Aim
Start a project from a known working robot baseline, then investigate one bounded question.
Time: about 120 minutes. Start with the linked lectures and the course notation contract. The reference program is complete and runnable; your work is to explain it, modify it, and test the results.
Run from this repository
uv sync --locked
uv run --locked python practices/p07-project/run.py
The program writes configuration, MuJoCo version, and metrics to outputs/p07-project/metrics.json. Experiments with a time history also write trajectory.csv. These are generated results and are not source material.
Use a separate configuration and output directory for each variation:
uv run --locked python practices/p07-project/run.py --config practices/p07-project/config.yml --output outputs/p07-project/baseline
Keep the original configuration as the baseline. Record all changes in your report; avoid changing several physical and numerical parameters at once.
Tasks
- Run this harness to reproduce the cart-pole and quadrotor LQR baseline from Practice 3.
- Choose one task and define success, failure, actuator limits, and evaluation conditions before modification.
- Implement one controller change and one ablation. Keep paired test cases.
- Submit the final project brief's report, commands, configuration, metrics, and a short demonstration.
Expected checks
The supplied harness is a starting baseline, not a completed project. Your assessed contribution must answer a new question and include failures and limitations.
Submission
Submit your YAML configuration, a short explanation of the equations and assumptions, labeled plots or numerical comparisons with units, and a paragraph explaining a failure or limitation. Include the exact run command. The weekly practice rubric awards 40% for correctness and checks, 30% for the controlled comparison, 20% for interpretation, and 10% for reproducibility.
Extension
Use a contact-rich robot only after a small CPU experiment runs reliably. Simulator success alone does not establish hardware safety.