Research On Mobile Terminal Adaptation And Optimization Of Live Video Transmission Algorithm

With the rapid development of the Internet and video technology,many applications such as live broadcasting and online education have emerged,which have changed our life and work.At the same time,the huge traffic generated by real-time video has brought great challenges to the Internet transmission,and the problems of excessive video lag and high delay affect the user’s viewing experience.Therefore,the video stream transmission algorithm and its optimization become a research hotspot.We design and implement an evaluation platform for mobile video broadcast transmission algorithms,and use real mobile network data to simulate links to analyze the performance of different algorithms in different network scenarios.Experimental results show that the algorithm based on machine learning can achieve higher throughput than the GCC algorithm,but there are problems such as training difficulties and algorithm performance instability.In view of the above problems,we conduct an in-depth study on the algorithm based on machine learning,and explore the influence of different factors on the algorithm performance through comparative experiments.The experimental results show that the length of historical information has an effect on the judgment of network congestion by the congestion control algorithm,while the structure of neural network has little effect.According to the experimental results,LVCC algorithm is proposed based on DQN algorithm,and the network condition detection algorithm is proposed to ensure the robustness for the unstable online environment.Finally,based on the above algorithm,we design and implement the adaptive system of the video broadcast transmission algorithm in the mobile terminal,and the experiment is verified in the real live broadcast scene.The results show that the optimized LVCC algorithm has some improvements in video RTT,video lag rate and other indexes,and improves the video transmission quality.