Incremental online sparsification for model learning in real-time robot control

Nguyen-Tuong, Duy; Peters, Jan

doi:10.1016/j.neucom.2010.06.033

Cited by 51 publications

(30 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…two datasets from the same model but having different density distributions will be classified as two different models. The authors of [16], [17] proposed an online sparsification method by using an independency measure to control the sample pruning. This method nevertheless suffers from high computational demand and memory overhead since for a dictionary of saved data points of size n, a matrix of size (n − 1) has to be saved for every sample.…”

Section: A Related Workmentioning

confidence: 99%

A Framework for On-line Learning of Underwater Vehicles Dynamic Models

Wehbe

Hildebrandt

Kirchner

2019

2019 International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Learning the dynamics of robots from data can help achieve more accurate tracking controllers, or aid their navigation algorithms. However, when the actual dynamics of the robots change due to external conditions, on-line adaptation of their models is required to maintain high fidelity performance. In this work, a framework for on-line learning of robot dynamics is developed to adapt to such changes. The proposed framework employs an incremental support vector regression method to learn the model sequentially from data streams. In combination with the incremental learning, strategies for including and forgetting data are developed to obtain better generalization over the whole state space. The framework is tested in simulation and real experimental scenarios demonstrating its adaptation capabilities to changes in the robot's dynamics.

show abstract

Section: A Related Workmentioning

confidence: 99%

A Framework for On-line Learning of Underwater Vehicles Dynamic Models

Wehbe

Hildebrandt

Kirchner

2019

2019 International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

show abstract

“…Thus, we use measures that allow the robot to quantify where in the search domain there exists useful data that needs to be collected. While there exists many measures that can select important data subject to a learning task, we use a measure of linear independence [7,19,20]. This measure is often used in sparse Gaussian processes [7,20] where a data set D = {x i , y i } M i=1 is comprised of M input measurements x i ∈ R v and M output measurements y i ∈ R c such that each data point maximizes the measure of linear independence.…”

Section: Definition 2 a Dynamical Systemmentioning

confidence: 99%

“…While there exists many measures that can select important data subject to a learning task, we use a measure of linear independence [7,19,20]. This measure is often used in sparse Gaussian processes [7,20] where a data set D = {x i , y i } M i=1 is comprised of M input measurements x i ∈ R v and M output measurements y i ∈ R c such that each data point maximizes the measure of linear independence. We use this measure of independence, also known as a measure of importance, to create a distribution for which the robot will provide area coverage in the search domain for active data collection.…”

Section: Definition 2 a Dynamical Systemmentioning

confidence: 99%

“…We then generate a LQR controller with the approximate dynamics which we use as the equilibrium policy, The input data we collect is the state x(t i ) = x i and control u(t i ) = u i and the output data is (x i+1 − x i )/(t i+1 − t i ) which approximates the function ∆x ∆t ≈ẋ = f (x, u) [24]. An incremental sparse Gaussian process [7] with a radial basis function kernel is used to generate a learned model of the dynamics using a data set of M = 80 and to specify the importance measure (3). Figure 4 (a) and Figure 4 (b) illustrates the modeling error and the minimum importance value within the data set using our method and the equilibrium policy with uniformly added added noise at 33% of the saturation limit.…”

Section: Simulated Examplesmentioning

confidence: 99%

See 1 more Smart Citation

Active Area Coverage from Equilibrium

Abraham¹,

Prabhakar²,

Murphey

2019

EasyChair Preprints

View full text Add to dashboard Cite

This paper develops a method for robots to integrate stability into actively seeking out informative measurements through coverage. We derive a controller using hybrid systems theory that allows us to consider safe equilibrium policies during active data collection. We show that our method is able to maintain Lyapunov attractiveness while still actively seeking out data. Using incremental sparse Gaussian processes, we define distributions which allow a robot to actively seek out informative measurements. We illustrate our methods for shape estimation using a cart double pendulum, dynamic model learning of a hovering quadrotor, and generating galloping gaits starting from stationary equilibrium by learning a dynamics model for the half-cheetah system from the Roboschool environment.

show abstract

“…Another model learning method Sparse Incremental Gaussian Process Regression (SI-GPR) was developed and demonstrated to learn and adapt to changing dynamics on a 7-DoF rigid body robot arm [10]. This method used a learned dynamics model as a feed forward term and used a PD feedback controller to stabilize the system.…”

Section: Introductionmentioning

confidence: 99%

Locally Weighted Learning Model Predictive Control for nonlinear and time varying dynamics

Lehnert

Wyeth

2013

2013 IEEE International Conference on Robotics and Automation

View full text Add to dashboard Cite

This paper proposes an online learning control system that uses the strategy of Model Predictive Control (MPC) in a model based locally weighted learning framework. The new approach, named Locally Weighted Learning Model Predictive Control (LWL-MPC), is proposed as a solution to learn to control robotic systems with nonlinear and time varying dynamics. This paper demonstrates the capability of LWL-MPC to perform online learning while controlling the joint trajectories of a low cost, three degree of freedom elastic joint robot. The learning performance is investigated in both an initial learning phase, and when the system dynamics change due to a heavy object added to the tool point. The experiment on the real elastic joint robot is presented and LWL-MPC is shown to successfully learn to control the system with and without the object. The results highlight the capability of the learning control system to accommodate the lack of mechanical consistency and linearity in a low cost robot arm.

show abstract

Incremental online sparsification for model learning in real-time robot control

Cited by 51 publications

References 14 publications

A Framework for On-line Learning of Underwater Vehicles Dynamic Models

A Framework for On-line Learning of Underwater Vehicles Dynamic Models

Active Area Coverage from Equilibrium

Locally Weighted Learning Model Predictive Control for nonlinear and time varying dynamics

Contact Info

Product

Resources

About