Learning Inverse Kinematics: Reduced Sampling Through Decomposition Into Virtual Robots

Angulo, Vicente Ruiz de; Torras, Carme

doi:10.1109/tsmcb.2008.928232

Cited by 20 publications

(17 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The presented approach offers gains similar to those obtained with the previous ones in [8] and [11], because they all rely on approximating the kinematics of chains having half of the number of joints of the robot. However, there are more criteria to be evaluated in the comparison of these approaches.…”

Section: Discussionmentioning

confidence: 59%

“…This is the same equality used in [11]. The first subchain of the robot must be the same as the last one reverted and transformed to the desired pose.…”

Section: Kinematics Compositionmentioning

confidence: 99%

“…This was the motivation for this work since the number of movements required in our humanoid robot [10] was impractical, even using state-of-theart methods to learn FK. An effective way to palliate this problem are decomposition techniques [8], [11]. Hereby, the robot is virtually divided into two (or more) subchains with fewer DoF each.…”

Section: Introductionmentioning

confidence: 99%

“…A second approach that is general w.r.t. the choice of the robot architecture has been presented in [11]. However, it requires the ability to observe the spatial pose of all subchains' end-effectors at the same time in order to be able to learn.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

General Robot Kinematics Decomposition Without Intermediate Markers

Ulbrich

Angulo

Asfour

et al. 2012

IEEE Trans. Neural Netw. Learning Syst.

Self Cite

View full text Add to dashboard Cite

Abstract-The calibration of serial manipulators with high numbers of degrees of freedom by means of machine learning is a complex and time-consuming task. With the help of a simple strategy, this complexity can be drastically reduced and the speed of the learning procedure can be increased: When the robot is virtually divided into shorter kinematic chains, these subchains can be learned separately and, hence, much more efficiently than the complete kinematics. Such decompositions, however, require either the possibility to capture the poses of all endeffectors of all subchains at the same time, or they are limited to robots that fulfill special constraints. In this work, an alternative decomposition is presented that does not suffer from these limitations. An offline training algorithm is provided in which the composite subchains are learned sequentially with dedicated movements. A second training scheme is provided to train composite chains simultaneously and online. Both schemes can be used together with many machine learning algorithms. In the simulations, an algorithm using Parameterized Self-Organizing Maps (PSOM) modified for online learning and Gaussian Mixture Models (GMM) were chosen to show the correctness of the approach. The experimental results show that, using a two-fold decomposition, the number of samples required to reach a given precision is reduced to twice the square root of the original number.

show abstract

Section: Discussionmentioning

confidence: 59%

“…This is the same equality used in [11]. The first subchain of the robot must be the same as the last one reverted and transformed to the desired pose.…”

Section: Kinematics Compositionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

General Robot Kinematics Decomposition Without Intermediate Markers

Ulbrich

Angulo

Asfour

et al. 2012

IEEE Trans. Neural Netw. Learning Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Instead of finding a direct approximation of inverse kinematics, it is also possible to model the joint probability distribution of end-effector positions and joint angles [5]. Robot manipulator can be decomposed into two or more virtual robot arms to speed up the learning of the inverse kinematics [6]. These solutions are not applicable to mobile robots, because mobile robots are not fixed.…”

Section: Related Workmentioning

confidence: 99%

Use of Learning Methods to Improve Kinematic Models

Priedniece¹,

Ņikitenko²,

Liekna³

et al. 2013

Applied Computer Systems

View full text Add to dashboard Cite

-Kinematic model is the basic aspect in robot design and motion planning. Kinematic models are idealized, however there exist certain specific aspects of particular robot or environment, so that during navigation, the robot can significantly deviate from the planned trajectory. To increase the accuracy of motions, kinematic model can be improved and to achieve that the artificial intelligence methods can be used. In case of fixed base robots different approaches are used to train kinematics, at the same time, for the mobile base robots it proves to be a more complicated task. The reason is that a mobile robot can move unbound with respect to environment thus it is difficult to control the platform without deviation from the target position, which leads to inaccuracy in the position estimate. This paper presents the method meant for improvement of the accuracy of motion of differential drive platform. Genetic programming is used to obtain the wheel velocity function, from which the coefficient, which describes different factor influence on motion, is obtained. As a result, the kinematic model of a particular platform for a particular task is obtained. This method is effective because the developed kinematic model is more specific than the general one.

show abstract

State of the Art

Colomé¹,

Torras²

2019

Springer Tracts in Advanced Robotics

View full text Add to dashboard Cite

Learning Inverse Kinematics: Reduced Sampling Through Decomposition Into Virtual Robots

Cited by 20 publications

References 8 publications

General Robot Kinematics Decomposition Without Intermediate Markers

General Robot Kinematics Decomposition Without Intermediate Markers

Use of Learning Methods to Improve Kinematic Models

State of the Art

Contact Info

Product

Resources

About