ROBOTICS · ZHEJIANG UNIVERSITY

Yang Liu 刘洋

Learning to perceive.
Planning to act.

I am an undergraduate in Robotics Engineering (Honors Program) at Zhejiang University, expecting to graduate in June 2027.

My research at Differential Robotics and FAST Lab, Zhejiang University is advised by Prof. Fei Gao. Previously, I worked on quadrupedal locomotion at ZJU–Hangzhou Global Scientific and Technological Innovation Center, contributing to the White Rhino robot.

I study deep reinforcement learning, embodied perception and state estimation, and motion planning and control, with an emphasis on algorithms that work on real robots under practical sensing and computational constraints.

Portrait of Yang Liu
Yang Liu / 刘洋Hangzhou, China
Perception → Planning → Control
RESEARCH FOCUS

Deep reinforcement learning

Embodied perception

Autonomous robots

SELECTED RESEARCH

Publications & preprints03

From sparse sensing to agile motion.

† Co-first authors · * Corresponding authors · Manuscripts under review

TIO-Former real flight, six-direction ToF sensor arrangement, and trajectory comparisonSTATE ESTIMATION
2026 ICRA 2027 · Under review

TIO-Former: Ultra-Lightweight 6-Directional ToF-Inertial Odometry for Nano-UAVs via a Streaming Causal Transformer

Yang Liu†, Yifan He†, Wenhao Zhao†, Xiangyu Mo, Yang Xu, Hao Wei, Mingze Ma, Huan Li, Yifan Wu, Fei Gao, Zipeng Dai, Xin Zhou*

Streaming range–inertial odometry for nano-UAVs, combining sparse ToF measurements and IMU data with bounded inference cost and memory.

54.4% lower position error vs. optical-flow baseline · 10.466 ms P95 latency

White Rhino quadruped sprinting on a track and carrying a 100 kg payload during testingLEGGED LOCOMOTION
Manuscript IEEE RA-L · Under review · Rebuttal submitted

White Rhino: Exploiting Foot–Ground Impact for High-Speed and High-Payload Quadrupedal Locomotion

Shaowen Cheng†, Yucheng Tao†, Yang Liu, Ze Wang, Xianwei Liu, Yanyan Yuan, Yu Qi*, Yongbin Jin*, Hongtao Wang*

Exploiting foot–ground impact to support fast locomotion and heavy payloads through the joint design of robot mechanics and motion.

8 m/s top speed · >100 kg dynamic payload

Manuscript under review; public paper link forthcoming.

BUILDING & EXPLORING

Research projects01

Perception meets interaction.

Conceptual grasp generation pipeline, not an experimental resultCONCEPTUAL ILLUSTRATION
2024–2026 Project lead

Dexterous grasping in cluttered scenes

Grasp-pose generation for dexterous hands, spanning simulation, dataset curation, model training, and evaluation of object relationships and contact states.

Provincial Undergraduate Innovation Program

RESEARCH IN PRACTICE

Experience

Differential Robotics

Robotics Algorithm Intern · Advised by Prof. Fei Gao

Lightweight multimodal odometry and autonomous UAV navigation.

ZJU–Hangzhou Global Scientific and Technological Innovation Center

Motion Control Algorithm Engineer · Advised by Prof. Hongtao Wang

Reinforcement learning for quadrupedal locomotion; simulation-to-reality transfer and hardware testing.

FAST Lab, Zhejiang University

Research Intern · Advised by Prof. Fei Gao

High-speed target perception and trajectory prediction.

ACADEMIC FOUNDATION

Education

Zhejiang University

B.Eng. in Robotics Engineering
Honors Program · Sep 2023 — Jun 2027 (expected)

LET’S CONNECT

Interested in robots that learn and act.

1211175035@qq.com ↗