An LMI approach for observer design for Takagi-Sugeno descriptor models

This work presents a novel observer design for nonlinear descriptor systems using their Takagi-Sugeno representation. The approach allows obtaining more relaxed results than previous literature by changing the extended estimated state vector; this step permits using a full observer gain. The obtained conditions are pure LMI. Moreover, previous results are always included by the new ones. An example illustrates the benefits of the proposed method.