Two parameter estimation approaches have been developed for system modeling and health monitoring of multi-story
buildings based on structural vibration responses under seismic excitation and/or applied forces. In the first approach,
parameter estimation is performed in a decentralized fashion, i.e. parameters are estimated for each story individually,
while in the second approach, the parameters are estimated in a centralized manner for the entire structure. In order to
reduce the effect of random noise, drift with D.C. off-set as well as low-frequency disturbances in the measured data,
three kinds of filter (i.e., high-pass filter, moving average filter and Kalman filter) have been used for data preprocessing.
Numerical simulations based on lumped parameter models and a finite element model of a realistic building have been
conducted. The results show that, the performance of the decentralized approach is better than the centralized one in the