Abstract
This paper develops an enhanced teaching interface tested on both a Baxter robot and a KUKA iiwa robot. Movements are
collected from a human demonstrator by using a Kinect v2 sensor, and then the data is sent to a remote PC for the teleoperation
with Baxter. Meanwhile, data is saved locally for the playback process of the Baxter. The dynamic movement primitive
(DMP) is used to model and generalize the movements. In order to learn from multiple demonstrations accurately, dynamic
time warping (DTW), is used to pretreat the data recorded by the robot platform and Gaussian mixture model (GMM), aiming
to generate multiple patterns after the teaching process, are employed for the calculation of the DMP. Then the Gaussian
mixture regression (GMR) algorithm is applied to generate a synthesized trajectory with smaller position errors in 3D space.
This proposed approach is tested by performing two tasks on a KUKA iiwa and a Baxter robot.
collected from a human demonstrator by using a Kinect v2 sensor, and then the data is sent to a remote PC for the teleoperation
with Baxter. Meanwhile, data is saved locally for the playback process of the Baxter. The dynamic movement primitive
(DMP) is used to model and generalize the movements. In order to learn from multiple demonstrations accurately, dynamic
time warping (DTW), is used to pretreat the data recorded by the robot platform and Gaussian mixture model (GMM), aiming
to generate multiple patterns after the teaching process, are employed for the calculation of the DMP. Then the Gaussian
mixture regression (GMR) algorithm is applied to generate a synthesized trajectory with smaller position errors in 3D space.
This proposed approach is tested by performing two tasks on a KUKA iiwa and a Baxter robot.
| Original language | English |
|---|---|
| Pages (from-to) | 110-121 |
| Number of pages | 12 |
| Journal | International Journal of Intelligent Robotics and Applications |
| Volume | 2 |
| Issue number | 1 |
| DOIs | |
| Publication status | Published - 8 Mar 2018 |