M.S. candidate in Mechatronics and Robotics at NYU Tandon School of Engineering.
Expected graduation: May 2027.
I build learning-enabled robotic systems that remain useful beyond idealized simulation by accounting for state-estimation error, contact uncertainty, actuator limits, and hardware constraints.
My current work focuses on legged locomotion, state estimation, and deployment-oriented validation for quadrupedal and humanoid robots.
- Legged and humanoid locomotion — reinforcement-learning-based control, gait shaping, and robustness under real-world constraints.
- State estimation for legged robots — IMU integration, leg odometry, contact constraints, and Kalman filtering.
- Learning-based control under imperfect observations — delayed, noisy, and partially observable sensing.
- Deployment-oriented robotics — embedded control, sensor integration, and hardware-aware validation.
A platform-adapted reinforcement-learning locomotion and deployment framework for a 12-DOF, MG996R-actuated SpotMicro-class quadruped.
- Isaac Lab-based PPO locomotion training.
- MG996R-inspired torque-speed saturation modeling.
- ONNX export and PyBullet cross-simulator evaluation.
- Jetson Orin Nano to STM32 deployment interface.
- Servo calibration, joint-limit clipping, communication watchdog, and safe-pose recovery.
A research project investigating locomotion policies trained without direct privileged base-velocity observations.
- Replaces simulator ground-truth base velocity with an estimated observation.
- Fuses IMU integration with stance-leg odometry using a per-axis Kalman filter.
- Uses IMU, joint-state, and noisy contact observations.
- Investigates robustness under accelerometer bias and contact-label noise.
Robot learning
Python · PyTorch · PPO · Isaac Lab · Isaac Sim · domain randomization
Estimation and robotics
Kalman filtering · IMU integration · leg odometry · forward/inverse kinematics ·
Jacobians · contact constraints
Simulation and deployment
MuJoCo · PyBullet · ONNX · Jetson Orin Nano · STM32 · PCA9685
Embedded systems
C/C++ · Python · I²C · SPI · UART · sensor integration
- Email: donguyenbao0710@gmail.com | bd2688@nyu.edu
- LinkedIn: linkedin.com/in/bao-do-86829a222