Tianci Yin
MPhil Researcher at HKUST
Generalist robot learning · Visuotactile perception
I’m an MPhil researcher at HKUST, developing robot learning frameworks that investigate how perceptual information can be used effectively for planning. My previous work focused on visuotactile perception for robotic manipulation. I received my BEng in Computer Engineering from HKUST. I also enjoy tinkering with virtual agents. Physical or virtual, I like hatching agents—and seeing what they get up to.
tyin at connect dot ust dot hk
Research
I work on generalist robot learning and visuotactile perception at the Mevita Robotics Lab.
Generalist robot learning
I’m drawn to the idea of agents as companions: capable of helping, adapting, and bringing some character to the interaction. My research focuses on a practical part of that vision—how robots can learn to extract and represent task-relevant information from rich sensory observations, and use it to guide planning.
Visuotactile perception
My previous work explored vision-based tactile sensing at both the fingertip and across the arm. For fingertip sensing, we developed a structural-color-based tactile sensor that measures indentation depth across the sensing surface through a direct mechano-optical relationship, rather than reconstructing surface geometry from illumination-dependent image intensities, as in photometric stereo.
Selected publication
Embodied sensorimotor integration for whole-arm tactile sensing and adaptive robotic manipulation
Yifeng Tang, Tianci Yin, Tieshan Zhang, Liu Yang, Gen Li, Hao Ren, Zhengrong Ling, Haoxiang Zhao, Ruijia Zhang, and Yajing Shen.
Nature Sensors