44,878 Hits in 5.1 sec

Derivative-free online learning of inverse dynamics models [article]

Diego Romeres, Mattia Zorzi, Raffaello Camoriano, Silvio Traversaro, Alessandro Chiuso
2018 arXiv   pre-print
This paper discusses online algorithms for inverse dynamics modelling in robotics.  ...  An extensive experimental study with real data from the right arm of the iCub robot is presented, comparing different model classes and estimation procedures, showing that the proposed 'derivative-free  ...  In this paper we shall be concerned with online learning of so called inverse dynamics models: these models have joint trajectories as inputs and joint torques as outputs.  ... 
arXiv:1809.05074v1 fatcat:yzaz2kazqvdlna7udqigap5nzy

A State Space Modeling Method for Aero-Engine Based on AFOS-ELM

Hongyi Chen, Qiuhong Li, Shuwei Pang, Wenxiang Zhou
2022 Energies  
The structure of the extreme learning machine (ELM) is determined according to the form of the state space model, and the inverse-free ELM algorithm is used to automatically select the appropriate number  ...  Then, according to the analytical equation of the ELM model, the state space model of an aero-engine at each sampling time is obtained by using the partial derivative method.  ...  Then, the SSM is derived by applying the partial derivative method to the analytical model of the AFOS-ELM. (2) The inverse-free extreme learning machine (ELM) is used to initialize the neural network  ... 
doi:10.3390/en15113903 fatcat:rsd5qirm3fbqppop5k2slii6iu

Cascaded Gaussian Processes for Data-efficient Robot Dynamics Learning [article]

Sahand Rezaei-Shoshtari, David Meger, Inna Sharf
2019 arXiv   pre-print
The learned modeling is tested in conjunction with the classical inverse dynamics model (semi-parametric) and on its own (non-parametric) in the context of feed-forward control of the arm.  ...  Motivated by the recursive Newton-Euler formulation, we propose a novel cascaded Gaussian process learning framework for the inverse dynamics of robot manipulators.  ...  Our model learning framework learns offline but predicts the torques online. B.  ... 
arXiv:1910.02291v1 fatcat:idfmp74ubjfcxdx56di3ntnbre

Online-Learning Deep Neuro-Adaptive Dynamic Inversion Controller for Model Free Control [article]

Nathan Lutes and K. Krishnamurthy and Venkata Sriram Siddhardh Nadendla and S. N. Balakrishnan
2021 arXiv   pre-print
The type of controller designed is an adaptive dynamic inversion controller utilizing a modified state observer in a secondary estimation loop to train the network.  ...  The deep neural network learns the entire plant model on-line, creating a controller that is completely model free. The controller design is tested in simulation on a 2 link planar robot arm.  ...  This paper presented a novel model-free dynamic inversion controller featuring a DNN that was trained online using the multiplicative weight update training algorithm.  ... 
arXiv:2107.10383v1 fatcat:dzwr57pl7bblhfmuxv55dfpuge

Online semi-parametric learning for inverse dynamics modeling [article]

Diego Romeres and Mattia Zorzi and Raffaello Camoriano and Alessandro Chiuso
2016 arXiv   pre-print
This paper presents a semi-parametric algorithm for online learning of a robot inverse dynamics model. It combines the strength of the parametric and non-parametric modeling.  ...  The former exploits the rigid body dynamics equa- tion, while the latter exploits a suitable kernel function.  ...  We aim to test the models of Section II for learning online the inverse dynamics of its right arm.  ... 
arXiv:1603.05412v2 fatcat:qjpw22mis5am3ktvfawmxqy67q

Nonparametric Online Learning Control for Soft Continuum Robot: An Enabling Technique for Effective Endoscopic Navigation

Kit-Hang Lee, Denny K.C. Fu, Martin C.W. Leong, Marco Chow, Hing-Choi Fu, Kaspar Althoefer, Kam Yim Sze, Chung-Kwong Yeung, Ka-Wai Kwok
2017 Soft Robotics  
In this study, we propose a generic control framework based on nonparametric, 17 online, as well as local training, in order to learn the inverse model directly, without prior knowledge of the 18 robot's  ...  However, previous model-based control approaches often require simplified geometric 15 assumptions on the soft manipulator, but which could be very inaccurate in the presence of unmodeled 16 external interaction  ...  of the online 35 learning. 21 34inverse model can quickly adapt the inverse mapping upon contact with the external interaction at 36s.  ... 
doi:10.1089/soro.2016.0065 pmid:29251567 pmcid:PMC5734182 fatcat:7oqh6houvzdaxmmncu57bhsg4i

Context Estimation and Learning Control through Latent Variable Extraction: From discrete to continuous contexts

Georgios Petkos, Sethu Vijayakumar
2007 Engineering of Complex Computer Systems (ICECCS), Proceedings of the IEEE International Conference on  
Recent advances in machine learning and adaptive motor control have enabled efficient techniques for online learning of stationary plant dynamics and it's use for robust predictive control.  ...  This work refines the multiple model formalism to bootstrap context separation from context-unlabeled data and enables simultaneous online context estimation, dynamics learning and control based on a consistent  ...  This framework allows us to estimate the context online based only on the learned inverse dynamics models using Markovian filtering.  ... 
doi:10.1109/robot.2007.363634 dblp:conf/icra/PetkosV07 fatcat:mrngukch7na75kx2cy3ypj2bgy

A Survey for Machine Learning-Based Control of Continuum Robots

Xiaomei Wang, Yingqi Li, Ka-Wai Kwok
2021 Frontiers in Robotics and AI  
In this article, we attempt to overview the current state of kinematic/dynamic model-free control schemes for continuum manipulators, particularly by learning-based means, and discuss their similarities  ...  To this end, data-driven modeling strategies making use of machine learning algorithms would be an encouraging way out for the control of soft continuum robots.  ...  Reinforcement Learning With Kinematics/Dynamics Model The policy trained on kinematics/dynamics model-based methods can perform stably, while in model-free methods, the derivation may be large when the  ... 
doi:10.3389/frobt.2021.730330 pmid:34692777 pmcid:PMC8527450 fatcat:p4yeo5jqajfhphzsdbiu746swa

A Robust Neural Fuzzy Petri Net Controller For A Temperature Control System

Ali A. Abed, Abdul Adhem A. Ali, Nauman Aslam, Ali F. Marhoon
2011 Procedia Computer Science  
After offline learning to get the initial weights, the RNFPN is online constructed by concurrent structure/parameter learning.  ...  The RNFPN has many advantages when applied to temperature control plants such as: high learning ability which reduces the controller training time, no a priori knowledge of the plant is required which  ...  These types of computing techniques have many advantages, such as: it is used for nonlinear systems, model-free systems, and good self learning abilities.  ... 
doi:10.1016/j.procs.2011.07.122 fatcat:qko4j342nffshbw27te3ktmabm

OSCAR: Data-Driven Operational Space Control for Adaptive and Robust Robot Manipulation [article]

Josiah Wong, Viktor Makoviychuk, Anima Anandkumar, Yuke Zhu
2021 arXiv   pre-print
In this work, we propose OSC for Adaptation and Robustness (OSCAR), a data-driven variant of OSC that compensates for modeling errors by inferring relevant dynamics parameters from online trajectories.  ...  OSCAR decomposes dynamics learning into task-agnostic and task-specific phases, decoupling the dynamics dependencies of the robot and the extrinsics due to its environment.  ...  This design factorizes the dynamics model into a canonical task-agnostic component learned online from scratch using free-space motion and a constrained task-specific residual, allowing for fast adaptation  ... 
arXiv:2110.00704v1 fatcat:r6kia5op7fbexeldh55e5d74em

Online Model Selection Based on the Variational Bayes

Masa-aki Sato
2001 Neural Computation  
In this article, we derive an online version of the VB algorithm and prove its convergenc e by showing that it is a stochastic approximation for nding the maximum of the free energy.  ...  By combining sequential model selection procedures, the online VB method provides a fully online learning method with a model selection mechanism.  ...  Online Variational Bayes Method Expectation Value of the Free Energy. In this section, we derive an online version of the VB algorithm.  ... 
doi:10.1162/089976601750265045 fatcat:wzqipqgz6be23nhc7hl3cmc7nu

Stable Neuro-Flight-Controller Using Fully Tuned Radial Basis Function Neural Networks

Yan Li, Narasimhan Sundararajan, Paramasivan Saratchandran
2001 Journal of Guidance Control and Dynamics  
Gomi and Kawato’ proposed a feedback-error-learning con- trol strategy, where a Gaussian RBFN is used for online learning of the inverse dynamics of the system.  ...  The RBFN controller, consisting of variable Gaussian functions, uses only online learning to represent the local inverse dynamics of the aircraft system.  ... 
doi:10.2514/2.4793 fatcat:53xwbw4okbdevin5lgrysy4cw4


D.T. Pham, A.A. Fahmy, E.E. Eldukhri
2008 IFAC Proceedings Volumes  
An inductive learning algorithm is applied to generate the required inverse modelling rules from the robot's input/output records.  ...  This paper presents a new systematic adaptive fuzzy neural network for inverse modelling of robot manipulators.  ...  ACKNOWLEDGEMENT The research described in this paper was supported by the EC FP6 Innovative Production Machines and Systems (I*PROMS) Network of Excellence.  ... 
doi:10.3182/20080706-5-kr-1001.00893 fatcat:o3qunfdxtffgpkgwxpaooxtfla

Biologically Plausible Online Principal Component Analysis Without Recurrent Neural Dynamics [article]

Victor Minden, Cengiz Pehlevan, Dmitri B. Chklovskii
2018 arXiv   pre-print
Artificial neural networks that learn to perform Principal Component Analysis (PCA) and related tasks using strictly local learning rules have been previously derived based on the principle of similarity  ...  Further, during these fast dynamics such networks typically "disable" learning, updating synaptic weights only once the fixed-point iteration has been resolved.  ...  ACKNOWLEDGMENTS The authors thank Mariano Tepper and Anirvan Sengupta for useful discussion that contributed to the quality of this manuscript.  ... 
arXiv:1810.06966v2 fatcat:ul7clkmgtrhxtll7mk4wpcg4gm

Learning tracking control with forward models

Botond Bocsi, Philipp Hennig, Lehel Csato, Jan Peters
2012 2012 IEEE International Conference on Robotics and Automation  
We use sparse online Gaussian process inference to obtain a flexible probabilistic forward model and second order optimization to find the inverse mapping.  ...  We propose an adaptive learning algorithm for tracking control of underactuated or non-rigid robots where the physical model of the robot is unavailable.  ...  Finally, the inverse dynamics model is used for computing the torques for the actuated DoFs of the robot.  ... 
doi:10.1109/icra.2012.6224831 dblp:conf/icra/BocsiHCP12 fatcat:wuzem7owwfdgja7xmtjpadiy4a
« Previous Showing results 1 — 15 out of 44,878 results