A new method for instantaneous signal period identification by repetitive pattern matching

Identifying reoccurring patterns within a signal set often yields important information about the signal. Fourier analysis is typically used as a method to determine periodic sinusoidal elements of a given signal; however, it is not able to properly identify near-periodic patterns with small instantaneous changes in the periodicity of the input waveform. To that end, a new transform has been developed which performs a form of high gain correlation on a selected subset of time-localized data points. This transform yields a periodicity vector which accurately identifies a set of samples which are found to be approximately repetitious over a two period timeframe. The transform has been extended to the problem of heart beat detection, referred to as QRS Complex detection in the literature.