Low latency is critical for edge-enabled autonomous drones to do real-time computer vision tasks like target tracking effectively. We discuss our work on analyzing the latency of the real-time drone video stream using the SteelEagle autonomous drone system. By varying factors such as the cloudlet location, network type, and the video decoding library used, we show that the latency is bottlenecked not by the network but by the decoding of the RTSP video stream generated by the drone.

11 pages

Return to: SCS Technical Report Collection
School of Computer Science

This page maintained by reports@cs.cmu.edu