Additional methods for the class DiscreteTimeTrajectory

[wei-chen-li]

Is your feature request related to a problem? Please describe.

1. When a function returns a DiscreteTimeTrajectory, the std::vector used for its construction is unnecessarily copied.
2. There is no convenient method to sample from a Trajectory and generate a DiscreteTimeTrajectory.

Describe the solution you'd like
Implement the following three methods in the DiscreteTimeTrajectory class

DiscreteTimeTrajectory(const std::vector<T>& times, std::vector<MatrixX<T>>&& values, const double time_comparison_tolerance)

DiscreteTimeTrajectory(std::vector<T>&& times, std::vector<MatrixX<T>>&& values, const double time_comparison_tolerance)

static DiscreteTimeTrajectory<T> SampleFrom(const Tajectory<T>& traj, double time_period, double time_offset=0)

Additional Considerations

A potential issue with SampleFrom is handling cases where traj.start_time() == -inf or traj.end_time() == inf. Possible solutions include:

• Returning an empty trajectory
• Throwing an exception
• Defining a reasonable default start time and end time

Feedback on the preferred behavior for these edge cases would be appreciated.

[cohnt]

Wouldn't the Trajectory::vector_values method do what you need? I guess this doesn't work for trajectories that are matrix-value yet.