Adaptive Clustered SFU Architecture for User-Generated Video Transmission over LTE Networks

This study presents an adaptive user-generated video distribution system designed for large-scale arenas, where audiences can transmit live videos from smartphones over standard LTE networks. When hundreds of users attempt simultaneous video transmission, the limited wireless bandwidth causes severe uplink contention, resulting in unstable and delayed streams. To address this issue, the proposed system employs a clustered Selective Forwarding Unit (SFU) architecture combined with a statistically weighted, priority-based transmission control mechanism. Each participating device reports statistical metrics such as transmission performance and network stability, after which a priority-sorted selection process identifies candidate transmitters. Subsequently, a PING–PONG round-trip time (RTT) probing procedure selects the devices with the lowest latency for video transmission. This approach maintains stable and efficient operation of the user-generated video system under congested LTE network conditions, enabling real-time display of selected videos on large venue screens.