Fuzzy neural networks for identification and control of DC drive systems

This paper demonstrates the application of fuzzy neural networks (FNN´s) in identification and control of DC motor drive system. This technique compensates the drawbacks of the fuzzy-logic controllers (FLC) with fixed membership function and quantization levels. The membership function and quantization levels are adapted according to the system operating condition changes. Two FNN are proposed with different learning rates. The first is FNN identifier to provide the sensitivity inference about the drive system changes. The second is FNN controller with adaptive ability to regulate the drive system against the operating condition changes and disturbances. An online backpropagation algorithm is used to achieve both FNN identifier and controller objectives. Experimental setup of the suggested technique is developed of the DC drive system. Comparison between the developed technique and FLC is highlighted and the experimental test results are listed.