A robust fuzzy credit assignment cerebellar model articulation controller (FCA-CMAC) with fast learning applied to control system

Authors:
Yan-Ping Wang;Zne-Jung Lee;Shun-Feng Su
Affiliations:
Dept. of Electrical Engineering, National Taiwan University of Science and Technology;Dept. of Information Management, Kang-Ning Junior College of Medical Care and Management, Nei-Hu, Taipei, Taiwan, R.O.C.;Dept. of Electrical Engineering, National Taiwan University of Science and Technology
Venue:
ICECS'03 Proceedings of the 2nd WSEAS International Conference on Electronics, Control and Signal Processing
Year:
2003

Citing 2
Cited 0

Combining evolution with credit apportionment: a new learning algorithm for neural nets

Neural Networks
Neural Networks for Statistical Modeling

Neural Networks for Statistical Modeling

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, a robust fuzzy credit assignment CMAC (FCA-CMAC) is proposed to speed up the learning ability and to increase the robust capability of CMAC. The FCA-CMAC is to combine the fuzzy logic concept and credit assignment ideas to provide fast and accurate learning for CMAC. With fuzzified blocks, FCA-CMAC can increase its precision and resolution; Meanwhile, the learning approach is to replace the logic AND and OR operations in the CMAC with the commonly used t-norm and t-conorm operations, respectively. Furthermore, the calculated errors are assigned proportional to the inverse of learning times, which are viewed as the creditability of addressed hypercubes. Besides, we also embed the robust learning algorithm (i.e. M-estimators) and tune the learning constant in the CMAC learning algorithm to control system of an inverse dynamic plant. From simulation results, it shows that the proposed algorithm of FCA-CMAC has better learning capability and performance than other CMACs.