An Investigation of Wavelet Design using Genetic Algorithms
Yvette J. Hill, Steven G. O’Keefe, David Victor Thiel · 2001
This paper presents the development of a lightning stroke identification wavelet basis set using genetic algorithms. A basic introduction of the wavelet transform is presented followed by an overview of a genetic algorithm. The frequency content of the best wavelets produced coincided with the spectral content of the spherics examined. As the method is computationally inexpensive, development of the wavelets can be carried out on site. This tailors the identification wavelet to suit spherics at particular localities and allows the wavelet to address changes in frequency due to propagation paths and surface impedance. I. INTRODUCTION AVELETS are mathematical transient functions that when combined form basis sets representing an original time domain signal. Like the Windowed Fourier Transform (WFT), the output coefficients of the Wavelet Transform (WT) indicate the magnitude of the respective wavelet function required to reconstruct the original signal. The sum of the products of each wavelet function in the basis set and its respective coefficient will complete the reconstruction.