A Peak-Trough Detection Algorithm Based on Momentum
Karl Harmer, Gareth Howells, Weihua Sheng, M.C. Fairhurst, Farzin Deravi · 2008
This paper presents a simple, yet novel, approach to peak-trough detection using a rudimentary model of Newtonian mechanics. Based on the line-searching technique also employed in artificial neural network technology to determine global minima, the momentum is used to find both peaks and troughs of a signal. This algorithm provides a fast alternative to the traditional techniques, which uses contextual information in order to determine prominent peaks and troughs without requiring smoothing or thresholding.