Learning to Control an Octopus Arm with Gaussian Process Temporal Difference Methods

By Yaakov Engel, Peter Szabó, Dmitry Volkinshtein. 2005

75 citations recorded by the source corpus.

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): “Octopus arm control involves moving the tip of a flexible arm to a goal position (Engel et al., 2005).”
Convex Optimization: Algorithms and Complexity (2015, unknown): “BAYESIAN MULTI-TASK REINFORCEMENT LEARNING115 adopt the GPTD value function model ([Engel et al., 2005a], also see §5.1.1) for each task, model the distribution over the value functions using an HBM, and develop solutions to the following problems: (i) joint learning of the value functions…”
Learning Neural Network Policies with Guided Policy Search under Unknown Dynamics (2014, methodology): “Octopus arm control involves moving the tip of a flexible arm to a goal position [6].”
A Review of Robot Learning for Manipulation: Challenges, Representations, and Algorithms (2019, methodology): “This representational paradigm has been successful in manipulation tasks in which very little task knowledge is given a priori to the robot, such as an 88dimensional octopus arm control problem in which Gaussian Process Temporal Difference learning is used [100], and 50-dimensional control of a SARC”
Learning GP-BayesFilters via Gaussian process latent variable models (2009, background): “Over the last years, Gaussian processes (GPs) have been applied with great success to robotics tasks such as reinforcement learning (Engel et al. 2006) and learning of prediction and observation models (Ferris et al. 2006; Ko et al. 2007; Plagemann et al. 2007).”
Autonomous blimp control using model-free reinforcement learning in a continuous state and action space (2007, background): “Their approach was successfully applied to the problem of learning complex manipulation policies [4].”
Bayesian Reinforcement Learning (2010, methodology): “Finally, the GPTD model can be used to derive a SARSA-type algorithm, called GPSARSA (Engel et al, 2005a; Engel, 2005), in which state-action values are estimated using GPTD and policies are improved by a ε -greedily strategy while slowly decreasing ε toward 0.”
Control Strategies for Soft Robotic Manipulators: A Survey (2018, unknown): “In the domain of reinforcement learning, a simulated multisegmented dynamical planar model of the octopus arm was developed in Ref.(68) The authors then addressed the task of reaching a point by modeling the problem as a Hidden Markov Model that was solved online through a nonparametric Gaussian tem”
Autonomous reinforcement learning with experience replay. (2013, methodology): “The action space consists of a set of 6 actions each of which pre-defines the activation level in the arm’smuscles, as used in Engel et al. (2005).”
GP-BayesFilters: Bayesian filtering using Gaussian process prediction and observation models (2008, background): “In (Engel et al. 2006), Gaussian processes are applied to reinforcement learning for mobile robots.”
Stability of Controllers for Gaussian Process Forward Models (2016, unknown): “Many successful applications impressively demonstrate the efficiency of this approach (Deisenroth et al., 2015; Engel et al., 2006; Pan & Theodorou, 2014; Kocijan et al., 2004; Klenske et al., 2013; Maciejowski & Yang, 2013; Nguyen-Tuong & Peters, 2011).”
Learning state representation for deep actor-critic control (2016, methodology): “The octopus arm [23] consists of 13 segments, each of which contains multiple muscles.”
Deterministic Policy Gradient Algorithms (2014, methodology): “Finally, we tested our algorithms on an octopus arm (Engel et al., 2005) task.”