Advantages and limitations of reservoir computing on model learning for robot control
(2015) 2nd IROS Workshop on Machine Learning in Planning and Control of Robot Motion- Abstract
- In certain cases analytical derivation of physicsbased models of robots is difficult or even impossible. A potential workaround is the approximation of robot models from
sensor data-streams employing machine learning approaches.
In this paper, the inverse dynamics models are learned by
employing a learning algorithm, introduced in [1], which is
based on reservoir computing in conjunction with self-organized
learning and Bayesian inference. The algorithm is evaluated
and compared to other state of the art algorithms in terms
of generalization ability, convergence and adaptability using
five datasets gathered from four robots in order to investigate
its pros and cons. Results show that the proposed... (More) - In certain cases analytical derivation of physicsbased models of robots is difficult or even impossible. A potential workaround is the approximation of robot models from
sensor data-streams employing machine learning approaches.
In this paper, the inverse dynamics models are learned by
employing a learning algorithm, introduced in [1], which is
based on reservoir computing in conjunction with self-organized
learning and Bayesian inference. The algorithm is evaluated
and compared to other state of the art algorithms in terms
of generalization ability, convergence and adaptability using
five datasets gathered from four robots in order to investigate
its pros and cons. Results show that the proposed algorithm
can adapt in real-time changes of the inverse dynamics model
significantly better than the other state of the art algorithms. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/ec41820b-706f-4c41-9340-a9c47ca51692
- author
- Polydoros, Athanasios ; Nalpantidis, Lazaros and Krüger, Volker LU
- publishing date
- 2015
- type
- Contribution to conference
- publication status
- published
- subject
- conference name
- 2nd IROS Workshop on Machine Learning in Planning and Control of Robot Motion
- conference location
- Hamburg, Germany
- conference dates
- 2015-10-02 - 2015-10-02
- language
- English
- LU publication?
- no
- additional info
- IROS Workshop on Machine Learning in Planning and Control of Robot Motion ; Conference date: 02-10-2015
- id
- ec41820b-706f-4c41-9340-a9c47ca51692
- date added to LUP
- 2019-05-16 21:22:58
- date last changed
- 2019-05-18 02:18:35
@misc{ec41820b-706f-4c41-9340-a9c47ca51692, abstract = {{In certain cases analytical derivation of physicsbased models of robots is difficult or even impossible. A potential workaround is the approximation of robot models from<br/>sensor data-streams employing machine learning approaches.<br/>In this paper, the inverse dynamics models are learned by<br/>employing a learning algorithm, introduced in [1], which is<br/>based on reservoir computing in conjunction with self-organized<br/>learning and Bayesian inference. The algorithm is evaluated<br/>and compared to other state of the art algorithms in terms<br/>of generalization ability, convergence and adaptability using<br/>five datasets gathered from four robots in order to investigate<br/>its pros and cons. Results show that the proposed algorithm<br/>can adapt in real-time changes of the inverse dynamics model<br/>significantly better than the other state of the art algorithms.}}, author = {{Polydoros, Athanasios and Nalpantidis, Lazaros and Krüger, Volker}}, language = {{eng}}, title = {{Advantages and limitations of reservoir computing on model learning for robot control}}, year = {{2015}}, }