An Adaptive Video Data Representation Model to Increase Delivery Efficiency in Next-Generation Networks

This study proposes an adaptive video stream representation model that provides dynamic adjustment of bitrate, frame rate, compression ratio, and frame structure based on a comprehensive analysis of network conditions, content priorities, and technical features of endpoint equipment. The research methodology includes mathematical modeling of video data transmission processes, analysis of radio channel noise immunity, algorithmic formalization of adaptive optimization of video parameters, and simulation modeling in Wi-Fi, 4G/5G, and PON networks. The results show that the proposed model provides a bandwidth reduction of 22 - 30% compared to static coding and classical ABR algorithms, reduces buffer time by 40 - 60%, increases delay stability to 150 ms in 5G and to 300 ms in 4G networks, and decreases packet loss rate to 1 - 3%. Its PSNR and SSIM metrics remain stable and device load is reduced by 15 - 20%.