Discriminative and adaptive imitation in uni-manual and bi-manual tasks

By Aude Gemma Billard, Sylvain Calinon, Florent Guenter. 2006

225 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering, Human-Computer Interaction

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

In Human-Computer Interaction

Citations in the corpus

An Incremental Self-Organizing Architecture for Sensorimotor Learning and Prediction (2017, background): “A number of other works have also adapted HMMs to the problem of incremental learning of human motion [30][31][32][33].”
Model-based imitation learning by probabilistic trajectory matching (2013, background): “, motion capture [28], kinesthetic teaching [8]), the level at which imitation happens (e.”
On Learning, Representing, and Generalizing a Task in a Humanoid Robot (2007, methodology): “Metric of imitation We proposed in [24] a general formalism for evaluating the reproduction of a task.”
Trajectories and keyframes for kinesthetic teaching: A human-robot interaction perspective (2012, background): “Some also consider the position of the end-eﬀector with respect to the target object of the skill [5, 11].”
Improved human–robot team performance through cross-training, an approach inspired by human team training practices (2015, methodology): “Billard et al. (2006) reduce the dimensionality of the demonstrations to a subset of relevant features using a hidden Markov model (HMM).”
Keyframe-based Learning from Demonstration (2012, background): “Most studies subsample the recorded data with a ﬁxed rate [4, 6].”
Humanoid Robot's Autonomous Acquisition of Proto-Symbols through Motion Segmentation (2006, background): “Billard et al have proposed an imitation framework, in which an humanoid robot can reproduce the specific gesture by memorizing the demonstrated hand path and extracting the relative importance of individual joint’s movement in each given task [6].”
A survey of robot learning from demonstration (2009, unknown): “In kinesthetic teaching, a humanoid robot is not actively controlled but rather its passive joints are moved through desired motions [Billard et al., 2006].”
Incremental learning of full body motion primitives and their sequencing through human motion observation (2012, methodology): “…2006; Ito and Tani 2004; Andry et al. 2004), stochastic models such as hidden Markov models (HMMs) (Inamura et al. 2004; Takano et al. 2006, 2005; Billard et al. 2006, 2004; Asfour et al. 2006), phase oscillators for rhythmic motions (Nakanishi et al. 2004), dynamical models (Ijspeert et al.…”
Dynamical System Modulation for Robot Learning via Kinesthetic Demonstrations (2008, unknown): “The solution to this minimization problem is given by [20]”
Automatic selection of task spaces for imitation learning (2009, methodology): “This basic idea was already evaluated in our previous work [1] and other imitation learning approaches that use probabilistic representations [5], [7], [14].”
GELLO: A General, Low-Cost, and Intuitive Teleoperation Framework for Robot Manipulators (2023, unknown): “This means that the joints and links of GELLO correspond directly to those of the target arm, allowing the user to control the GELLO manipulator as if they were directly controlling the target arm, as in kinesthetic teaching [59].”
Full Paper Reinforcement Learning for Imitating Constrained Reaching Movements (unknown): “This application has been described in detail in Refs [26, 28 ].”
Task-Specific Generalization of Discrete and Periodic Dynamic Movement Primitives (2010, background): “It is also possible to generalize the example movements to new situations by specifying constraints and weighting them with respect to their importance for the task [22].”
Incremental Learning, Clustering and Hierarchy Formation of Whole Body Motion Patterns using Adaptive Hidden Markov Chains (2008, methodology): “When automatically selecting the number of states based on BIC (Billard et al. 2006) or AIC (Kulic et al. 2007a), under 10 states are usually selected for typical human motion patterns such as walking, kicking, punching, etc.”