Authors:
Guillaume Avrin
1
;
Maria Makarov
2
;
Pedro Rodriguez-Ayerbe
2
and
Isabelle A. Siegler
3
Affiliations:
1
CentraleSupélec - CNRS - Univ. Paris-Sud, Université Paris-Saclay, Univ. Paris-Sud, Université Paris-Saclay and Université d’Orléans, France
;
2
CentraleSupélec - CNRS - Univ. Paris-Sud and Université Paris-Saclay, France
;
3
Univ. Paris-Sud, Université Paris-Saclay and Université d’Orléans, France
Keyword(s):
Hybrid System, Bouncing Ball, Stability, Impact Map, Biologically-inspired Control.
Related
Ontology
Subjects/Areas/Topics:
Informatics in Control, Automation and Robotics
;
Modeling, Analysis and Control of Discrete-event Systems
;
Modeling, Analysis and Control of Hybrid Dynamical Systems
;
Signal Processing, Sensors, Systems Modeling and Control
;
System Modeling
Abstract:
This interdisciplinary study aims to understand and model human motor control principles using automatic control methods, with possible applications in robotics for tasks involving a rhythmic interaction with the environment. The paper analyses the properties of a candidate model for the visual servoing of the 1D bouncing ball benchmark task in humans. The contributions are twofold as they i/ enable a computationally efficient way of testing hypotheses in human motor control modeling, and ii/ will allow to export and adapt the lessons learned from this modeling of human behavior for more robust and less model-dependent robotic control methods. Three hypotheses about the sensorimotor couplings involved during the task, i.e. three control structures are analyzed from the point of view of task stability by means of Poincaré maps. Obtained results are used to refine the proposed models of sensorimotor couplings. It is shown that the fixed points of the Poincaré maps are stable and that t
he obtained linear approximation, derived on these equilibrium points, can be viewed as a state-feedback. The human-like controller gains are then retrieved with a Linear Quadratic control method, thus showing its inherent robustness.
(More)