Learning Motor Skills

Kober, Jens; Peters, Jan

doi:10.1007/978-3-319-03194-1

Cited by 27 publications

(7 citation statements)

References 171 publications

(415 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a matter of fact, after reducing the frequency from 1.0 to 0.5, the robot's body dynamics change and the crawling in experiment 1 cannot be properly presented. The solution to this problem might be to use hybrid learning based on eNAC by counting in the frequency parameters (Kober, 2012 ). Thirdly, the implementation of CPGs is not energy efficient.…”

Section: Discussionmentioning

confidence: 99%

A novel approach to locomotion learning: Actor-Critic architecture using central pattern generators and dynamic motor primitives

Lowe

Ziemke

2014

Front. Neurorobot.

View full text Add to dashboard Cite

In this article, we propose an architecture of a bio-inspired controller that addresses the problem of learning different locomotion gaits for different robot morphologies. The modeling objective is split into two: baseline motion modeling and dynamics adaptation. Baseline motion modeling aims to achieve fundamental functions of a certain type of locomotion and dynamics adaptation provides a “reshaping” function for adapting the baseline motion to desired motion. Based on this assumption, a three-layer architecture is developed using central pattern generators (CPGs, a bio-inspired locomotor center for the baseline motion) and dynamic motor primitives (DMPs, a model with universal “reshaping” functions). In this article, we use this architecture with the actor-critic algorithms for finding a good “reshaping” function. In order to demonstrate the learning power of the actor-critic based architecture, we tested it on two experiments: (1) learning to crawl on a humanoid and, (2) learning to gallop on a puppy robot. Two types of actor-critic algorithms (policy search and policy gradient) are compared in order to evaluate the advantages and disadvantages of different actor-critic based learning algorithms for different morphologies. Finally, based on the analysis of the experimental results, a generic view/architecture for locomotion learning is discussed in the conclusion.

show abstract

Section: Discussionmentioning

confidence: 99%

A novel approach to locomotion learning: Actor-Critic architecture using central pattern generators and dynamic motor primitives

Lowe

Ziemke

2014

Front. Neurorobot.

View full text Add to dashboard Cite

show abstract

“…The required turning angle of a robot for different line angle with varying the distance between sensor array and rear wheel/axel and the distance between sensor array and front wheel/axel was observed [2]. Some of the observations have been given in the data tables (see table 1 and table 2 …”

Section: Experimental Methodologymentioning

confidence: 99%

Optimizing the Turning Velocity in a Line Follower Robot

Hassan¹,

Zahir²

2015

IJCA

View full text Add to dashboard Cite

In this paper, the analysis for data collected for two different line follower robots vehicles are presented. Keeping the hardware identical, the two vehicles were programmed with two different algorithms. When taking a turn, a correlation between its turning angle and the angle of line turn was observed. To find that correlation, experimental data was collected after varying critical parameters. After that, a mathematical representation was derived to calculate the required turning angle for any line angle.

show abstract

“…The idea to use Artificial Neural Networks (ANN) in the control handling systems is well established [6][7][8][9].The central approach applied in this paper is to use ANNs to predict error magnitudes and use them in plant control for error compensation [10]. The ANN is used to approximate linear, as well as non-linear, multi-dimensional influencing variables or combinations of influencing variables.…”

Section: Improvement Of a Pick And Place Process Using An Artificial mentioning

confidence: 99%

Development of an Automated Assembly Process Supported With an Artificial Neural Network

Bobka

Heyn²,

Henningson³

et al. 2018

Journal of Machine Engineering

View full text Add to dashboard Cite

A central problem in automated assembly is the ramp-up phase. In order to achieve the required tolerances and cycle times, assembly parameters must be determined by extensive manual parameter variations. Therefore, the duration of the ramp-up phase represents a planning uncertainty and a financial risk, especially when high demands are placed on dynamics and precision. To complete this phase as efficiently as possible, comprehensive planning and experienced personnel are necessary. In this paper, we examine the use of machine learning techniques for the ramp-up of an automated assembly process. Specifically we use a deep artificial neural network to learn process parameters for pick-and-place operations of planar objects. We describe how the handling parameters of an industrial robot can be adjusted and optimized automatically by artificial neural networks and examine this approach in laboratory experiments. Furthermore, we test whether an artificial neural network can be used to optimize assembly parameters in process as an adaptive process controller. Finally, we discuss the advantages and disadvantages of the described approach for the determination of optimal assembly parameters in the ramp-up phase and during the utilization phase.

show abstract

Learning Motor Skills

Cited by 27 publications

References 171 publications

A novel approach to locomotion learning: Actor-Critic architecture using central pattern generators and dynamic motor primitives

A novel approach to locomotion learning: Actor-Critic architecture using central pattern generators and dynamic motor primitives

Optimizing the Turning Velocity in a Line Follower Robot

Development of an Automated Assembly Process Supported With an Artificial Neural Network

Contact Info

Product

Resources

About