Module dynamic_movement_primitives 

Dynamic Movement Primitives (DMP)

Learns a trajectory from demonstration data by modelling the forcing function as a weighted sum of Gaussian basis functions. The learned weights can then be used to recreate the trajectory with different start/goal positions or durations.

References:

• https://arxiv.org/abs/2102.03861
• https://www.frontiersin.org/articles/10.3389/fncom.2013.00138/full

Structs

Dmp

A learned DMP model that can reproduce trajectories.

DmpConfig

Configuration for the DMP learner.

DmpTrajectory

Result of trajectory recreation.