|"Interaction in Dynamic Scenarios of HRI" |
|Konuşmacı ||: ||Dr. Zeynep Yücel |
|Tarih ||: ||19 Ağustos 2019 (Pazartesi) |
|Saat ||: ||15:30 - 16:30 |
|Yer ||: ||Bilgisayar ve Bilişim Fakültesi, Oda: 4305 |
Design of social interaction robots is a very challenging research problem. In this talk, we will consider dynamic settings where a robot is employed in a public space and possibly acts as part of a -social- pedestrian group. In such settings, the robot is required to act in a socially acceptable way while moving both alone and as part of a pedestrian group. In this respect, we will firstly discuss how to identify pedestrian groups in low to mid-density crowds. Then, we will have a closer look into group motion. Namely, we will present how intensity of interaction, gestures accompanying interaction, and social relation between the peers affect group dynamics. In doing that, we will present several recognition algorithms or motion models, which potentially enable the robot to understand the crowd better and replicate human-like locomotion/interaction.
Zeynep Yucel got her BS degree from Bogazici University, Istanbul, Turkey in 2003 in Electrical and Electronics Engineering, and her MS and PhD degrees from Bilkent University, Ankara, Turkey in 2005 and 2010, respectively. She worked as a postdoctoral researcher at ATR labs in Kyoto, Japan for 5 years. She was awarded a JSPS fellowship in 2016. Since April 2017, she serves as an assistant professor at Okayama University, Japan. Her research interests are investigation of human behavior and social interaction robots through signal processing, computer vision, pattern recognition techniques.