CIS: Bridging The Gap Between Robots and Humans Through Explainability
Abstract:
Robots are traditionally placed behind fences for autonomous pick-and-place tasks. Today, we are witnessing robots moving to work with humans. They will not only need to manipulate but also hand objects to people and explain when autonomy fails. In this talk, I will start with how I tackled challenges in mobile manipulation tasks in a collaborative environment and a better way to let objects go for fluent human-robot handovers. Then I will focus on high-level hierarchical explanation generation algorithms using behavior trees and finer-level references for preferred concise explanations. This includes reference production through a cognitive status-informed approach and mixed-reality robot behavior replays to reason about objects no longer present. Finally, I will discuss how augmented reality can enable physically limited robots to gesture.
Date and Time
Location
Hosts
Registration
- Date: 17 May 2023
- Time: 06:00 PM to 08:00 PM
- All times are (UTC-06:00) Mountain Time (US & Canada)
-
Add Event to Calendar
- 2155 East Wesley Avenue
- Denver, Colorado
- United States 80208
- Building: Ritchie Engineering
- Room Number: 410
- Contact Event Hosts
- Co-sponsored by Eric Ericson
- Starts 16 April 2023 05:04 PM
- Ends 17 May 2023 07:00 PM
- All times are (UTC-06:00) Mountain Time (US & Canada)
- No Admission Charge
Speakers
Dr. Zhao Han
Bridging The Gap Between Robots and Humans Through Explainability
Robots are traditionally placed behind fences for autonomous pick-and-place tasks. Today, we are witnessing robots moving to work with humans. They will not only need to manipulate but also hand objects to people and explain when autonomy fails. In this talk, I will start with how I tackled challenges in mobile manipulation tasks in a collaborative environment and a better way to let objects go for fluent human-robot handovers. Then I will focus on high-level hierarchical explanation generation algorithms using behavior trees and finer-level references for preferred concise explanations. This includes reference production through a cognitive status-informed approach and mixed-reality robot behavior replays to reason about objects no longer present. Finally, I will discuss how augmented reality can enable physically limited robots to gesture.
Biography:
Dr. Zhao Han
- Denver IEEE Computer Society Guest Lecturer
- Postdoctoral Research Fellow, Colorado School of Mines
Dr. Han holds a Ph.D. in Computer Science from UMass Lowell in the HRI Lab, directed by AAAI Fellow Dr. Holly Yanco. Previously, he spent time in Canada and received his M.S. (advised by Dr. Carson Leung in the Database and Data Mining Lab) and B.S. degrees in Computer Science from the University of Manitoba. Dr Han is currently a Post-Doctoral Fellow at Colorado School of Mines and will start as an Assistant Professor in the Computer Science and Engineering department at the University of South Florida in August 2023.
Dr. Han received the best long-paper award at INLG 2022 and the best late-breaking report third prize at HRI 2022. Previously, he had led teams to win multiple robot competitions, including first place in the Panasonic Prototype 3D LiDAR Challenge and second place in the FetchIt Mobile Manipulation Challenge at ICRA 2019. In addition to research, he received a university-wide Mines Diversity, Inclusion, and Access (DI&A) Award in 2022.