A very simple 2D rocket landing using NL-MPC. Has a bad saw-tooth response behaviour, should be corrected.
Things I aim to do at some point.
- change reference trajectory generation cost
- switch to linear rate based MPC for trajectory tracking
- quantify stability using monte-carlo runs
- make the problem 3D