# Dejun Guo

> Staff Engineer @ XPENG | Robotics

Location: Mountain View, California, United States
Profile: https://flows.cv/dejunguo

10+ years of hands-on experience in R&D of robotics, visual servo control, and computer vision. Expertise and demonstrated capability in design, modeling, and control of robotic systems including autonomous ground vehicles, aerial robots (e.g. quadrotors), and manipulators. Enthusiastic about innovations towards elegant design and product development in robotics and autonomous systems. 

Highlights:
- Published over 15 papers in highly-ranked journals/conferences such as IJRR, ICRA, IROS, T-MECH, RA-L, etc. Google Scholar: https://scholar.google.com/citations?user=0BVUJtMAAAAJ&hl=en
- Some papers are among the most highly cited in the field.
- Won 2017 Student Best Paper Award at ASME Dynamic Systems and Control Conference (DSCC)

Specialties: 
- Expertise: Humanoid & Quadruped Robots, Deep Learning, Reinforcement Learning, Motion Planning & Control, Optimization, Computer Vision
- Tool: C++/Python, PyTorch, ROS/ROS2/lcm, Docker, Git
- Simulation: IsaacGym&Sim, MuJoCo, Pybullet, Gazebo

## Work Experience
### Staff Engineer @ XPENG
Jan 2022 – Present
Humanoid RL, Trajectory optimization, Whole-body control

### Staff Engineer @ UBTECH Robotics
Jan 2021 – Jan 2022 | Pasadena, California, United States
Navigation: SLAM, motion planning, and control

### Senior Engineer @ UBTECH Robotics
Jan 2019 – Jan 2021 | Greater Los Angeles Area
Navigation: SLAM, motion planning, and control

### Research Assistant @ University of Utah
Jan 2015 – Jan 2019 | Greater Salt Lake City Area
Achievements:  
- 1 first author publication on IJRR - the most top journal in the Robotics community
- Vision-based Localization: Developed and implemented a new visual localization method for underactuated aerial robots in GPS-denied environments through a single onboard camera and IMU by designing a new light-computational nonlinear observer based on Lyapunov method.
- Motion Planning: Developed and implemented a new image-based motion planning method for underactuated aerial robots through a single onboard camera based on the homography concept and differentially flat property to generate an optimally smooth and dynamically feasible trajectory for aggressive maneuvers such as flights through several narrow windows.
- Flight Control: Developed and implemented new control approaches for aerial robots through an onboard camera or a high-flying camera exploiting adaptive, repetitive, backstepping, and cascaded quaternion control to endow the ability of hovering, aggressive maneuvers, and flights with a cable-suspended payload, respectively, in GPS-denied environments with system uncertainties, disturbance, and underactuated issue.
- Accomplished 7 journal publications
- Presented in top robotics and control related conferences, e.g., ICRA, IROS, ACC, AIM, DSCC (Won Student Best Paper Award).

Practical Experience: 
- Established simulation environments in MATLAB and Gazebo as well as experimental platforms using a Pixhawk flight controller with an Odroid XU4 single board computer; Realized proposed algorithms using C++ in ROS. All computations were performed onboard in the embedded system.
- Developed image processing modules for point detection, object segmentation, and template matching adapted from OpenCV.
- Mentoring the senior design for tracking a mobile robot using a camera on quadcopter

### Reserach Assistance @ Shanghai Jiao Tong University
Jan 2012 – Jan 2015 | Shanghai City, China
Achievements:    
- Developed and implemented a new adaptive controller-estimator scheme for ground robots through an uncalibrated onboard perspective/omnidirectional camera to track a leader robot.
- Developed a new adaptive dynamic controller-observer scheme for space manipulators with 12 DOFs through a camera on the end-effector to grasp a moving object without prior knowledge of its motion and manipulator dynamics.
- Implemented a new trajectory tracking control for a mobile robot algorithm through an overhead camera.
- Accomplished 2 journal publications

Practical Experience:
- Established simulation environments in MATLAB; Realized algorithms using C++ with GUI in Visual Studio. All computations were performed onboard in the embedded system.
- Developed and maintained an indoors visual positioning system through two wide-angle lens cameras. It can precisely localize at most 16 different ground robots moving in a 6mx9m area with ±1cm and ±1° positioning error.


## Education
### Doctor of Philosophy - PhD in Robotics and Automation Engineering
University of Utah

### Master's degree in Control Engineering
Shanghai Jiao Tong University

### Bachelor's degree in Electrical and Electronics Engineering
Northwestern Polytechnical University


## Contact & Social
- LinkedIn: https://linkedin.com/in/dejun-guo-982796a2

---
Source: https://flows.cv/dejunguo
JSON Resume: https://flows.cv/dejunguo/resume.json
Last updated: 2026-04-12