ros_dmp

Dynamic Motion Primitives for Learning from Demonstration

This package implements Dynamic Motion Primitives for Learning from Demonstration. This package was developed during the R and D project as part of academics. Detailed report on analysis, implementation and use of this package can be found at https://github.com/abhishek098/r_n_d_report/blob/master/PadalkarAbhishek-%5BRnD%5DReport.pdf .

Reference : Auke Jan Ijspeert, Jun Nakanishi, Heiko Hoffmann, Peter Pastor, and Stefan Schaal. Dynamical movement primitives: learning attractor models for motor behaviors. Neural computation, 25(2):328{373, 2013}.

How to use the package

Clone the repository:

git clone https://github.com/abhishek098/ros_dmp.git

Build the package:

Inside the package

catkin build --this

This package provides two services:

1. Service for learning the DMP
2. Service for genearating motion from already learnt DMP

Geerated motion is published as navigation path on the topic:

/generate_motion_service_node/cartesian_path

And as cartesian trajectory on the topic:

/generate_motion_service_node/cartesian_trajectory

For more details about using these services, have a look at example clients in example folder.

Example clients

Running learn DMP and generate motion clients

Launch the DMP simulation launch file in one terminal using following command.

roslaunch ros_dmp dmp_sim.launch

Run rviz to visualize learned and demonstrated path published on following topics which can be visualized in rviz.

/learn_dmp_service_node/demonstrated_path

/learn_dmp_service_node/demonstrated_path

In another terminal, run learn client script in example folder

python learn_client.py

This script calls the learn DMP service for imitating a artificial trajectory.

Observe the imitated and demonstrated trajectories in rviz.

In another terminal, run generate motion client

python motion_generation_client.py

Genrated motion can be visualized in rviz on following topic:

/generate_motion_service_node/cartesian_path