A tutorial on task-parameterized movement learning and retrieval

Calinon, Sylvain

doi:10.1007/s11370-015-0187-9

Cited by 333 publications

(364 citation statements)

References 95 publications

Supporting

Mentioning

361

Contrasting

Unclassified

Order By: Relevance

“…With this probabilistic form of DMP, the basis functions can also be learned from the demonstrations. GMM/GMR have been successful in representing sets of demonstrated examples that are timeindexed, this way using time as the Gaussian conditioning variable to perform the regression, see Calinon (2016) for an overview. Such systems are often used for learning models from a set of demonstrations but are somewhat restrictive in their generalization capability.…”

Section: Related Workmentioning

confidence: 99%

“…Such systems are often used for learning models from a set of demonstrations but are somewhat restrictive in their generalization capability. An extension to the traditional GMM-based learning approach is the parametrization of the problem with a set of different coordinate systems (Calinon 2016). In this setting, learning is performed in multiple coordinate systems, whose information is fused through products of Gaussians used to generate the final motion.…”

Section: Related Workmentioning

confidence: 99%

“…Right, the developed ROV with prototype manipulators and prototype underwater vision system visible between the arms (LQT) with a double integrator system. Our approach leverages the key advantages of two models, the flexibility of a task-parametrized mixture model encoding (Calinon 2016) alongside with the ability of the HSMM to accurately capture motion duration and generate novel trajectories.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Learning from demonstration for semi-autonomous teleoperation

Havoutis

Calinon

2018

Auton Robot

Self Cite

View full text Add to dashboard Cite

Teleoperation in domains such as deep-sea or space often requires the completion of a set of recurrent tasks. We present a framework that uses a probabilistic approach to learn from demonstration models of manipulation tasks. We show how such a framework can be used in an underwater ROV teleoperation context to assist the operator. The learned representation can be used to resolve inconsistencies between the operator's and the robot's space in a structured manner, and as a fall-back system to perform previously learned tasks autonomously when teleoperation is not possible. We evaluate our framework with a realistic ROV task on a teleoperation mock-up with a group of volunteers, showing a significant decrease in time to complete the task when our approach is used. In addition, we illustrate how the system can execute previously learned tasks autonomously when the communication with the operator is lost.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning from demonstration for semi-autonomous teleoperation

Havoutis

Calinon

2018

Auton Robot

Self Cite

View full text Add to dashboard Cite

show abstract

“…We assume that the demonstrations are available under various contexts s i . In this case, we can model the conditional distribution of the demonstrated trajectories given the context in order to generalize the demonstrated trajectories to new situations [11], [24], [25]. Here, we use Locally Weighted Regression (LWR) to model this distribution [26], [27].…”

Section: Modelling Demonstrated Trajectory Distributionsmentioning

confidence: 99%

A learning-based shared control architecture for interactive task execution

Abi-Farraj

Osa

Peters

et al. 2017

2017 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Abstract-Shared control is a key technology for various robotic applications in which a robotic system and a human operator are meant to collaborate efficiently. In order to achieve efficient task execution in shared control, it is essential to predict the desired behavior for a given situation or context to simplify the control task for the human operator. To do this prediction, we use Learning from Demonstration (LfD), which is a popular approach for transferring human skills to robots. We encode the demonstrated behavior as trajectory distributions and generalize the learned distributions to new situations. The goal of this paper is to present a shared control framework that uses learned expert distributions to gain more autonomy. Our approach controls the balance between the controller's autonomy and the human preference based on the distributions of the demonstrated trajectories. Moreover, the learned distributions are autonomously refined from collaborative task executions, resulting in a master-slave system with increasing autonomy that requires less user input with an increasing number of task executions. We experimentally validated that our shared control approach enables efficient task executions. Moreover, the conducted experiments demonstrated that the developed system improves its performances through interactive task executions with our shared control.

show abstract

“…e optimal trajectory can be retrieved iteratively using dynamic programming [5,9], or in batch form by solving a large regularised least squares problem. Here we describe the la er, which is more compact and allows a straightforward probabilistic interpretation of the result.…”

Section: Least Squares Solutionmentioning

confidence: 99%

Dynamic Graffiti Stylisation with Stochastic Optimal Control

Berio

Calinon

Leymarie

2017

Proceedings of the 4th International Conference on Movement Computing

Self Cite

View full text Add to dashboard Cite

We present a method for the interactive generation of stylised letters, curves and motion paths that are similar to the ones that can be observed in art forms such as gra ti and calligraphy. We de ne various stylisations of a le er form over a common geometrical structure, which is given by the spatial layout of a sparse sequence of targets. Di erent stylisations are then generated by optimising the trajectories of a dynamical system that tracks the target sequence.e evolution of the dynamical system is computed with a stochastic formulation of optimal control, in which each target is de ned probabilistically as a multivariate Gaussian. e covariance of each Gaussian explicitly de nes the variability as well as the curvilinear evolution of trajectory segments. Given this probabilistic formulation, the optimisation procedure results in a trajectory distribution rather than a single path. It is then possible to stochastically sample from the distribution an in nite number of dynamically and aesthetically consistent trajectories which mimic the variability that is typically observed in human drawing or writing. We further demonstrate how this system can be used together with a simple user interface in order to explore di erent stylisations of interactively or procedurally de ned le ers. CCS CONCEPTS•Computing methodologies →Procedural animation; •Humancentered computing →Gestural input; KEYWORDSHuman hand-writing movement modeling; procedural calligraphy; gra ti and tags generation; model predictive control; stochastic optimal control; smoothing splines; iconic and kinemic le er forms.

show abstract

A tutorial on task-parameterized movement learning and retrieval

Cited by 333 publications

References 95 publications

Learning from demonstration for semi-autonomous teleoperation

Learning from demonstration for semi-autonomous teleoperation

A learning-based shared control architecture for interactive task execution

Dynamic Graffiti Stylisation with Stochastic Optimal Control

Contact Info

Product

Resources

About