Implementation of fuzzy control approach for MIMO robotics systems
Youcef, T.
Yacine, A.
Comput. Sci. & Robotics Lab., Paris-12 Val de Marne Univ., Vitry-sur-Seine, France;
This paper appears in: Robotics, Automation and Mechatronics, 2004 IEEE Conference on
Publication Date: 1-3 Dec. 2004
Volume: 2,
On page(s): 619- 625 vol.2
ISSN:
ISBN: 0-7803-8645-0
INSPEC Accession Number: 8412006
Current Version Published: 2005-06-13
Abstract
This paper proposes an effective approach of fuzzy logic controller (FLC) design and optimization methodology for Cartesian robot control. The FLC is based on a Takagi Sugeno (TS) model. It consists on MISO-controllers subsets decomposition. The FLC optimization methodology is implemented offline and proceeds in three phases: A first set of rules is extracted automatically from training data using rapid prototyping algorithm (RPA). In the second phase, the positioning of all membership functions in the universe of discourse is then optimized by Solis and Wetts (1981) method in conjunction with RPA algorithm. Finally, a stochastic gradient method is implemented to modify the conclusions and then to increase the optimization quality and performances. Once the resulting FLC is generated and optimized, it is implemented, online, in an external position/force control structure and tested on an experimental cell following circular trajectory under force constraints. In this case, a back-propagation (BP) method is implemented. To show the effectiveness of our approach, experimental results are described and discussed.
Index
Terms
Available to subscribers and IEEE members.
References
Available to subscribers and IEEE members.
Citing Documents
Available to subscribers and IEEE members.
Cart
