Dynamic Movement Primitives -A Framework for Motor Control in Humans and Humanoid Robotics

By Stefan Schaal. 2006

Subfields: Neuroscience, 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 Neuroscience

In Robotics & Control

Citations in the corpus

Data-efficient Deep Reinforcement Learning for Dexterous Manipulation (2017, background): “in the form of dynamic movement primitives [30], or in the form of strong teaching signals, e.”
Accelerating Reinforcement Learning with Learned Skill Priors (2020, background): “Prior work has learned skill libraries from data collected by humans [27, 22, 23, 29, 19] or by agents autonomously exploring the world [12, 30].”
Learning Action Representations for Reinforcement Learning (2019, methodology): “In the ﬁeld of robotics, dynamical system based models have been used to construct dynamic movement primitives (DMPs) for continuous control [Ijspeert et al., 2003, Schaal, 2006].”
Movement reproduction and obstacle avoidance with dynamic movement primitives and potential fields (2008, methodology): “We address this problem by using the dynamic movement primitive (DMP) framework, which has been successfully applied to humanoid robotics [3], [4].”
Learning grounded finite-state representations from unstructured demonstrations (2015, background): “LfD has had many successes in teaching robots tennis swings (Schaal, 2003), walking gaits (Nakanishi et al.”
Spatial adaption of robot trajectories based on laplacian trajectory editing (2015, methodology): “Instead of programming every single movement by hand, the robot imitates a demonstrated movement either through physical guidance also known as kinesthetic teaching (Lee and Ott 2011) or through suitable learning and adaption meth-ods (Schaal 2006; Hoffmann et al. 2009).”
A Review of Robot Learning for Manipulation: Challenges, Representations, and Algorithms (2019, methodology): “Dynamic Movement Primitives [326, 288] use a simple spring-mass-damper system to smoothly adjust to new initial and goal locations.”
Learning legged swimming gaits from experience (2015, unknown): “In order to maintain a tractable learning problem complexity, we expose a naturally useful set of low-level controls to the learning mechanism.”
Incremental learning of full body motion primitives and their sequencing through human motion observation (2012, background): “Much of human movement is thought to be composed of motion primitives, which are combined and sequenced to generate more complex behavior (Schaal 2006).”
Unsupervised Learning of Sensorimotor Affordances by Stochastic Future Prediction (2018, background): “Methods for aggregating low-level controls into higher-level representations on which planning and RL can be performed are well-studied in the robotics literature: notable examples include the options framework for hierarchical RL [43, 3], dynamic motion primitives for manipulation [38, 37, 31] and ”
Learning Navigation Subroutines from Egocentric Videos (2019, unknown): “Dynamic Motion Primitives (DMPs) propose a framework for specifying macro actions [9], which can be learned from a demonstration [10] for a speciﬁc macro action.”
Biologically-inspired dynamical systems for movement generation: Automatic real-time goal adaptation and obstacle avoidance (2009, background): “A realization of such dynamical systems are dynamic movement primitives [1], [2], [6].”
How to train your robot with deep reinforcement learning: lessons we have learned (2021, background): “On-policy algorithms such as policy gradient methods (Peters and Schaal 2006; Schulman et al. 2015) have recently become popular owing to their stability and their ability to learn policies for a wide variety of tasks.”
MimicPlay: Long-Horizon Imitation Learning by Watching Human Play (2023, background): “Traditional IL algorithms such as DMP and PrMP [25, 35, 36, 41] enjoy high learning sample efficiency but are limited in their ability to handle high-dimensional observations and settings that require closed-loop control.”
Robot Learning by Demonstration with local Gaussian process regression (2010, methodology): “3Actually we store only the result of the covariance matrix Cholesky decomposition, L, solving (LT )−1(L−1y) [16] to equip the robot with self improvement capabilities.”