Dynamical System Modulation for Robot Learning via Kinesthetic Demonstrations

By Micha Hersch, Florent Guenter, Sylvain Calinon, Aude Gemma Billard. 2008

288 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Control and Systems Engineering, Reinforcement Learning, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Control and Systems Engineering

In Reinforcement Learning

In Robotics & Control

Citations in the corpus

Value Iteration Networks (2016, background): “, 2007; Nowozin & Lampert, 2011) are important for learning from demonstration in robotics (Hersch et al., 2008; Ratliff et al., 2009; Sung et al., 2015), activity forcasting (Kitani et al.”
BM: An iterative algorithm to learn stable non-linear dynamical systems with Gaussian mixture models (2010, methodology): “Existing approaches to the statistical estimation of f̂ use either Gaussian Process Regression (GPR) [10], Gaussian Mixture Regression (GMR) [4] or Locally Weighted Projection Regression (LWPR) [11].”
Novel Interaction Strategies for Learning from Teleoperation (2012, background): “In LfD, demonstrations are often represented as arm joint trajectories and/or end-effector path [8, 12].”
Task-Specific Generalization of Discrete and Periodic Dynamic Movement Primitives (2010, background): “is described in [8], where the final trajectory is generated as a linear combination of attractor dynamics and Gaussian mixture models that were learned from example trajectories.”
A Dynamical System-based Approach to Modeling Stable Robot Control Policies via Imitation Learning (2012, methodology): “During the last decade, the IL-based approach to estimate DS model of robot motions has become increasingly popular as it offers a mechanism to exploit the advantages of both robot learning and DS modeling (Dixon & Khosla, 2004a; Kober, Mulling, et al., 2010; Ude et al., 2010; Hersch et al., 2008; K”
Task adaptation through exploration and action sequencing (2009, background): “Such problems can be avoided by kinesthetic guiding [3], where the robot arm is led through the action by a human teacher, but this method is not applicable to every robot.”
Learning and Reproduction of Gestures by Imitation (2010, unknown): “In [23], we considered a second-order model.”
Learning Stable Nonlinear Dynamical Systems With Gaussian Mixture Models (2011, methodology): “DS have been advocated as a powerful alternative to modeling robot motions [5], [14].”
Learning to search: Functional gradient techniques for imitation learning (2009, background): “Additionally, (Calinon et al., 2007) and (Hersch et al., 2008) have developed techniques that allow generalization to alternative goals within a given domain.”
Imitation learning of globally stable non-linear point-to-point robot motions using nonlinear programming (2010, background): “Thus parameters (cid:18) need to be determined that lead to an estimate ^ f ( ξ ; (cid:18) ) with a single asymptotically stable attractor.”
Keyframe-based Learning from Demonstration (2012, background): “Demonstrations often consist of arm joint and/or end-effector trajectories [9, 15].”
A tutorial on task-parameterized movement learning and retrieval (2015, methodology): “as an autonomous system with ξ=[x, ẋ⊤] ⊤ , by learning P(x, ẋ) with a GMM, with x and ẋ representing position and velocity of the system (either in task space or joint space), and by retrieving iteratively during reproduction a series of velocity commands by estimating P(ẋ|x) with GMR [38, 39, 17, 4”
Statistical Learning by Imitation of Competing Constraints in Joint Space and Task Space (2009, methodology): “, by encapsulating position and velocity in GMM, and by retrieving recursively through GMR a velocity command, given the current position) [34, 35].”
Trajectories and keyframes for kinesthetic teaching: A human-robot interaction perspective (2012, background): “Kinesthetic teaching may be challenging for everyday users who do not have experience manipulating a robot arm with many degrees of freedom.”