Tobia Marcucci

Currently we ask `d^n r / d s^n = 0` for `n=2, 3, …` at the starting and ending times. But this is incorrect if the boundary velocities are nonzero.

In the exercise the compass gait `MultibodyPlant` is translated `ToAutoDiffXd` in order to use the `MultibodyPlant` methods to build constraints for an optimization. This however makes `GetInfeasibleConstraints` fail because `MultibodyPlant`...

Make the auto grading look nicer

2

@wilkosch: I think it'd be nice to make the autograding cell at the end of the notebooks look nicer. What do you think? For example, would the addition of some...

IndexError Traceback (most recent call last) in () 7 y_min = np.array([-d]) 8 y_max = np.array([d]) ----> 9 plot_output_trajectory(C, x, h, (y_min, y_max)) 10 plt.show() /Users/tobia/codes/pympc/pympc/plot.pyc in plot_output_trajectory(C, x, h,...

As discussed with @RussTedrake, the following preprocessing stage can be useful before looking for a shortest path in a `GraphOfConvexSets`. Currently we are enforcing the following constraints separately: ``` y_e...

MakeSemidefiniteRelaxation: additional tightening constraints for convex quadratics

11

As by the discussion with @RussTedrake @AlexandreAmice @bernhardpg @iamsavva this morning, we can improve `MakeSemidefiniteRelaxation ` as follows. For any (convex) quadratic constraint of the form `x^T A x