Learning to Walk in 20 Minutes

By Russ Tedrake, H. Seung. 2005

128 citations recorded by the source corpus.

Subfields: 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 Reinforcement Learning

In Robotics & Control

Citations in the corpus

Reinforcement learning in robotics: A survey (2013, methodology): “…(2007), Erden and Leblebicioğlu (2008), Martínez-Marín and Duckett (2005), Rosenstein and Barto (2004), Smart and Kaelbling (1998), Tedrake (2004); Tedrake et al. (2005), and Wang et al. (2006) Prior knowledge through task structuring Approach Employed by… Hierarchical Daniel et al. (2012),…”
Evaluation of Policy Gradient Methods and Variants on the Cart-Pole Benchmark (2007, background): “, anthromorphic robot arms [4], [8] or legged robots [2], [6], [13], require highly task specific adaptation of both the policy and the algorithm.”
Learning to Walk via Deep Reinforcement Learning (2018, background): “Real-world applications typically make use of simple and inherently stable robots [14] or low-dimensional gait parameterizations [8, 29, 36], or both [45].”
A Berkeley View of Systems Challenges for AI (2017, background): “RL has a long-standing tradition, with classical success stories including learning to play backgammon at level of the best human players [108], learning to walk [105], and learning basic motor skills [86].”
Policy transfer via modularity and reward guiding (2017, methodology): “Reinforcement learning has been used for tasks such as learning to dress, swing a bat, perform locomotion, stand up from a sitting position, drive a car, and many other tasks [11], [12], [13], [14], [15], [16], [17].”
Learning to Walk in the Real World with Minimal Human Effort (2020, background): “Researchers have investigated applying RL to real robotic systems directly, which is intrinsically free from the sim-to-Fig real gap [48, 40].”
How to train your robot with deep reinforcement learning: lessons we have learned (2021, background): “RL holds the promise to automatically design high performance locomotion controllers (Kohl and Stone 2004; Tedrake et al. 2015; Ha et al. 2018; Hwangbo et al. 2019; Lee et al. 2020a).”
On a Connection between Importance Sampling and the Likelihood Ratio Policy Gradient (2010, background): “They have been applied to a variety of complex real-world reinforcement learning problems, such as hitting a baseball with an articulated arm robot [1], constrained humanoid robotic motion planning [2], and learning gaits for legged robots [3, 4, 5].”
Feedback controller parameterizations for Reinforcement Learning (2011, background): “, [14], [24]), but not in situations in which instability greatly disrupts the learning process1.”
Automated Design of Adaptive Controllers for Modular Robots using Reinforcement Learning (2008, background): “2003), and bipedal walking (Tedrake et al. 2005), as well as simulated mobile manipulation (Martin 2004).”
Towards a theoretical foundation for morphological computation with compliant bodies (2011, background): “For example, Tedrake et al. (2005) showed that the complexity of the task to learn to walk was drastically reduced by the use of a passive walker.”
Learning to Drive a Real Car in 20 Minutes (2007, background): “Learning by interacting with the real system directly has an important advantage: the controller is tailored exactly to the behaviour of the real system at hand instead of a more or less exact model of it.”
Policy Gradient Methods for Robotics (2006, background): “This approach can be highly efﬁcient in simulation optimization of deterministic systems [21] or when a common history of random numbers [22] is being used (the later is known as PEGASUS in reinforcement learning [16]), and can get close to a convergence rate of O (cid:7) I − 1 / 2 (cid:8) [22].”
Learning model-free robot control by a Monte Carlo EM algorithm (2009, methodology): “Several successful model-free RL approaches in robotics are instances of actor-critic or policy gradient algorithms (Ng and Jordan, 2000; Tedrake et al., 2005; Peters and Schaal, 2008a).”
Policy transfer via modularity (2017, methodology): “has been used for tasks such as learning to swing a bat, perform locomotion, stand up from a sitting position, drive a car, and many other tasks [11], [12], [13], [14], [15], [16].”