CN110572699A

CN110572699A - Network-assisted video streaming media transmission optimization method based on multi-cell cluster

Info

Publication number: CN110572699A
Application number: CN201910955331.3A
Authority: CN
Inventors: 刘奕彤; 何伟; 沈玉博; 杨鸿文
Original assignee: Beijing University of Posts and Telecommunications
Current assignee: Beijing University of Posts and Telecommunications
Priority date: 2019-10-09
Filing date: 2019-10-09
Publication date: 2019-12-13

Abstract

The invention discloses a network-assisted video streaming media transmission optimization method based on a multi-cell cluster, which comprises the following steps: the method comprises the steps of carrying out a multi-cell cluster network bandwidth bearing capacity estimation process before a video playing client selects the next video segment code rate, estimating by using the position and the speed of a current cell where a user is located in the estimation process, giving the length of time to be optimized, estimating the number of cells to be crossed by the user in the period of time, obtaining the average network bandwidth bearing capacity estimation of the multi-cell cluster in the period of time in the future by collecting the wireless link state or the network capacity in the cells to be crossed by the user according to a certain calculation mode, and giving the next video segment code rate selection by the video playing client according to the estimation.

Description

Network-assisted video streaming media transmission optimization method based on multi-cell cluster

Technical Field

The invention relates to the technical field of multimedia transmission in a wireless network, in particular to a method for optimizing the playing process of video streaming media based on network bottom layer information by means of network assistance for a video user moving at a high speed.

Technical Field

The performance of the wireless network is improved, so that online multimedia services are more and more common, and a large proportion of video watching scenes occur under the condition of high-speed movement, such as watching live video and short video on high-speed rails and subways. Smooth and clear video playing service on a high-speed mobile user side becomes a demand.

The problems are that: a user quickly crosses a plurality of cells in a short time, experiences a rapid signal strength change in each cell, and quickly crosses a plurality of cells with completely different loads. In turn, base stations carrying these high speed users, such as base stations along the subway, are subject to rapid changes in their load capacity due to the high speed movement of the users. Therefore, in this scenario, even though the radio link condition of a single cell is better and the load is not high, the user still cannot see the clear and smooth video streaming media service.

The current situation is as follows: the owner of the wireless network, such as an operator or an owner of the wlan access, can only optimize and improve the capability of the wireless network, such as improving the user bandwidth, increasing the network speed, and reducing the network transmission delay; on the other hand, many video playing websites or clients use adaptive streaming media transmission, and adapt to the fluctuation of network bandwidth depending on adjustable picture coding. However, these video playing websites can only rely on the optimization means of the service layer, such as caching for a longer time in advance, reducing the time window of bandwidth estimation, increasing the chip length of the adaptive streaming media, and the like, to solve the video playing problem on the high-speed mobile user side.

Conventional adaptive streaming technologies and protocols, such as DASH, HLS, etc., do not cope well with scenarios where a user passes through multiple cells at high speed. They predict the network condition in the next period of time by estimating the network bandwidth in the past period of time, and select the corresponding video bitrate to download according to the predicted network condition. For example, in the kurto bandwidth estimation algorithm, the bandwidth carrying capacity of the network in a future period is estimated according to the size of a video clip obtained by an application layer in a past period and the downloading time of the video clip. This approach is effective when the user is traveling through a cell at low speed, because during the download time of a video clip the user is still within the cell or has traveled only through a cell, the underlying channel conditions change slowly and regularly, and it is feasible to use past network bandwidth values to predict future network bandwidth. However, in a scenario where a user passes through multiple cells at a high speed, this bandwidth estimation method is not suitable because the user may have passed through multiple cells during the downloading time of a video segment, the average bandwidth value in the past period is not correlated with the channel conditions of multiple cells to be passed through by the user in the future, and it is very inaccurate to predict the channel conditions in the future period, which causes waste of network bandwidth resources or poor video blocking, and greatly reduces the viewing experience of the user.

disclosure of Invention

The invention provides a method for selecting the code rate of the next video clip through the estimation process of the bearing capacity of a multi-cell network under the scene that a user passes through a plurality of cells at a high speed. The accuracy of the video playing client side for estimating the network bandwidth is greatly improved, and video blocking caused by the fact that the network bandwidth is estimated by mistake is reduced. The main principle is as follows: estimating the position and the speed of the current cell where the user is located, giving the time length to be optimized, estimating the number of cells to be crossed by the user in the period of time, estimating the network bandwidth bearing capacity of the cells by the upper network unit by collecting the wireless link conditions in the cells to be crossed by the user, and calculating the average network bandwidth bearing capacity of a plurality of cells in the future period of time. The difference from the existing mechanism is that the invention makes full use of the network bandwidth carrying capacity of each cell that the user is going to traverse by using a network-assisted mode.

The specific steps of the invention include cell number and time calculation, multi-cell network bearing capacity estimation and code rate selection:

1) Cell number and time calculation

Firstly, collecting the position (x, y) and the speed v of a current cell where a user is located, giving a time length T to be optimized according to the distribution of the cell on a map and the cell range, and predicting the number N of cells to be crossed by the user in the time T and the time tn of the user in each cell.

2) Multi-cell network bearer capability estimation

After calculating the number N of cells to be traversed by the user and the time tn elapsed by the user in each cell, the upper network element collects l N the network capacity in these cells, and estimates Bn the network bandwidth carrying capacity of these cells. And calculating to obtain the average network bandwidth bearing capacity estimation B of a plurality of cells in a future period of time according to the time of the user in each cell as the weight.

3) Code rate selection

after the previous multi-cell network bearing capacity estimation B is completed, the video playing client starts to select the next video clip code rate R, compares the multi-cell average network bandwidth bearing capacity estimation B with each video code rate grade (R1, R2, … … and Rm) stored by the server, collects the buffer length BS, and selects the appropriate next video clip code rate R according to the comparison result and the BS.

And when the next video clip is downloaded, continuously repeating the three processes.

Drawings

FIG. 1 is a schematic view of the scene of the present invention

FIG. 2 is a schematic flow chart of the method of the present invention

Detailed Description

According to the mechanism proposed by the present invention, it is assumed that there are 10 cells, cell a, cell B, … …, and node J, respectively, on the path that the user will traverse. There are a certain number of access users in each cell. At a certain time, the user is at a certain position (x, y) of cell a, whose velocity is v. Given a length of time, T, as the length of video clips stored within the service, it is calculated that the user will traverse 4 cells at time T, where time T is T1 for cell a, T2 for cell B, T3 for cell C, and T4 for cell D.

The network capacity in the cell A, B, C, D is collected by the upper network unit to be l 1, l 2, l 3, l4, and the network bandwidth carrying capacity of the four cells is estimated to be B1, B2, B3, B4. And calculating the network bandwidth carrying capacity B1, B2, B3 and B4 of each cell according to the time t1, t2, t3 and t4 of the user in the four cells as a weight to obtain an average network bandwidth carrying capacity estimation B of the four cells in a future period of time.

After the upper network unit finishes the estimation of the average network bearing capacity of the four cells, the video playing client starts to select the next video clip code rate R, firstly, the average network bandwidth bearing capacity estimation B is compared with each video code rate grade (R1, R2, … … and Rm) stored by the server, then the buffer length BS is collected, and the appropriate next video clip code rate R is selected according to the comparison result and the BS. And after the next video clip is downloaded, continuously repeating the three processes.

Claims

1. A network-assisted video streaming media transmission optimization method based on a multi-cell cluster is characterized by comprising the following steps:

Step one, collecting position and speed information of a current cell where a user is located, and estimating the number of cells which the user will pass through in the period of time according to a certain calculation method aiming at the time length to be optimized;

step two, the upper network unit collects the wireless link conditions in each cell which the user will pass through, estimates the network bandwidth bearing capacity of the cells, and enters the next code rate selection process after obtaining the network bandwidth bearing capacity estimation of the multi-cell cluster in the future period of time according to a certain calculation mode;

And step three, after the estimation of the bandwidth bearing capacity of the multi-cell cluster network is completed, the upper network unit can compare the estimated value with each video code rate grade stored by the server, and select a proper code rate grade as the next video segment code rate requested by the video playing client to the server according to the current buffer length of the video playing client and other requirements of the client.

2. The method as claimed in claim 1, wherein before the video playing client selects the next video segment bitrate, a multi-cell cluster network bearing capacity estimation process is performed first.

3. The method as claimed in claim 1, wherein the playing time to be optimized in step (a) is one or more video coding segments, or a time length unrelated to the duration of the coding segments.

4. The method as claimed in claim 1, wherein the calculation method for predicting the number of cells to be passed through is determined by the client according to the current location and speed of the user, or according to a prediction judgment algorithm of an upper unit of the network, or according to a behavior habit of the user.

5. The method as claimed in claim 1, wherein the radio link condition in each cell that the user will traverse is the network capacity of the cell, the distribution of access users, or the variation of received power of reference signals on the moving path of the user in the cell.

6. The method of claim 1, wherein the upper network element estimates the network bandwidth carrying capacity in a cell that the user is going to traverse.

7. The method as claimed in claim 1, wherein the calculation manner of the upper network unit in step two for estimating the average network bandwidth carrying capacity of the multi-cell cluster in a given time is that, according to the network capacity of each cell in the multi-cell cluster, the average network bandwidth carrying capacity can be calculated by taking the moving time of the user in each cell as a weight.

8. The method of claim 1, wherein the next video segment bitrate selected in step three refers to a comparison result between the multi-cell average network bandwidth carrying capacity estimation obtained by the upper network unit and each video bitrate level stored by the server.

9. the method of claim 1, wherein the next video segment bitrate selected in step three refers to a current buffer length of a video playing client.