Abstract The authors previously proposed a method to test different hypotheses concerning the human motor control strategies in a computationally efficient way. Its efficiency was demonstrated by modeling the human visual servoing of the 1D bouncing ball benchmark, which involves rhythmic interactions with the environment. Three candidate human-inspired control laws were discriminated based on their capacity to stabilize the task, as analyzed by means of Poincaré maps. It was shown that the linear approximation, derived on these equilibrium points, of the human-like controller could be viewed as a state-feedback. Thus, the human-like controller was compared to the Linear Quadratic controller around the equilibrium point. The present study extends the analyzes made in this previous study by taking into account the influence of the arm and neural dynamics on the behavior adaptation to perturbations and task constraints. Particularly, the influence of these dynamics on the task stability is studied. Obtained results contribute to a better understanding of the human motor control strategies and will help to replicate these performances in robotics thanks to a more robust and less model-dependent robotic control architecture.
Dynamic Stability of the Hybrid Ball-bouncing Task
Informatics in Control, Automation and Robotics ; 3 ; 719-738
Lecture Notes in Electrical Engineering ; 495 , 3
18.04.2019
20 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
A Change of Difficulty on "Wall-Bouncing" Task
British Library Online Contents | 2002
|Europäisches Patentamt | 2024
|