CS Events
PhD DefenseHuman Behavior-Driven Cyber-Physical Systems for On-demand Gig Delivery |
|
||
Tuesday, July 13, 2021, 10:30am - 12:30pm |
|||
Speaker: Yu Yang
Location : Remote via Zoom
Committee:
Prof. Desheng Zhang (chair)
Prof. Jie Gao
Prof. Hao Wang
Dr. Lu Han (outside member, Google)
Event Type: PhD Defense
Abstract: In this thesis, I will introduce my work on the foundations and applications of Human Behavior-Driven Cyber-Physical Systems with a concrete use case of on-demand delivery as part of the Gig Economy. The key challenge for on-demand gig delivery is to obtain real-time gig worker status in a cost-efficient approach to enable timely delivery and efficient order scheduling. However, the existing sensing approaches in both industry and academy are limited to achieve a practical balance between reliability and scalability due to cost issues and an uncertain operational environment. I design a human behavior-driven system based on transfer learning to infer the accurate workers' status based on worker reporting behavior. Then, I will talk about a smartphone-based approach to scale it up to a nationwide system. This system has been deployed and operated at a real-world industry platform via our collaboration with Alibaba On-demand Delivery Platform Eleme. Based on this deployment, I will provide some results on how gig workers and the proposed system interact with each other and improve each other’s behavior. Finally, I will talk about some opportunities and challenges for my future work.
: