 + 0.5 * (1 - 15 / 30) = 1.46

[0124] In this case, the cache size of the prediction cache will increase by about 46%, thus ensuring smooth video playback in the case of high interest and average channel quality.

[0125] Further, the method further comprises:

[0126] Collecting the historical behavior characteristics of the user, and collecting the following user behavior data for each type of video: drag frequency F drag , drag interval T drag , drag jump distance L drag , and viewing completion degree C complete .

[0127] Based on the collected historical behavior characteristic data of the user, a preset model is used to predict the user interest factor H u of the user for the video by considering the characteristics of different video types and user behaviors.

[0128] H u = σ (γ1·f type (T) + γ2·f drag (F drag , T drag , L drag ) + γ3·f completion (C complete ))

[0129] Wherein, σ (x) is a Sigmoid function that maps the score to the interval [0, 1], γ1, γ2, γ3 are adjustment coefficients that control the influence degree of video type, drag behavior and viewing completion degree on interest score, respectively, video type factor f type (T) represents the influence of video type on user interest;

[0130] Drag behavior factor f drag (F drag , T drag , L drag ) represents the influence of drag frequency, drag interval and drag jump distance on interest, which is calculated by the following formula:

[0131]

[0132] α1, α2, α3 are adjustment coefficients that control the influence of each factor on interest score;

[0133] Viewing completion degree factor f completion (C complete ) represents the influence of user viewing completeness on interest, which is calculated by the following formula:

[0134]

[0135] C complete is the completion degree of the user watching the video (between 0 and 1), if the user watches the video completely, C complete = 1.

[0136] Predict the user interest factor H of the user to the video u The purpose is to predict the user's interest score for the video being watched according to the user's historical behavior habits, and dynamically adjust the cache size according to the interest score, ultimately making the video play more smooth, and reducing the waste of traffic.

[0137] To effectively predict the user's interest score for the current video according to the user's historical behavior, the following key historical behavior features need to be considered, especially those related to the drag behavior:

[0138] Video type: (documentary, movie, TV series, variety show, short video, animation…) Each type of video has different appeal to users, and users' behavior when watching different types of videos will also be different.

[0139] For each type, the following user behavior data is collected

[0140] Drag frequency (F drag ):

[0141] The frequency of dragging the progress bar during video playback (unit: times / minute). High drag frequency usually indicates that the user is not interested or is quickly browsing the video.

[0142] Drag interval (T drag ):

[0143] The time interval between consecutive drags (unit: seconds). If the drag interval is short, it indicates that the user is quickly browsing the video and may be looking for interesting parts; if the interval is long, it indicates that the user is watching a certain part.

[0144] Drag jump distance (L drag ):

[0145] The jump distance of each drag (unit: seconds). A large drag jump distance indicates that the user is skipping certain parts and usually has low interest in that content; a small jump distance indicates that the user may be interested in certain segments.

[0146] Completion of watching (C complete ):

[0147] The user's past watching completion (video progress watching ratio reaches 80%). A high completion history usually means that the user has a high interest in the video.

[0148] Based on the collected data, a model is designed to predict the user's interest score H for a video by considering the characteristics of different video types and user behaviors u :

[0149] H u = σ(γ1·f type (T)+γ2·f drag (F drag , T drag , L drag )+γ3·f completion (C complete ))

[0150] σ(x) is a Sigmoid function that maps the score to the interval [0, 1].

[0151] γ1, γ2, γ3 are adjustment coefficients that control the influence of video type, dragging behavior, and viewing completion on the interest score, respectively.

[0152] Video type factor f type (T): represents the influence of video type on user interest. Different video types have different influences on user interest. For example, movies usually attract users more, while advertisements may have less appeal. Assign a basic interest factor to each video type, such as in an embodiment, it can be set as:

[0153] Documentary: f type (Documentary) = 0.8

[0154] Movie: f type (Movie) = 1.0

[0155] TV series: f type (TV series) = 0.9

[0156] Variety show: f type (Variety show) = 1.2

[0157] Short video: f type (Short video) = 0.7

[0158] Animation: f type (Animation) = 1.1

[0159] Dragging behavior factor f drag (F drag , T drag , L drag ): represents the influence of dragging frequency, dragging interval, and dragging jump distance on interest.

[0160]

[0161] α1, α2, α3 are adjustment factors, controlling the influence of each factor on the interest score.

[0162] Watching completion factor f completion (C complete ): represents the influence of the completeness of the user's watching of the video on the interest.

[0163]

[0164] C complete is the completion of the user's watching of the video (between 0 and 1), if the user watches the video completely, C complete = 1, indicating a higher interest.

[0165] Collect user behavior feature data as model input, make prediction, get user watching interest parameter H u as the basis for adjusting the caching strategy.

[0166] Further, the method further comprises:

[0167] Adjust the video download rate V adjust according to the difference between the cache amount and the playing demand; comprising:

[0168] When the cache amount C current reaches a threshold B thresh , the adjusted video download rate V adjust is calculated by the following formula:

[0169]

[0170] Wherein, M is the playing code rate of the video, C remaining is the remaining cache amount, C total is the total cache capacity of the cache area, V init is the initial download rate of the terminal given by the base station according to the CQI of the user terminal and the load of the base station.

[0171] The embodiments of the present disclosure also dynamically adjust the download rate according to the real-time network status (such as the download rate provided by the base station) and the cache amount. When the cache amount of the cache area reaches a threshold, the download rate can be reduced to avoid unnecessary waste of traffic, while ensuring that the cache data is sufficient to meet the user's demand when playing the video, avoiding lag.

[0172] Specific process: when starting caching, the terminal will send relevant information to the base station to inform the base station of the caching event. At the same time, the terminal will also send the measured channel quality (CQI) to the base station. The base station will determine the initial download rate of the terminal according to the CQI (channel quality indication) of the user terminal and the load of the base station, assuming that the initial rate given by the base station is V init . Note that V initThe value is determined by the base station according to its load, CQI, etc. The value will change according to the load, CQI, etc.

[0173] At this speed, the video cached in the cache reaches a certain threshold B thresh , indicating that the current cache amount is sufficient to meet the demand of video playback, that is, the download speed can be reduced.

[0174] The threshold can be set according to the clarity of the video content. For example, if the video clarity is high, the cache threshold should be set larger to ensure sufficient pre-caching during video playback. Specifically:

[0175] 1. The clarity of the video is divided into, for example, 360p, 480p, 720p, 1080p, 2K, 4K, etc. These numbers represent the resolution of the video, that is, the number of pixels of the video. Generally, the more the number of pixels, the higher the video quality. The code rate M of different clarity videos is also different.

[0176] The cache threshold B thresh is set according to experience: through experience and the association of video types, we can roughly set a preliminary range of the cache threshold according to the code rate M and T (if using the basic cache area, T=T init , if using the predictive cache area, T=T pred ). According to different video types, the cache threshold can be adjusted by the following proportion:

[0177] 360p: The initial threshold of the cache can be set to 10%-20% of the total duration T, that is:

[0178] B thresh =M×T×factor, factor=0.1

[0179] For example, for a 360p video with a T of 3 minutes and a code rate of 1 Mbps, the cache threshold is about:

[0180] B thresh =1×180×0.1=18MB

[0181] 480p: The initial threshold of the cache can be set to 20% of the total duration T:

[0182] B thresh =M×T×factor, factor=0.2

[0183] 720p: The initial threshold of the cache can be set to 30% of the total duration T:

[0184] B thresh =M×T×factor, factor=0.3

[0185] 1080p: The initial threshold for general caching can be set to 40% of Ttotal duration:

[0186] 2K: The initial threshold for general caching can be set to 50% of Ttotal duration:

[0187] 4K video: Due to the high code rate of high-definition and 4K video, more cache is usually needed to avoid stuttering. For 4K video, the cache threshold can be set to 60% of Tvideo duration.

[0188] When the cache amount C current reaches the threshold B thresh , the system will adjust the download rate V adjust according to the difference between the cache amount and the playback demand. The system will gradually reduce the download rate. Then how to set V adjust , specifically:

[0189] Assume the remaining cache amount is C remaining , the total cache capacity of the cache area is C total , and the current cache amount is C current .

[0190] The proportion of the remaining cache amount to the total cache capacity is:

[0191]

[0192] According to this proportion, the download rate V adjust needs to be reduced proportionally. To ensure that the download rate is adjusted proportionally to the remaining cache amount, but not lower than the playback code rate M of the video, we can design the following adjustment formula:

[0193]

[0194] The download rate will be reduced proportionally according to the proportion of the remaining cache amount, but will not be lower than the playback code rate M of the video, ensuring that even if the cache space is small, the download rate can meet the video playback demand.

[0195] Application example:

[0196] Assume:

[0197] The initial download rate V init = 50 Mbps.

[0198] The current cache amount C current = 2000 MB.

[0199] The total cache capacity C total = 3000 MB.

[0200] Video code rate M = 8 Mbps (i.e. the download rate cannot be lower than 8 Mbps).

[0201] Remaining buffer amount C remaining = C total - C current = 3000 - 2000 = 1000

[0202] C remaining / C total = 0.333

[0203] Adjusted download rate: V adjust = max (50 x 0.3333, 8), so V adjust is 16.65 Mbps

[0204] In this case, the remaining buffer amount accounts for 33.33%, and the download rate is adjusted to about 16.67 Mbps. Assuming that the buffer amount continues to decrease, the current buffer amount C current = 2900 MB, at this time the remaining buffer amount is C remaining = 100 MB. C remaining / C total = 0.0333

[0205] V adjust = max (50 x 0.0333, 8), so V adjust is 8 Mbps

[0206] Further, the method further comprises:

[0207] When it is detected that the user has not performed the dragging behavior for a certain period of time and the buffer area starts to cache new videos, the basic buffer area is switched to, and the event is sent to the base station, so that the base station automatically restores the terminal to the normal download rate.

[0208] The embodiment of the disclosure further designs an automatic speed limit recovery algorithm. When it is detected that the user has not performed the dragging behavior for a long time (a dragging detection threshold J seconds can be set, which can be 30 seconds, 40 seconds or other values) and the buffer area starts to cache new videos, the event is sent to the base station, and the base station is automatically restored to the normal rate.

[0209] Suppose the time when the speed limit starts is t0, if Δt ≥ J, the original rate V restore is restored: V restore = V init , where Δt = t - t0.

[0210] The whole process is as follows: the user opens the app to watch videos, uses the basic buffer area for caching, and at the same time calculates the size of the basic buffer area according to the basic buffer area capacity algorithm, and at the same time the download speed is downloaded at the normal speed provided by the base station as Vinit , the cache amount of the basic cache area reaches a threshold B thresh , the event is sent to the base station, and the base station starts to limit the speed of the user terminal. The limited speed is:

[0211]

[0212] If the dragging is once, the video is cached in the manner of the basic cache area. When the user drags the video twice within 15 seconds, the system considers that a frequent dragging event occurs, the video is cached in the prediction cache area, and the size of the prediction cache area is calculated according to the prediction cache area capacity algorithm. Meanwhile, the download speed is Vinit according to the normal speed provided by the base station, and when the cache amount of the prediction cache area reaches a threshold B thresh , the event is sent to the base station, and the base station starts to limit the speed of the user terminal. The limited speed is:

[0213]

[0214] When the frequent dragging event occurs, it is detected that the user does not drag for a long time (a dragging detection threshold J seconds is set), the basic cache area is switched to, and the event is sent to the base station to start to cache a new video segment, and the base station automatically restores to the normal speed V restore = V init .

[0215] The embodiments of the disclosure reduce the waste of traffic caused by video progress bar dragging and frequent jumping and over-caching of videos by dynamically adjusting the caching strategy and intelligently adjusting the speed, and optimize the caching speed through real-time feedback of network conditions, so that the smooth viewing experience is ensured. The method has strong flexibility and adaptability, can not only reduce unnecessary caching traffic, but also fully utilize the high bandwidth advantage of the 5G network, optimize the video playing performance, and improve the user satisfaction.

[0216] The second embodiment of the disclosure also provides a video caching control system, which realizes accurate caching speed control by analyzing the viewing behavior of the user, the network state and the video content characteristics in real time, reduces unnecessary traffic waste, improves the resource utilization rate of the 5G network, and ensures the smooth viewing experience of the user. The system consists of four modules:

[0217] 1. Video information acquisition and intelligent cache area setting module

[0218] In the caching process, the module adopts a multi-level cache area for the intelligent terminal, which is divided into a "basic cache area" and a "prediction cache area". The two correspond to the currently played video segment and the cached segment of the dragged video to be played, so as to realize accurate control, and both are intelligent mixed access.

[0219] 2. User video interest degree prediction analysis module

[0220] The module aims to predict the user's interest score for the video being watched according to the user's historical behavior habits, and dynamically adjust the cache size according to the interest score, ultimately making the video play more smooth and reducing the waste of traffic.

[0221] 3. Download rate adjustment module

[0222] Main idea: This module dynamically adjusts the download rate according to real-time network conditions (such as the download rate provided by the base station) and the cache amount. When the cache amount in the cache area reaches a threshold, the download rate can be reduced to avoid unnecessary traffic waste, while ensuring that the cache data is sufficient to meet the user's needs when playing the video, avoiding lag.

[0223] 4. Automatic speed limit recovery module

[0224] Description: An automatic speed limit recovery algorithm is designed. When the module detects that the user has not performed a drag behavior for a long time (set a drag detection threshold J seconds) and the cache area starts to cache new videos, the event is sent to the base station, and the base station automatically recovers to the normal rate.

[0225] The embodiments of the present disclosure can reduce unnecessary traffic waste, improve the resource utilization rate of the 5G network, ensure the user's smooth viewing experience, more scientifically and efficiently manage and optimize the mobile communication network, effectively cope with network challenges, reduce user complaints, and improve overall operational efficiency and service quality.

[0226] Embodiment three of the present disclosure also provides a video cache control system, as shown in Figure 2 The system comprises:

[0227] A cache setting module 11 is configured to use a multi-level cache area for the intelligent terminal, which is divided into a basic cache area and a prediction cache area, corresponding to the cache of the currently played video segment and the cache of the segment to be played by the dragged video, respectively;

[0228] A basic cache area cache module 12 is configured to dynamically calculate the cache size of the basic cache area of the intelligent cache area according to the video code rate, signal conditions and cache time;

[0229] A prediction cache area cache module 13 is configured to determine a dynamic adjustment factor according to the viewing habit parameters and the network condition parameters, to determine the cache size of the prediction cache area; and

[0230] According to the video code rate, the dynamic adjustment factor and the future viewing time when the frequent video dragging situation occurs, the cache size of the prediction cache area of the intelligent cache area is calculated.

[0231] Further, the base cache area caching module 12 is specifically configured as:

[0232] The cache amount B of the base cache area is calculated by the following formula init :

[0233]

[0234] And f(RSRP, SINR) is calculated by the following formula:

[0235]

[0236] Wherein, M is the video code rate, the signal condition is determined by f(RSRP, SINR), f(RSRP, SINR) is a function based on signal quality RSRP and SINR, which describes the influence of signal quality on the cache amount, RSRP thresh And SINR thresh are the threshold values of RSRP and SINR, β1 and β2 are adjustment coefficients, used to adjust the sensitivity of the function to the signal quality, and α is a constant coefficient, used to further adjust the change of the cache amount under different signal quality.

[0237] Further, the prediction cache area caching module 13 is specifically configured as:

[0238] The prediction cache area size B is calculated by the following formula pred :

[0239]

[0240] Wherein, T pred is the future viewing time when the frequent dragging video situation occurs, and α(Hu, N) is a dynamic adjustment factor, which depends on the user interest factor H u and the network condition parameter N, used to determine the prediction cache area cache amount.

[0241] Further, the prediction cache area caching module 13 is specifically configured as:

[0242] The dynamic adjustment factor α(Hu, N) is calculated by the following formula:

[0243]

[0244] Wherein, H u is the user interest factor, indicating the user's interest in the current video, ranging from 0 to 1, and the coefficient δ indicates the influence degree of H u on the cache amount, the channel quality indication CQI represents the network quality, CQI max represents the maximum CQI value, and β represents the CQI adjustment coefficient.

[0245] Further, the system further comprises a user interest analysis module 14;

[0246] The user interest analysis module 14 is configured to collect the historical behavior characteristics of the user, and for each type of video, collect the following user behavior data: drag frequency F drag , drag interval T drag , drag jump distance L drag , viewing completion degree C complete ; and

[0247] Based on the collected historical behavior characteristic data of the user, through a preset model, considering the characteristics of different video types and user behavior, the user interest factor H u of the user to the video is predicted:

[0248] H u = σ (γ1·f type (T) + γ2·f drag (F drag , T drag , L drag ) + γ3·f completion (C complete ))

[0249] Wherein, σ (x) is a Sigmoid function, which maps the score to the interval [0, 1], γ1, γ2, γ3 are adjustment coefficients, which respectively control the influence degree of video type, drag behavior and viewing completion degree on interest score, video type factor f type (T) represents the influence of video type on user interest;

[0250] Drag behavior factor f drag (F drag , T drag , L drag ) represents the influence of drag frequency, drag interval and drag jump distance on interest, which is calculated by the following formula:

[0251]

[0252] α1, α2, α3 are adjustment coefficients, which control the influence of each factor on interest score;

[0253] Viewing completion degree factor f completion (C complete ) represents the influence of user viewing video completion degree on interest, which is calculated by the following formula:

[0254]

[0255] C completeis the completion degree of the user watching the video (between 0 and 1), if the user watches the video completely, C complete = 1.

[0256] Further, the system further comprises a download rate adjustment module 15;

[0257] The download rate adjustment module 15 is configured to adjust the video download rate V adjust according to the difference between the buffer amount and the playing demand;

[0258] When the buffer amount C current reaches a threshold B thresh , the adjusted video download rate V adjust is calculated by the following formula:

[0259]

[0260] Wherein, M is the playing code rate of the video, C remaining is the remaining buffer amount, C total is the total buffer capacity of the buffer area, and V init is the initial download rate of the terminal given by the base station according to the CQI of the user terminal and the load of the base station.

[0261] Further, the system further comprises an automatic speed limit recovery module 16;

[0262] The automatic speed limit recovery module 16 is configured to switch to the basic buffer area when it is detected that the user does not have a drag behavior within a certain time and the buffer area starts to buffer new videos, and send the event to the base station, so that the base station automatically restores the terminal to the normal download rate.

[0263] The video buffer control system of the embodiment of the disclosure is used to implement the video buffer control method in the method embodiment one and the method embodiment two, so the description is relatively simple, and the specific description can be referred to the related description in the method embodiment, which will not be repeated here.

[0264] In addition, as Figure 3 shown, the fourth embodiment of the disclosure further provides an electronic device, comprising a memory 100 and a processor 200, the memory 100 stores a computer program, when the processor 200 runs the computer program stored in the memory 100, the processor 200 executes the above various possible methods.

[0265] Wherein, the memory 100 is connected with the processor 200, the memory 100 can adopt a flash memory or a read-only memory or other memories, and the processor 200 can adopt a central processing unit or a single-chip microcomputer.

[0266] In addition, the embodiments of the present disclosure further provide a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to perform the above various possible methods.

[0267] The computer readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, computer program modules or other data. The computer readable storage medium includes but is not limited to RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable read only memory), flash memory or other memory technology, CD-ROM (Compact Disc Read-Only Memory), digital versatile disc (DVD, Digital Video Disc) or other optical disc storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer.

[0268] It can be understood that the above embodiments are only exemplary embodiments adopted for illustrating the principles of the present disclosure, and the present disclosure is not limited thereto. Various modifications and improvements can be made by those of ordinary skill in the art without departing from the spirit and essence of the present disclosure, and these modifications and improvements are also considered to be within the protection scope of the present disclosure.

Claims

1. A video cache control method, characterized in that: The method comprises: A multi-level cache is used for smart terminals, which is divided into a basic cache and a prediction cache, corresponding to the cache of the currently playing video segment and the cache of the segment to be played when dragging the video; Dynamically calculate the cache size of the basic cache area of ​​the intelligent cache area based on the video bit rate, signal conditions and cache time; Determining a dynamic adjustment factor based on viewing habit parameters and network condition parameters to determine a cache capacity of a predicted cache area; Calculate the predicted cache size of the smart cache based on the video bitrate, dynamic adjustment factor, and predicted future viewing time when frequent video dragging occurs; Determining the dynamic adjustment factor according to the viewing habit parameter and the network condition parameter includes: The dynamic adjustment factor α(H u , N): Among them, H u is the user interest factor, which indicates the user's interest in the current video, ranging from 0 to 1. The coefficient δ represents the control of H u The impact on the amount of cache, the channel quality indicator CQI indicates the network quality, CQI max represents the maximum CQI value, and β represents the CQI adjustment coefficient.

2. The method according to claim 1, characterized in that The dynamic calculation of the buffer size of the basic buffer area of ​​the intelligent buffer area according to the video bit rate, signal condition and buffer time includes: The cache capacity B of the basic cache area is calculated by the following formula init : And calculate f(RSRP, SINR) using the following formula: Among them, M is the video bit rate, T init The signal condition is determined by f(RSRP, SINR). f(RSRP, SINR) is a function based on the signal quality reference signal received power RSRP and the signal to interference plus noise ratio SINR. It describes the impact of signal quality on the cache amount. RSRP thresh and SINR thresh are the thresholds of RSRP and SINR, β1 and β2 are adjustment coefficients used to adjust the sensitivity of the function to signal quality, and α is a constant coefficient used to further adjust the change of cache capacity under different signal qualities.

3. The method according to claim 1 or 2, characterized in that The calculation of the predicted buffer size of the smart buffer area according to the video bit rate, the dynamic adjustment factor, and the predicted future viewing time when the video is frequently dragged includes: The predicted buffer size B is calculated using the following formula pred : Among them, T pred To predict the future viewing time when videos are frequently dragged, α(H u , N) is a dynamic adjustment factor, which depends on the user interest factor H u and the network condition parameter N, which is used to determine the cache capacity of the prediction buffer area.

4. The method according to claim 3, characterized in that The method further comprises: Collect historical user behavior characteristics. For each video type, collect the following user behavior data: drag frequency F drag , drag interval T drag , drag jump distance L drag , viewing completion C complete ; Based on the collected historical behavioral feature data of users, the user interest factor H of the user for the video is predicted by the preset model, taking into account the characteristics of different video types and user behavior. u : H u =σ(γ1·f type (T)+γ2·f drag (F drag ,T drag ,L drag )+γ3·f completion (C complete )) Among them, σ(x) is the Sigmoid function, which maps the score to the interval [0, 1]. γ1, γ2, and γ3 are adjustment coefficients, which respectively control the influence of video type, dragging behavior, and viewing completion on the interest score. The video type factor f type (T): represents the impact of video type on user interest; Drag behavior factor f drag (F drag , T drag , L drag ): Indicates the impact of drag frequency, drag interval, and drag jump distance on interest, calculated using the following formula: α1, α2, and α3 are adjustment coefficients that control the impact of various factors on interest scores; Viewing completion factor f completion (C complete ): Indicates the impact of the user's viewing completeness on their interest in the video, calculated using the following formula: C complete Is the degree of completion of the user watching the video, between 0 and 1. If the user watches the video completely, C complete =1.

5. The method according to claim 1, wherein The method further comprises: Adjust the video download rate V according to the difference between the cache amount and the playback demand adjust ;include: When the cache size C current Reaching threshold B thresh When the video download rate V is adjusted, the following formula is used to calculate the video download rate V adjust : Among them, M is the video playback bit rate, C remaining is the remaining cache size, C total is the total cache capacity of the cache area, V init The initial download rate of the terminal is given by the base station based on the CQI of the user terminal and the load of the base station.

6. The method according to claim 5, characterized in that The method further comprises: When it is detected that the user has not performed any dragging behavior within a certain period of time and the cache area starts to cache new videos, it switches to the basic cache area and sends an event to the base station, so that the base station automatically restores the terminal to a normal download rate.

7. A video cache control system, characterized in that: The system comprises: A cache setting module is configured to use a multi-level cache area for the smart terminal, which is divided into a basic cache area and a prediction cache area, corresponding to the cache of the currently playing video segment and the cache of the dragged video segment to be played; A basic buffer zone caching module is configured to dynamically calculate the cache size of the basic buffer zone of the intelligent buffer zone based on the video bit rate, signal conditions and cache time; A prediction buffer area caching module configured to determine a dynamic adjustment factor based on viewing habit parameters and network condition parameters to determine a cache capacity of the prediction buffer area; and Calculate the predicted cache size of the smart cache based on the video bitrate, dynamic adjustment factor, and predicted future viewing time when frequent video dragging occurs; Determining the dynamic adjustment factor according to the viewing habit parameter and the network condition parameter includes: The dynamic adjustment factor α(H u , N): Among them, H u is the user interest factor, which indicates the user's interest in the current video, ranging from 0 to 1. The coefficient δ represents the control of H u The impact on the cache capacity, the channel quality indicator CQI indicates the network quality, CQI max represents the maximum CQI value, and β represents the CQI adjustment coefficient.

8. An electronic device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the processor runs the computer program stored in the memory, the processor executes the video cache control method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the video cache control method according to any one of claims 1 to 6 is implemented.

Images