Edge Media Server for Real-Time 4K Video Streaming with Multiple 5G-Enabled Drones

Recently, there has been a significant surge in the development of services utilizing multiple drones, particularly in domains such as detect and rescue operations. These advanced services leverage the capabilities of the high-speed 5G network to capture 4K videos from drones. Once these videos are sent through the MEC (Mobile Edge Computing) server, they are utilized for functions such as object detection, tailored to particular needs. To achieve this, the development of a media server capable of receiving, storing, and relaying 20 or more channels of 4K videos along with geometry information, is needed. During flight, over 20 drones simultaneously transmit data, including 4K videos at 8-20MBps bandwidth and 30 FPS(Frames Per Second), along with metadata containing sensor and gimbal information. This data is meticulously synchronized and formatted into MPEG-2 TS before being sent to a media server. Once transmitted, the media server receives and relays the videos to connected data servers, BM(Business Model) monitoring servers, and AI servers. These servers perform tasks such as storing, monitoring, and AI analyzing, enhancing the efficiency and relevance of the transmitted data.In a typical commercial 5G network, an upload speed of 100Mbps is common. However, in the case of the private "e-um 5G" network, upload speeds of up to 300Mbps or more are achievable. This enhanced upload capability enables up to 20 drones to transmit high-resolution video simultaneously. The media server developed in this paper is designed to receive and relay drone videos in the "e-um 5G" environment. To validate its capabilities, the media server was tested in a simulated environment with up to 50 drones connected and operational.In this paper, the edge media server processes 4K videos and metadata transmitted from drones and relays them for display on the web through various protocols. Furthermore, the GPS information extracted from the streamed video on the web is utilized to annotate the locations covered by drones on digital mapping platforms like Cesium and VWorld. The system demonstrates the capability to present four video streams along with their associated metadata within a single web page.