Information Technology Reference
In-Depth Information
Figure 7.9 a, b represent the controlled force behaviour (simulated and real system
response) and the feed rate variations for the optimal fuzzy controller (i.e., obtained
applying CE method) when drilling GGG40. In order to suppress the cutting force
increase, the feed rate is decreased gradually as the drilling depth increases, and the
cutting force is quite well regulated at the given setpoint.
7.6 Conclusions
The intelligent tuning of fuzzy and neurofuzzy systems is an active research and
development topic despite the growing and world-wide number of applications
already reported in the literature. In this chapter two complementary strategies are
presented and discussed. The first strategy suggests the use of learning by an evolving
clustering method to setup a transductive neuro-fuzzy inference system. The para-
digm of internal model control can then be applied on the basis of the direct and
inverse models obtained from input-output data. The second strategy is focused on
the most popular fuzzy control system that uses the error and its derivative as main
inputs. An optimization criterion based on cross entropy method is then applied for
optimal tuning of input scaling factors of the fuzzy controller. A rough process model
and a performance index are necessary to apply this method.
The experimental study based on drilling force control is reported. Both strategies
yield good results. If you have input-output data the first strategy is a good choice.
On the other hand, if rough mathematical description of the process and a perfor-
mance index are available, you can run an optimization for optimal tuning of a fuzzy
controller already designed by a template-based method.
Acknowledgments This work was supported by the Ministry of Economy and Competitive-
ness through its DPI2012-35504 CONMICRO research project and the European project 295372
DEMANES. The authors wish to thank the reviewers and the topic editors for their useful sug-
gestions. We also gratefully acknowledge the collaboration of Antony Price in the preparation
of this paper. J. Godoy wants to especially thank the JAE program (Spanish National Research
Council—CSIC) for its support in the development of this work.
References
Bezdek, J. C., Ehrlich, R., & Full, W. (1984). Fcm: The fuzzy c-means clustering algorithm. Com-
puters and Geosciences , 10 , 191-203.
Bonissone, P. P. (2000). Hybrid soft computing systems: Where are we going? Proceedings
of the 14th European Conference on Artificial Intelligence (ECAI 2000) (Berlin, Germany)
(pp. 739-746).
Bonissone, P. P., Badami, V., Chiang, K. H., Khedkar, P. S., Marcelle, K. W., & Schutten, M. J.
(1995). Industrial applications of fuzzy logic at general electric. Proceedings of the IEEE , 83 ,
450-465.
Search WWH ::




Custom Search