Adaptive neuro-fuzzy control of a flexible manipulator
Adaptive neuro-fuzzy control of a flexible manipulator
Adaptive neuro-fuzzy control of a flexible manipulator
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Mechatronics Journal., Volume 15, Issue 10, Pages 1159-1320 (December 2005)<br />
<strong>Adaptive</strong> <strong>neuro</strong>-<strong>fuzzy</strong> <strong>control</strong> <strong>of</strong> a <strong>flexible</strong> <strong>manipulator</strong><br />
By: Lianfang Tian, Curtis Collins<br />
Abstract<br />
This paper describes an adaptive <strong>neuro</strong>-<strong>fuzzy</strong> <strong>control</strong> system for <strong>control</strong>ling a <strong>flexible</strong> <strong>manipulator</strong><br />
with variable payload. The <strong>control</strong>ler proposed in this paper is comprised <strong>of</strong> a <strong>fuzzy</strong> logic <strong>control</strong>ler (FLC)<br />
in the feedback configuration and two dynamic recurrent neural networks in the forward path. A dynamic<br />
recurrent identification network (RIN) is used to identify the output <strong>of</strong> the <strong>manipulator</strong> system, and a<br />
dynamic recurrent learning network (RLN) is employed to learn the weighting factor <strong>of</strong> the <strong>fuzzy</strong> logic. It<br />
is envisaged that the integration <strong>of</strong> <strong>fuzzy</strong> logic and neural network based-<strong>control</strong>ler will encompass the<br />
merits <strong>of</strong> both technologies, and thus provide a robust <strong>control</strong>ler for the <strong>flexible</strong> <strong>manipulator</strong> system. The<br />
<strong>fuzzy</strong> logic <strong>control</strong>ler, based on <strong>fuzzy</strong> set theory, provides a means for converting a linguistic <strong>control</strong><br />
strategy into <strong>control</strong> action and <strong>of</strong>fering a high level <strong>of</strong> computation. On the other hand, the ability <strong>of</strong> a<br />
dynamic recurrent network structure to model an arbitrary dynamic nonlinear system is incorporated to<br />
approximate the unknown nonlinear input–output relationship using a dynamic back propagation learning<br />
algorithm. Simulations for determining the number <strong>of</strong> modes to describe the dynamics <strong>of</strong> the system and<br />
investigating the robustness <strong>of</strong> the <strong>control</strong> system are carried out. Results demonstrate the good<br />
performance <strong>of</strong> the proposed <strong>control</strong> system.
Keywords: Flexible <strong>manipulator</strong>; Fuzzy logic; Dynamic recurrent neural network; Neuro-<strong>fuzzy</strong><br />
<strong>control</strong>ler