Dynamic Movement Primitives on Sawyer

This project implemented the DMP algorithms onto the Sawyer robot in the ROS framwork

Getting Started

The DMP is a method in generating human-like nonlinear and high-dimensional movements, referenced from the paper below for the algorithm implementation (http://www-clmc.usc.edu/publications/P/pastor-ICRA2009.pdf)

Prerequisites

The project is implemented on ubuntu 16.04 with ROS kinetics 1.Create a new ros workspace 2.Git clone the repository into the newly created workspace 3.Build the ros workspace

Running the tests

1.Run the dmp server for the calculation

$ roslaunch dmp_sawyer dmp.launch 

2.Run the data recorder to record robot joint space data, when finished, ctrl+c to exit, ignore errors

$ rosrun dmp_sawyer joints_recorder.py

3.Specify the new initial joint state and goal joint state in the dmpMassage.py script 4.Run the dmpMassage.py to generate DMP plan and move the robot arm accordingly

$ rosrun dmp_sawyer dmpMassage.py

Built With

• rospy - The robotics framework used
• Sawyer - Robot hardware

Authors

• Taozheng Yang - Initial work -