A Method for Signal Change Detection via Short-Time Conditional Local Peaks Rate Feature

Abstract

In this paper, we present a method for signal change/event/anomaly detection based on a novel time-domain feature termed “conditional local peaks rate” (CLPR). First, the CLPR feature is described and further the method is introduced based on it. The CLPR calculation algorithm is implemented in the Matlab® software environment as a user-defined function and several numerical experiments are conducted with real-world data for sake of verification and validation. The performance of the proposed method is compared with three other classic detection methods based on the short-time energy, short-time zero-crossing rate and short-time kurtosis and the obtained results indicate its advantages. The accessibility of the Matlab® implementation allows repeatability of the experiments and facilitates the real practical application of the method.

Keywords

• time domain
• signal
• data
• change
• event
• anomaly
• detection