Biped dynamic walking using reinforcement learning

By Hamid Benbrahim, Judy A. Franklin. 1997

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

Citations in the corpus

End-to-End Training of Deep Visuomotor Policies (2015, background): “, 2011), locomotion (Benbrahim and Franklin, 1997; Kohl and Stone, 2004; Tedrake et al., 2004; Geng et al., 2006; Endo et al., 2008), and flight (Ng et al.”
Reinforcement learning in robotics: A survey (2013, methodology): “Model-Free Asada et al (1996); Bakker et al (2003); Benbrahim et al (1992); Benbrahim and Franklin (1997); Birdwell and Livingston (2007); Bitzer”
Supervised Reinforcement Learning with Recurrent Neural Network for Dynamic Treatment Recommendation (2018, background): “ combing SL and RL utilize expert behaviors to accelerate reinforcement learning and avoid the risks of actions. Common examples are imitation learning [1, 12, 19, 28, 44] and supervised actor-critic [5, 6, 10]. Given the samples of expert trajectories, imitation learning requires th skill of estima”
Sim-to-Real: Learning Agile Locomotion For Quadruped Robots (2018, background): “Much research has applied reinforcement learning to locomotion tasks [2, 47, 11, 22, 31].”
Policy Networks with Two-Stage Training for Dialogue Systems (2016, background): “Supervised actor-critic architectures following this idea have been proposed in the past (Benbrahim and Franklin, 1997; Si et al., 2004); the actor works together with a human supervisor to gain competence on its task even if the critic’s estimations are poor.”
Machine learning of motor skills for robotics (2007, background): “Examples for learning legged locomotion with reinforcement learning can be found in Morimoto and colleagues (Jun Morimoto, 2000; Morimoto, 2002), Benbrahim & Franklin (Benbrahim & Franklin, 1997; Benbrahim, 1996), Zhang & Meng (C.”
Solving Deep Memory POMDPs with Recurrent Policy Gradients (2007, background): “see [1–5]) as they can deal with continuous states and actions.”
Example-Driven Model-Based Reinforcement Learning for Solving Long-Horizon Visuomotor Tasks (2021, background): “Reinforcement learning (RL) has a rich history of being used for robotic control problems on tasks from locomotion [11, 12, 13] to object manipulation [14, 15, 16].”
HumanPlus: Humanoid Shadowing and Imitation from Humans (2024, background): “…humanoid robots [13, 32, 37, 54, 65, 89], learning-based meth-ods can achieve robust locomotion performance for humanoids [39, 67, 68, 85, 91, 94, 103] and biped robots [4, 42, 44, 45, 83, 84, 93, 100] due to their training on highly randomized environments in simulation and their ability to adapt.”
Abandoning Objectives: Evolution Through the Search for Novelty Alone (2011, methodology): “…Faloutsos, 2009; Hein et al., 2007; Ok and Kim, 2005; Paul, 2003) or oscillation (Hein et al., 2007; Ishiguro et al., 2003), simplifying the task (Benbrahim, 1996; McHale and Husbands, 2004), or initially making the task easier (Allen and Faloutsos, 2009; Ishiguro et al., 2003; Van de Panne and…”
Reinforcement learning of motor skills with policy gradients (2008, methodology): “roll-outs needed. Finally, this approach has yielded the most realworld robot motor learning results ( Benbrahim & Franklin, 1997; Endoetal.,2005;Gullapallietal.,1994;KimuraK Mori et al., 2004; Nakamura et al., 2004; Peters et al., 2005a).”
Policy Gradient Methods for Robotics (2006, background): “…in reinforcement learning is to optimize the policy parameters θ ∈ R K so that the expected return is optimized where a k denote time-step dependent weighting factors, often set to a k = γ k for discounted reinforcement learning (where γ is in [0 , 1] ) or a k = 1 /H for the average reward case.”
Learning CPG Sensory Feedback with Policy Gradient for Biped Locomotion for a Full-Body Humanoid (2005, background): “Our goals in this paper are to acquire a successful walking pattern through learning and to achieve walking with a hardware 3D full-body humanoid robot (Fig. 1). While many attempts have been made to investigate learning algorithms for simulated biped walking, there are only a few successful impleme”
Policy Gradient Methods for Robot Control (2003, background): “policy gradients. Gullapalli demonstrated that policy gradient methods can be used to learn fine manipulation control as it was required for performing the classical ‘peg in a hole’ task [13]; Benbrahim and Franklin showed that policy gradient methods can be used for learning biped walking with inte”
Reinforcement Learning in Continuous State and Action Spaces (2012, background): “However, some promising results have been obtained, for instance in robotics (Benbrahim and Franklin, 1997; Peterset al, 2003).”
Parameter-exploring policy gradients (2010, unknown): “For example, in biped walking Benbrahim and Franklin (1997), object manipulation Peters and Schaal (2006) and other robot control tasks Müller et al.”