A Simple, Adaptive Locomotion Toy-System
Jonas Buchli, Auke Jan Ijspeert
- 发表年份
- 2004
- 引用次数
- 37
摘要
In order to successfully transfer biological principles to engineering problems, it is important to study the fundamental properties of biological systems. The goal is to arrive with useful abstractions that are (1) implementable, (2) testable by experiments and sample implementations, (3) retain the, for an engineering point of view, essential good properties of the biological systems. At BIRG, we are interested in the fundamentals of locomotion control and their possible applications to robotics. In this article, we present a simple, adaptive locomotion toy system that is of oscillatory nature. It is composed of two parts: an adaptive controller based on a nonlinear oscillator, and a mechanical system made of two blocks attached by an active and a passive spring. The controller is designed to be robust against perturbations, and to adapt its locomotion control to changing body kinematics or added external load. The tools to develop such a toy-system are a 2-scale nonlinear dynamical system, namely a Hopf oscillator with adaptive frequency, and a understanding of synchronization behavior of oscillators. A further central ingredient that will be discussed is the concept of asymmetric friction forces. We show that the system possesses several critical parameters. It is illustrated that the bifurcations connected with some of these parameters can be identified as non-smooth phase transitions and power law behavior. Links to biology and possible applications to robotics are discussed. 1.
关键词
相关论文
Statistical Learning Theory
Yuhai Wu, Vladimir Vapnik
1999
Artificial intelligence: a modern approach
1995
Applied Nonlinear Control
Jean-Jacques Slotine, Weiping Li
1991
A new optimizer using particle swarm theory
R.C. Eberhart, James Kennedy
2002