A self-adaptive learning network real-time video transmission method and device

Abstract

The invention relates to an adaptive learning network real-time video transmission method and device. This method is mainly aimed at end-to-end video transmission application scenarios under wireless networks. In this method, the data tracking module at the sending end, the network state perception module and the Bayesian controller cooperate to achieve the purpose of controlling the sending rate and avoiding congestion. The receiving end is mainly responsible for information feedback and assisting the data collection of the sending end. This method can sense the network state in time, and determine the optimal transmission rate through the rate decision mechanism based on historical data and Bayesian theorem. During the entire decision-making process, the present invention fully considers the highly variable characteristics of wireless networks, perceives long-term changes in the network environment through historical data, and adjusts the transmission rate with the help of link delays to adapt to instantaneous fluctuations in the network, thereby improving bandwidth The purpose of reducing the utilization rate and link delay is to improve the user experience of real-time video applications.

Description

technical field [0001] The invention relates to the field of multimedia, in particular to an adaptive learning network real-time video transmission method and device. Background technique [0002] With the development of mobile network and video technology, coupled with the upcoming 5G era, wireless network has become an increasingly popular network access method, and real-time video applications are also ushering in explosive growth. Whether in industry or academia, real-time video transmission methods for wireless network environments have received extensive attention. [0003] Compared with the traditional Ethernet network environment, the network environment of the wireless network has obvious differences, which is mainly reflected in the stability of the network link. For a wireless network link, whether it is bottleneck bandwidth, link propagation delay, or link packet loss rate, they are all in the process of changing in real time, and sometimes even link interruptio...

AI Technical Summary

Problems solved by technology

In addition, in order to ensure the picture quality, it is necessary to ensure high throughput, high stability, and good fairness when multiple data streams coexist. These demanding requirements are difficult to guarantee by many traditional transmission methods.

[0005] Therefore, there are still severe challenges in the transmission method for real-time video applications in wireless networks: 1) Real-time video applications have strict requirements on delay characteristics, but the high-speed jitter characteristics of wireless networks increase the difficulty of ensuring delay

At any time, the bottleneck bandwidth of the wireless network may decrease. If the bottleneck bandwidth drops below the sending rate, it will cause an increase in delay and a decrease in user experience.

2) High channel utilization is difficult to guarantee

It is ideal that the sending rate is exactly equal to the bottleneck bandwidth of the link, but due to the irregular jitter characteristics of the wireless network, it is difficult for the application to observe the real bottleneck bandwidth, which increases the difficulty of ensuring the optimal sending rate

3) There will be competition when multiple data streams coexist

When there are multiple data streams in a wireless link, the ideal situation is to allocate the bottleneck bandwidth fairly. However, for these data streams, it is difficult to observe the real upper limit of the bottleneck bandwidth and the number of coexisting data streams, which is prone to uneven distribution.

Images