| Type | Semester project |
| Split | 30% theory, 50% software, 20% simulation-based evaluation |
| Knowledge | Python; data analysis; basic linear algebra Bonus: control systems, simulation (MuJoCo or similar) |
| Subjects | Software development, simulation, human motion analysis |
| Supervision | Hanli Zhang |
| Published | 12.01.2026 |
Studying multi-limb coordination requires not only modeling techniques, but also robust and reusable tools for data processing, simulation, and evaluation.
This project focuses on developing a lightweight, modular platform that integrates motion data processing with simulation-based evaluation of given dynamical system controllers.
The student will not design new control or learning algorithms, but will instead focus on building infrastructure that allows existing dynamical system models to be systematically tested, visualized, and compared in multi-limb scenarios.
Approach
- Review basic concepts of goal-directed motion and dynamical systems
- Design a clean data structure for multi-limb motion datasets
- Implement reusable modules for:
- Data preprocessing and visualization
- Trajectory playback and perturbation injection
- Simulation wrappers for multi-effector tasks
- Integrate provided reference controllers into the simulation pipeline
- Evaluate coordination behavior under different task settings
Expectation
- Clean, modular, and well-documented Python code
- Reproducible simulation experiments
- Clear visualizations illustrating coordination behavior
- Concise technical report describing system design and evaluation results