Hardware implementation of a Kullback-Leibler Divergence based signal anomaly detector

Abstract

A field-programmable gate array (FPGA) based implementation of the Kullback-Leibler Divergence (KLD) algorithm for real-time detection of anomalies in periodic signals is presented. To aid in the development of the hardware platform, an optimised, application specific, differential-equation version of the KLD equation is also proposed. Successful tests of the implementation against 10 MHz wireless signals show that the algorithm is indeed capable of real-time analysis of high speed test vectors.