A method is proposed for parameter tuning of PID controllers. Unlike the traditional techniques, the proposed tuning procedure is expressed in terms of fuzzy rules, in which the input variables are the error signal and its derivative, while the output variables are the PID gains. A genetic algorithm (GA) is then used to search for the optimal PID gains that will minimize the integral of the squared error. Meanwhile, to prevent damaging the plant due to abrupt changes in PID gains, a trained neural network will also be used to represent the plant in the genetic searching procedure. After determining the PID gains in this off-line manner, these gains are then applied to the real plant for on-line control. In training the neural network model, generation of training data and adjustment of neuron weights will be performed simultaneously in an on-line manner. To show the validity of the proposed method, a seesaw system with one input and two outputs will be used for experimental evaluation.
|Number of pages||12|
|Journal||Journal of the Chinese Institute of Engineers, Transactions of the Chinese Institute of Engineers, Series A/Chung-kuo Kung Ch'eng Hsuch K'an|
|Publication status||Published - 2002 Jan 1|
All Science Journal Classification (ASJC) codes