A novel hybrid T-S model identification algorithm.

To overcome the drawback of regular T-S model identification techniques, such as the FCM and least-squares method, a new Hybrid Identification Algorithm (HIA) is proposed in this paper. The HIA can simultaneously optimize all the model parameters and avoid being trapped into the local minima by merging the FCM, Harmony Search (HS) and the least-squares method together and using the error feedback mechanism. Our HIA is employed in the T-S modeling of the Gyro-stabilized platform. By comparing the MSE performance, the HIA can indeed yield a superior MSE performance over the conventional identification methods. The identification results show that the HIA can effectively overcome the incomplete optimization problem of the conventional identification methods. [ABSTRACT FROM AUTHOR]