Intelligent signal processing for the use in device identification using smart sockets

Abstract

In an era, that has seen an increase in smart socket adoption in homes, greater sensor data acquisition and data analytics within the Internet of things (IoT) platforms; new developments in hardware design and converging sensor data with big data introduces new research opportunities in the energy sector. Smart meters currently provide an overall energy usage for a household, by introducing socket level identification of electrical devices an itemised bill or detailed breakdown for device type or category can be achieved. Voltage and current waveforms extracted from sensors within a smart socket is processed using signal processing techniques for the use of pattern recognition. Experimental results for single device identification show that a low equal error rate can be achieved, therefore, increasing the likelihood of a successful device recognition.