A new design methodology is proposed for the optimal control of nonlinear s
ystems described by the Takagi-Sugeno (TS) fuzzy model. The TS fuzzy system
s are first classified into two families based on how diverse their input m
atrices are and then a controller synthesis procedure based on the inverse
optimal approach is given for each family, We also show that the optimal co
ntroller can he found by solving a linear matrix inequality problem, The op
timal controllers have robustness with respect to a class of input uncertai
nties. The proposed method is applied to the attitude control of a rigid sp
acecraft to demonstrate its validity.