About Me
My name is Chenghao Wang. I hold my Bachelor of Engineer at Tongji University in China. During that time, I minored in mathematics and studied basic theories of electricity and control.
Currently, I’m a master student at the Electrical and Computer Engineering Department, Technical University of Munich(TUM). Throughout my master’s education, I studied robotics, machine learning, control theory, and completed various projects with different kinds of robots.
Research Interest
- Robotics: Motion Control, Trajectory Planning
- Machine Learning: Reinforcement Learning
Work Experience
- Tutor for course “Robocup@Home”
- TUM, 10/2019-01/2020
- Assistance to Robocup@Home project
- Tutor for course “Multidimentional Signal Processing” (MDSP)
- TUM, 11/2018-02/2019
- Assistance to MATLAB homework
Projects
Whole-body Hierarchical task execution for mobile service robots (master thesis)
This project proposed a torque-based general framework ros_task to realize whole-body hierarchical tasks execution. This framework is based on ROS and suitable for various ROS-based robots. Some basic tasks designed for mobile service robots are applied in this framework. They can achieve some specific behaviors for the robot, such as picking up objects while moving.Reproducing of handwriting trajectory with robotic arm
This project focused on motion planning and control of the robot. Model predictive control (MPC) is used to generate feasible trajectories from the handwriting and an adaptive torque controller based on dynamic model of the robot is designed for control. The LED on robot gripper can show the reproduced trajectory by time-lapse photography.Handwriting Reproduced Trajectory 
Gait planning of quadruped robot based on reinforcement learning
This project applied reinforcement learning to do gait planning for a quadruped robot. SARSA algorithm is used to adjust gait parameters based on the sensor data so that the robot can climb up different obstacles.
SimulationDemo