Task Dependent Trajectory Learning from Multiple Demonstrations Using Movement Primitives. Vidaković, J., Jerbić, B., Šekoranja, B., Švaco, M., & Šuligoj, F. In Berns, K. & Görges, D., editors, Advances in Service and Industrial Robotics, pages 275–282, Cham, 2020. Springer International Publishing. doi abstract bibtex We propose a model for learning robot task constrained movements from a finite number of observed human demonstrations. The model uses the variation between demonstrations to extract important parts of the movements and reproduce trajectories accordingly. Regions with low variability are reproduced in a constrained manner, while regions with higher variability are approximated more loosely to achieve shorter trajectories. The demonstrations are sampled into states and an initial state sequence is chosen by a minimum distance criterion. Then, a method for state variation analysis is proposed that weights the states according to its similarity to all the other states. A custom function is constructed based on the state-variability information. The time function is then coupled with a state driven dynamical system to reproduce the trajectories. We test the approach on typical two-dimensional task constrained trajectories with constrains on the beginning, in the middle and the end of the movement. The approach is further compared with the case of using a standard exponentially decayed time function.
@inproceedings{vidakovic_task_2020,
address = {Cham},
title = {Task {Dependent} {Trajectory} {Learning} from {Multiple} {Demonstrations} {Using} {Movement} {Primitives}},
isbn = {978-3-030-19648-6},
doi = {10.1007/978-3-030-19648-6_32},
abstract = {We propose a model for learning robot task constrained movements from a finite number of observed human demonstrations. The model uses the variation between demonstrations to extract important parts of the movements and reproduce trajectories accordingly. Regions with low variability are reproduced in a constrained manner, while regions with higher variability are approximated more loosely to achieve shorter trajectories. The demonstrations are sampled into states and an initial state sequence is chosen by a minimum distance criterion. Then, a method for state variation analysis is proposed that weights the states according to its similarity to all the other states. A custom function is constructed based on the state-variability information. The time function is then coupled with a state driven dynamical system to reproduce the trajectories. We test the approach on typical two-dimensional task constrained trajectories with constrains on the beginning, in the middle and the end of the movement. The approach is further compared with the case of using a standard exponentially decayed time function.},
language = {en},
booktitle = {Advances in {Service} and {Industrial} {Robotics}},
publisher = {Springer International Publishing},
author = {Vidaković, Josip and Jerbić, Bojan and Šekoranja, Bojan and Švaco, Marko and Šuligoj, Filip},
editor = {Berns, Karsten and Görges, Daniel},
year = {2020},
keywords = {Constrained trajectory, Dynamical movement primitives, Robot learning from demonstration},
pages = {275--282},
}
Downloads: 0
{"_id":"5gq4Gg5ZE57N6rDWK","bibbaseid":"vidakovi-jerbi-ekoranja-vaco-uligoj-taskdependenttrajectorylearningfrommultipledemonstrationsusingmovementprimitives-2020","author_short":["Vidaković, J.","Jerbić, B.","Šekoranja, B.","Švaco, M.","Šuligoj, F."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"Cham","title":"Task Dependent Trajectory Learning from Multiple Demonstrations Using Movement Primitives","isbn":"978-3-030-19648-6","doi":"10.1007/978-3-030-19648-6_32","abstract":"We propose a model for learning robot task constrained movements from a finite number of observed human demonstrations. The model uses the variation between demonstrations to extract important parts of the movements and reproduce trajectories accordingly. Regions with low variability are reproduced in a constrained manner, while regions with higher variability are approximated more loosely to achieve shorter trajectories. The demonstrations are sampled into states and an initial state sequence is chosen by a minimum distance criterion. Then, a method for state variation analysis is proposed that weights the states according to its similarity to all the other states. A custom function is constructed based on the state-variability information. The time function is then coupled with a state driven dynamical system to reproduce the trajectories. We test the approach on typical two-dimensional task constrained trajectories with constrains on the beginning, in the middle and the end of the movement. The approach is further compared with the case of using a standard exponentially decayed time function.","language":"en","booktitle":"Advances in Service and Industrial Robotics","publisher":"Springer International Publishing","author":[{"propositions":[],"lastnames":["Vidaković"],"firstnames":["Josip"],"suffixes":[]},{"propositions":[],"lastnames":["Jerbić"],"firstnames":["Bojan"],"suffixes":[]},{"propositions":[],"lastnames":["Šekoranja"],"firstnames":["Bojan"],"suffixes":[]},{"propositions":[],"lastnames":["Švaco"],"firstnames":["Marko"],"suffixes":[]},{"propositions":[],"lastnames":["Šuligoj"],"firstnames":["Filip"],"suffixes":[]}],"editor":[{"propositions":[],"lastnames":["Berns"],"firstnames":["Karsten"],"suffixes":[]},{"propositions":[],"lastnames":["Görges"],"firstnames":["Daniel"],"suffixes":[]}],"year":"2020","keywords":"Constrained trajectory, Dynamical movement primitives, Robot learning from demonstration","pages":"275–282","bibtex":"@inproceedings{vidakovic_task_2020,\n\taddress = {Cham},\n\ttitle = {Task {Dependent} {Trajectory} {Learning} from {Multiple} {Demonstrations} {Using} {Movement} {Primitives}},\n\tisbn = {978-3-030-19648-6},\n\tdoi = {10.1007/978-3-030-19648-6_32},\n\tabstract = {We propose a model for learning robot task constrained movements from a finite number of observed human demonstrations. The model uses the variation between demonstrations to extract important parts of the movements and reproduce trajectories accordingly. Regions with low variability are reproduced in a constrained manner, while regions with higher variability are approximated more loosely to achieve shorter trajectories. The demonstrations are sampled into states and an initial state sequence is chosen by a minimum distance criterion. Then, a method for state variation analysis is proposed that weights the states according to its similarity to all the other states. A custom function is constructed based on the state-variability information. The time function is then coupled with a state driven dynamical system to reproduce the trajectories. We test the approach on typical two-dimensional task constrained trajectories with constrains on the beginning, in the middle and the end of the movement. The approach is further compared with the case of using a standard exponentially decayed time function.},\n\tlanguage = {en},\n\tbooktitle = {Advances in {Service} and {Industrial} {Robotics}},\n\tpublisher = {Springer International Publishing},\n\tauthor = {Vidaković, Josip and Jerbić, Bojan and Šekoranja, Bojan and Švaco, Marko and Šuligoj, Filip},\n\teditor = {Berns, Karsten and Görges, Daniel},\n\tyear = {2020},\n\tkeywords = {Constrained trajectory, Dynamical movement primitives, Robot learning from demonstration},\n\tpages = {275--282},\n}\n\n","author_short":["Vidaković, J.","Jerbić, B.","Šekoranja, B.","Švaco, M.","Šuligoj, F."],"editor_short":["Berns, K.","Görges, D."],"key":"vidakovic_task_2020","id":"vidakovic_task_2020","bibbaseid":"vidakovi-jerbi-ekoranja-vaco-uligoj-taskdependenttrajectorylearningfrommultipledemonstrationsusingmovementprimitives-2020","role":"author","urls":{},"keyword":["Constrained trajectory","Dynamical movement primitives","Robot learning from demonstration"],"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"https://api.zotero.org/groups/5666190/collections/C5HS7N6U/items?key=v7LvdlDzMc6wuHelauUyxPKG&format=bibtex&limit=100","dataSources":["jtuwoJtMr7bLoczNh","8b2z9FJLYZT2ZydJQ","JYPZ6Z5hKfiFKzh3P"],"keywords":["constrained trajectory","dynamical movement primitives","robot learning from demonstration"],"search_terms":["task","dependent","trajectory","learning","multiple","demonstrations","using","movement","primitives","vidaković","jerbić","šekoranja","švaco","šuligoj"],"title":"Task Dependent Trajectory Learning from Multiple Demonstrations Using Movement Primitives","year":2020}