Evolution of a non-linear controller

被引:0
|
作者
Moore, WE [1 ]
Parker, R [1 ]
机构
[1] Charles Sturt Univ, Bathurst, NSW 2795, Australia
关键词
recursive neural network; evolutionary algorithm; non-linear control;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The use of new types of actuators such as bimetal strips and McKibben muscles in robotics has become popular. The use of these actuators can have the disadvantage of needing to design controllers for non-linearity's and time dependent parameters, which can arise in such systems. Other problems that occur in robotics are the increased problems associated with systems having large degrees of freedom. This paper presents a method of designing a controller that has the potential to overcome these problems. A nonlinear controller is evolved using an evolutionary algorithm for a two-link system and is shown to be robust and capable of controlling time varying parameter systems. The controller has the potential of being able to be extended to much larger systems with higher degrees of freedom.
引用
收藏
页码:1351 / 1354
页数:4
相关论文
共 50 条