Adaptive chirplet based signal approximation

The chirp function is one of the most fundamental functions in nature. Many natural events can be roughly approximated by a group of chirp functions. In this paper, we present a practical adaptive chirplet based signal approximation algorithm. Unlike the other chirplet decompositions known so far, the elementary chirplet functions employed in this algorithm are adaptive. Therefore, the resulting approximation could better match the underlying signal and uses fewer coefficients. The effectiveness of the algorithm is demonstrated by numerical simulations