Proactive Technique of Adaptive Video Streaming from Drones

Abstract

In recent years, recreational and commercial use of sUAS (small Unmanned Aircraft Systems) has increased dramatically. The emergence of sUAS equipped with high end cameras has paved the way to many novel applications, such as live-event broadcasts from drones. Ready-to-fly consumer drones traditionally use a fix-bitrate technique to encode video, for example, 1080p or 4K resolution. Unfortunately, fixed-rate video streaming performs poorly in high mobility cases where wireless channel conditions vary rapidly. An HD video stream might be a good fit in close proximity to the receiver, but as the drone moves away from the ground station, wireless link capacity can deteriorate quickly due to path loss, leading to buffer underflow and video playback interruptions. Adaptive bitrate streaming techniques are needed in uncertain wireless link streaming applications. ABR (Adaptive Bit Rate) video service is the new, widely adopted standard for many online video content platforms. However, existing ABR algorithms are reactive, meaning that they have to wait for wireless link conditions to change before they can adapt. Hence, they are not well suited for high-mobility systems where wireless link fluctuate constantly. Applications that deploy drones for live-event broadcasts have a need for a better system that can achieve high quality video streaming results in high mobility cases. Our proposed system is proactive in nature. It introduces the use of GPS (Global Positioning System) coordinates and future drone trajectory knowledge to improve the viewers' QoE (Quality of Experience), by increasing video quality and decreasing playback interruptions. Although far from complete, our system exhibits promise in adapting video streams better than traditional ABR algorithms for streaming from drones.