Cyclical Electromechanical Error Denial System Using Matrix Profile

Abstract

We propose that the Matrix Profile data structure, conventionally applied to large scale time-series data mining, is applicable to the analysis and suppression of cyclical error in electromechanical systems, paving the way for an intelligent family of adaptable control systems which respond to environmental error at a computational cost low enough to be practical in embedded applications. We construct and evaluate the efficacy of a control algorithm utilizing the Matrix Profile, which we call Cyclical Electromechanical Error Denial System (CEEDS).

Tasks

Reproductions