부산대학교 도서관

This paper presents a signal identification algorithm for signals composed of a sum of periodic signals. This algorithm is based on the internal model principle. By using several internal models paralleled with a tuning function, this algorithm can predict or identify signals composed of multiple harmonics with uncertain frequencies amplitudes and relative phases. A desired band-pass filter can be incorporated into algorithm by selecting appropriate coefficients of the tuning function and internal models, which can reject the noise better and improve the performance. This work is based on previous work in continuous time [1], [2]. However a discrete implementation will be much more practical for implementation. The simulation result shows a good tracking of the original signal without disturbances.