Experience-based Fuzzy Control of an Anthropomimetic Robot

@inproceedings{Potkonjak2012ExperiencebasedFC,
  title={Experience-based Fuzzy Control of an Anthropomimetic Robot},
  author={Veljko Potkonjak and Nenad Bascarevic and Predrag Milosavljevic and Kosta Jovanovi{\'c} and Owen Holland},
  booktitle={IJCCI},
  year={2012}
}
This paper aims to present a novel experience-based solution for a black-box control problem, applied to an anthropomimetic robot. The control method is tested on a point to point control problem of a multi-jointed robot arm. The model characteristics – dynamics, kinematics, and control parameters – are considered as unspecified, and therefore we deal with a machine learning approach that follows the cybernetic concept of black-box. The only available data of the system are those obtained from… 
2 Citations

Figures and Tables from this paper

Fictitious Fuzzy-Magnet Concept in Solving Mobile–Robot Target Navigation, Obstacle Avoidance and Garaging Problems
TLDR
This chapter presents a new approach to the modeling of driver skills, based on a fuzzy model and an original virtual fuzzy-magnet concept, and the proposed algorithms are very simple; they are not based on the WMR model.
C U R R I C U L U M V I T A E - Prof. Veljko Potkonjak
TLDR
The Vestibulo-ocular reflex and its correlation with dysequillibrium following hemispheric stroke can be linked with marked circadian fluctuations in childhood and infancy.

References

SHOWING 1-10 OF 12 REFERENCES
Black-box modeling of a 2-DOF manipulator in the image plane using recurrent neurofuzzy networks
TLDR
This work presents a black-box modeling of a 2-DOF planar robot in the image plane using recurrent neural networks with output feedback and a learning law inspired by adaptive observer theory, and proven to be convergent in the parameters and stable in the Lyapunov sense.
Biologically Inspired Control of a Compliant Anthropomimetic Robot
Anthropomimetics copying nature in order to construct and control similar mechanism in technical world is an increasingly attractive research topic. Its core objective is achieving human efficiency
Combining analysis, imitation, and experience-based learning to acquire a concept of reachability in robot mobile manipulation
TLDR
This paper uses capability maps as analytic models of the robot's dexterity to constrain the area in which the robot gathers training data and acquires a human model of reachability from human motion data and uses it to bias exploration.
The Anthropomimetic Principle
Most humanoid robots are essentially conventional robots that fit within the morphological envelope of a human. However, for robots that are intended to help in the understanding of human cognition
The Puller-Follower Control of Compliant and Noncompliant Antagonistic Tendon Drives in Robotic Systems
TLDR
A new control strategy for noncompliant and compliant antagonistic tendon drives based on a biologically inspired puller-follower concept is applied to a succession of increasingly complex single-joint systems, starting with a linear and non Compliant system and ending with a revolute, nonlinearly tendon coupled and compliant system.
MOSAIC Model for Sensorimotor Learning and Control
TLDR
Simulations of an object manipulation task prove that the MOSAIC architecture can learn to manipulate multiple objects and switch between them appropriately and shows generalization to novel objects whose dynamics lie within the polyhedra of already learned dynamics.
BLACK-BOX MODELING WITH STATE-SPACE NEURAL NETWORKS
TLDR
The goal of this paper is to show that there are advantages in using nonlinear state-space models, which constitute a larger class of nonlinear dynamical models, and their corresponding state- space neural predictors, to be potentially more efficient and more parsimonious than their conventional input-output counterparts.
Learning to Combine Motor Primitives Via Greedy Additive Regression
TLDR
It is concluded that learning techniques using local primitives and optimization procedures are viable and potentially important methods for motor control and possibly other domains, and that these techniques deserve further examination by the artificial intelligence and cognitive science communities.
Black-box models from input-output measurements
  • L. Ljung
  • Computer Science
    IMTC 2001. Proceedings of the 18th IEEE Instrumentation and Measurement Technology Conference. Rediscovering Measurement in the Age of Informatics (Cat. No.01CH 37188)
  • 2001
TLDR
Ways to fit black box structures to measured input-output data are described, as well as the more fundamental (statistical) properties of the resulting models.
...
...