Csenge Hubay

About

Hi! I'm Csenge, I'm completing my PhD degree in ethorobotics at ELTE.This webpage is mainly made to contain the theoretical background for my research. The intended use is personal, but if you're here, feel free to take a look. I'm also open to suggestions if you're missing something important:D

DISCLAIMER

Sorry for sounding square, this webpage is mainly edited with AI and by no means reflects my artistic aspirations.

Research themes

• Ethorobotics and function-orientedness — why robot form should follow function, and what the uncanny valley looks like when reread as a problem of agent categorisation.
• Emotion attribution to artificial agents — people ascribe emotions to robots spontaneously and act on the ascription; the affective layer is licensed by that, and undermined when it misreads.
• Grounding models — Panksepp's neural affective systems against Plutchik's adaptive continuous states, and the case for a hybrid.
• Perception-to-affect — reading affect from body pose and multimodal signals in real time. This is the open gap, in the field as much as in the thesis.
• Learning the behaviour layer — reward design for an agent whose reward signal is partly its own internal state.

Platform

Experiments run on Mecanumbot: a mecanum-drive mobile robot with a Jetson Orin Nano, lidar and camera, a 17-keypoint body-pose pipeline, a neck and grabbers, and LED panels used for expressive output. The perception stack fuses DR-SPAAM person detection with YOLO-pose; the behaviour layer is built on py_trees_ros.