Huawei · Embodied intelligence

Shiyue Zhao

赵世越

Learning to act in the physical world.

I am currently at Huawei, working on embodied intelligence. My research background connects vision-language-action systems, robot decision-making and control, and reinforcement learning.

Ph.D., Tsinghua University · Former visiting scholar, University of Michigan, Ann Arbor.

Portrait of Shiyue Zhao
Research · Learning · Real-world systems

Research interests

01 / Directions

My current focus is embodied intelligence, building on earlier work in autonomous driving, wheeled robots, and learning-based control.

01 — CURRENT FOCUS

Embodied intelligence

Intelligent systems that connect perception and action through interaction with the physical world.

02 — MODELS & DATA

Vision, language & action

VLA systems and multimodal driving data, connecting visual observations with decision-making and control.

03 — LEARNING & CONTROL

Reliable robot behavior

Reinforcement learning, safety-aware decisions, and control for wheeled robots and autonomous vehicles.

Selected work

02 / From methods to real vehicles

Open source & datasets

03 / Resources for research
Featured dataset · Multimodal driving

Extreme Driving Dataset

Driving observations paired with rich chassis dynamics, including raw IMU and GPS heading, for limit-handling research and VLA model training. Six scenario categories span normal driving, complex traffic, critical situations, low light, rain, and snow.

01 / Complex traffic
02 / Critical driving
03 / Snow driving
Simulation

Vehicle dynamics model

A Python model with validated high-fidelity tire simulation for conventional and drift driving.

View repository ↗
Reinforcement learning

CarSim–Python co-simulation

A Simulink bridge connecting CarSim and Python for vehicle reinforcement learning research.

View repository ↗
Visual perception

Extreme Road Image Dataset

Road surface images for research on tire–road friction coefficient estimation.

Explore dataset ↗

Research updates

04 / News

Current: I am at Huawei, working on embodied intelligence.

Research news archive · 2024–2026

Publications

05 / Journal & conference papers

Journal Papers (full text provided where possible)

Conference Proceedings

Background & community

06 / Academic journey

Experience & education

  1. HuaweiResearch in embodied intelligence
  2. Ph.D. · Tsinghua UniversitySchool of Vehicle and Mobility
    Advisor: Prof. Junzhi Zhang
  3. Visiting Scholar · University of MichiganNext Generation Mobility Systems Lab, Ann Arbor
    Host: Dr. Neda Masoud
  4. Bachelor of Engineering · Central South UniversitySchool of Traffic and Transportation Engineering

Academic service

Journal reviewer

  • ISA Transactions
  • IEEE Transactions on Industrial Electronics
  • Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering
  • Scientific Reports

Conference reviewer

  • Chinese Automation Congress
  • IEEE International Conference on Unmanned Systems

Honors & awards

  • [2023.11] The Best Paper Award at IEEE DSInS (2023)
  • [2023.09] Tsinghua University Comprehensive Outstanding First-Class Scholarship
  • [2022.05] Infineon Innovation Scholarship
  • [2022.04] Outstanding Volunteers of the 2022 Beijing Winter Olympics and Paralympics
  • [2021.06] Outstanding Graduates of Hunan Province
  • [2021.01] First Prize in the National Competition of Transport Science and Technology for China College Students (Top 1%)
  • [2018,2019,2020] National Scholarship

荣誉只是纸上之名,真正的价值在于无私的爱在世界留下的微小改变。
Honors are mere names on paper; their true value lies in the quiet changes my selfless love leaves in the world.

Beyond the papers

07 / Creative experiments

Small explorations in interactive graphics and generative art.

Get in touch

Let’s exchange ideas.

I welcome academic discussions on embodied intelligence, VLA systems, robotics, and reinforcement learning.

Stary132@163.com ↗