Computer Science Speaking Skills Talk

Monday, December 3, 2018 - 12:00pm to 1:00pm


Traffic21 Classroom 6501 Gates Hillman Centers


JUNJUE WANG, Ph.D. Student

Bandwidth-efficient Live Video Analytics for Drones via Edge Computing

Speaker: Junjue Wang

Location: GHC 6501

Bandwidth-efficient Live Video Analytics for Drones via Edge Computing

Real-time video analytics on small autonomous drones poses several difficult challenges at the intersection of wireless bandwidth, processing capacity, energy consumption, result accuracy, and timeliness of results. In response to these challenges, we describe four strategies to build an adaptive computer vision pipeline for search tasks in domains such as search-and-rescue, surveillance, and wildlife conservation. Our experimental results show that a judicious combination of drone-based processing and edge-based processing can save substantial wireless bandwidth and thus improve scalability, without compromising result accuracy or result latency.

In this talk, I will first introduce our vision of enabling live video analytics on drones and provide a motivating usage case in the search and rescue context. I will then present our experiments estimating how much processing capabilities a small drone can carry. These experiments show that onboard processing alone cannot satisfy the computational need and offloading the computation to an edge server is needed. After system design challenges are identified, I will talk in detail about four proposed optimizations to address these challenges.

This talk is based on a joint work with Ziqiang Feng, Zhuo Chen, Shilpa George, Mihir Bala, Padmanabhan Pillai, Shao-Wen Yang, and Mahadev Satyanarayanan.

Presented in Partial Fulfillment of the CSD Speaking Skills Requirement.

For More Information, Contact:


Speaking Skills