About
我是黄乐轩,港科大计算机工程大三学生,机器人副修,人工智能延伸主修。主要方向是嵌入式电控系统与机器人控制算法,能力覆盖整车嵌入式架构、控制算法、仿真验证、跨模块调试与团队协作。大一加入 RoboMaster 战队,后担任约 10 人嵌入式组组长,维护 RM2025-Core,并入选 RoboMaster 全明星赛。这个个人网站会比一页简历更完整地展示我的 RM 队内项目、课设、外包项目和开源控制算法。
Education
Hong Kong University of Science and Technology (HKUST)
2023.09 — Present
BEng in Computer Engineering · Minor in Robotics · Extended Major in Artificial Intelligence
Selected Coursework: System Modelling, Analysis and Control · Machine Learning · Embedded System · Robotics · Design and Analysis of Algorithm
Tech Stack
Embedded Systems
Robotics & Control
Simulation & ML
Languages & Tools
Spoken Languages
RoboMaster Team Projects
Engineering Robot
RM2025 · HKUST ENTERPRIZE · All-Star Game
5.8sCompetition Record: Fastest 4-Level Ore Exchange
#1, #2Ore Exchange Economy · Difficulty #1 · Success Count #2
All-StarRM2025 All-Star Game
Sole embedded engineer for the Engineering Robot group. Developed the full electro-mechanical control system for the team's first-ever 6-axis manipulator mounted on a mobile swerve-drive platform. Designed and iterated multiple versions of a custom isomorphic controller for intuitive teleoperation with haptic force feedback.
R&D Journey
- Phase 1 — 6-Axis Arm First MovementBuilt and validated the first prototype: mechanical structure, motor communication, and basic kinematics.▶ Watch
- Phase 2 — Custom Controller v1: Delta RobotDeveloped a Delta-robot-based haptic controller with impedance control for compliant, force-sensitive teleoperation.▶ Delta impedance control▶ 6-Axis teleop via Delta
- Phase 3 — Final Controller & Full IntegrationReplaced Delta with an isomorphic mini-arm controller for better motion mapping. Integrated arm + chassis into the complete robot. Delivered 5.8s task completion record and #1 match-average scoring efficiency.▶ Ore exchange (lab) ▶ Field test (1st-person POV) ▶ Stairs test (Bilibili)
Embedded Department Leader
RM2025 · HKUST ENTERPRIZE
Led the ~10-person embedded team. Designed and delivered a ~10-session training course (2h each) covering embedded fundamentals, GPIO, serial, CAN, motor control, and PID. Course videos published on Bilibili. Maintained RM2025-Core — the team's embedded library — managing feature development, module updates, urgent bug fixes, and pull request code reviews across all sub-projects.
- Delegated sub-projects: power control, motor drivers, serial communication protocol, DSP/matrix utility library
- Conducted onboarding training for all new embedded recruits
- Enforced coding standards and cross-module interface stability
Hero Robot
RM2024 · STM32F407
Full embedded system: omnidirectional chassis control (Mecanum), 2-axis gimbal, and friction-wheel launcher. Multi-motor real-time communication and control via CAN/FDCAN.

Dart System
RM2024 · STM32F427
Launch control system for both generations of the dart robot. Precision trajectory and timing control for the friction-wheel launch mechanism.
Swerve-Drive Chassis Exploration
RM2024
Early-stage exploration of swerve-drive chassis kinematics modeling and control algorithm prototyping. A technology tree the team had not yet unlocked.

Dual 6-Axis Mobile Manipulator Simulation
RM2026 · MuJoCo
Forward-looking simulation of dual 6-axis manipulators on a mobile base — laying groundwork for the team's future technology.
Course Project

4-Axis Exoskeleton Arm
HKUST ELEC 3300 · STM32H750 + FreeRTOS · 2025
4 joint actuators with wrist IMU for human intent prediction and dynamics feedback. Full kinematics & dynamics model with impedance control for real-time force feedback. Assisted lifting and arm motion augmentation.
Contract Projects
Biped Robot RL Locomotion
Contract · Isaac Sim + MuJoCo · PPO · 2025
Sim-to-Sim and Sim-to-Real training pipeline for biped locomotion. PPO-trained policy (Actor-Critic). Responsible for both embedded systems and RL control policy.

Squash Ball Serving Machine
Contract · STM32H723 · 2026 · 3 Weeks · Delivered
Production-grade embedded control system: precise regulation of launch frequency, angle, and velocity. Delivered as a complete prototype within a tight 3-week schedule.
Open Source
Chassis Power Control Algorithm
RM2024 · STM32F407
A chassis power management framework for RoboMaster competition robots. Uses Recursive Least Squares (RLS) to automatically identify motor power model parameters — no manual tuning required. A power ring constrains output to stay within the competition power limit without compromising chassis maneuverability. An energy ring with PD controller manages supercapacitor charge/discharge to maximize burst power and energy utilization. Supports mecanum, omni, swerve, and balance chassis. Robust error handling covers sensor disconnections. All team robots achieved zero health deduction from over-power throughout the season.
Note: the majority of RoboMaster codebases reside in the team's private GitHub organization. Course and contract project repositories are also private.
Honors
- 2025.07 — HKSAR Talent Development Scholarship TDS (2024/25)
- 2025.08 — RoboMaster 2025 Super Final National Championship: All-Star Game selectee; ranked #1 in the competition metric
- 2025.08 — RoboMaster 2025 Super Final Revival Tournament, Second Prize
- 2025.04 — RoboMaster 2025 University League Zhejiang Station Champion
- 2024.08 — RoboMaster 2025 Super Final Revival Tournament, Second Prize
- 2024.04 — RoboMaster 2024 University League Guangdong Station, First Prize
Journey
- 2023Two introductory mini-cars (team project). Passed the RM entry exam. Started embedded work on the Hero Robot.
- 2024Hero Robot. Dart System. Swerve-drive chassis exploration. Open-source chassis power-control algorithm. RM2024 University League Guangdong Station First Prize.
- 2025Embedded Department Leader (~10 ppl). RM embedded training course (Bilibili). RM2025-Core maintainer. Engineering Robot (6-axis arm + custom controller). All-Star. 5.8s WR. Economy #1. RL biped locomotion (contract). 4-axis exoskeleton arm (course). Zhejiang Station Champion.
- 2026Squash ball serving machine — delivered in 3 weeks (contract). FYP: modular snake robot. Seeking internship in embedded / robotics.