On-line modification of T-S fuzzy models based on gradient descent algorithm
Feng Haodong, Wei Gou Dong, Jiao Huanyan · 2021
The gradient descent algorithm is applied to the model parameter correction algorithm in view of the shortcoming that T-S fuzzy model cannot be used for real-time adaptive parameter correction. In this paper, an online model correction method combining T-S fuzzy model with gradient descent algorithm is proposed, and the operation data of a subway air-conditioning system in Wuhan is used for verification. The experimental results show that the T-S fuzzy model modified by gradient descent algorithm has better accuracy and higher operational efficiency.