Model Learning with Local Gaussian Process Regression

By Duy Nguyen-Tuong, Matthias W. Seeger, Jan Peters. Adv. Robotics, 2009

345 citations recorded by the source corpus.

Precise models of robot inverse dynamics allow the design of significantly more accurate, energy-efficient and compliant robot control. However, in some cases the accuracy of rigid-body models does not suffice for sound control performance due to unmodeled nonlinearities arising from hydraulic cable dynamics, complex friction or actuator dynamics. In such cases, estimating the inverse dynamics model from measured data poses an interesting alternative. Nonparametric regression methods, such as Gaussian process regression (GPR) or locally weighted projection regression (LWPR), are not as restrictive as parametric models and, thus, offer a more flexible framework for approximating unknown nonlinearities. In this paper, we propose a local approximation to the standard GPR, called local GPR (LGP), for real-time model online learning by combining the strengths of both regression methods, i.e., the high accuracy of GPR and the fast speed of LWPR. The approach is shown to have competitive learning performance for high-dimensional data while being sufficiently fast for real-time learning. The effectiveness of LGP is exhibited by a comparison with the state-of-the-art regression techniques, such as GPR, LWPR and ν-support vector regression. The applicability of the proposed LGP method is demonstrated by real-time online learning of the inverse dynamics model for robot model-based control on a Barrett WAM robot arm.

Subfields: Artificial Intelligence, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/b02229b1b92aced66dcaa6286eaf24b241ae8f7d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

Citations in the corpus

Online Sparse Multi-Output Gaussian Process Regression and Learning (2019, methodology): “The local GP (LGP) technique [48], [49], on the other hand, partitions the streaming data into non-overlapping groups, each of which is exploited to train a separate single-output GP model.”
Mercury: Metro density prediction with recurrent neural network on streaming CDR data (2016, background): “Stochastic Models [5], Artificial Neural Networks (ANNs) [2], [6] Support Vector Machines (SVM) [15], and Gaussian Process [16], [13] are four schools of mod-”
Gradient Weights help Nonparametric Regressors (2012, neutral): “The first two datasets describe the dynamics of 7 degrees of freedom of robotic arms, Barrett WAM and SARCOS [9, 10].”
Approximate real-time optimal control based on sparse Gaussian process models (2014, background): “Gaussian Processes (GPs) are a popular choice for model learning [15], [16], [2], which are extremely data efficient (utilizing available training data in a Bayes-optimal way) and hence are well suited for our needs for learning a dynamics model.”
A Survey on Policy Search for Robotics (2013, methodology): “Therefore, local models, such as LWR or local GP models [55], should be employed if the data sets are huge.”
Task-Specific Generalization of Discrete and Periodic Dynamic Movement Primitives (2010, background): “The release times were also estimated by GPR.”
Learning Physical Collaborative Robot Behaviors From Human Demonstrations (2016, methodology): “process regression [48], GMR exploits the joint probability density function of the data modeled by TP-GMM for deriving the regression.”
Learning to Control a Low-Cost Manipulator using Data-Efficient Reinforcement Learning (2011, methodology): “Typically, the training data for the GP dynamics models are obtained either by motor babbling [9] or by demonstrations [14].”
Model learning for robot control: a survey (2011, methodology): “Other methods combine nonparametric probabilistic regression, such as Gaussian process regression, with the local approaches while exploiting the strength of probabilistic methods for model learning in high dimensions (Nguyen-Tuong and Peters 2009; Ting et al. 2008; Urtasun and Darrell 2008).”
When Gaussian Process Meets Big Data: A Review of Scalable GPs (2018, methodology): “To relieve the problem of MILE, the third thread is to prepartition the input space by clustering techniques and assign points to the experts before training [138], [139].”