An approach to analyse energy consumption of an IoT system

Abstract

Internet of Thing (IoT) has emerged as the one of best solution to provide service for different applications such as smart cities and precision agriculture. In the infrastructure based on an IoT network, multiple sensors or smart devices are linked to the IoT gateway. These devices consume considerably more power during the transmission or reception of data from the transceiver to the gateway, in comparison to sensing the data through sensors or processing data. The electronic devices used in smart buildings, smart cities, and smart agricultural system consumes more energy than traditional electronic equipment. To sustain the growing demand of smart appliances, optimization of smart electronic devices in terms of power consumption is essential. Many researchers and scientists are now working on optimizing energy usage as a central focus, along with providing comfort atmosphere to smart application projects. In this paper, energy consumption for sensors and system on chips (SoCs) has been calculated with respect to the duty cycle. The threshold region for operating the device has been obtained by finding the minimum eigenvalue. Further, the signal to noise (SNR) has been estimated using the pathloss equation for the received signal. The central target of this work is to optimize power consumption by using an artificial neural network (ANN). The performance of ANN has been acquired by the mean square error (MSE) function, using the scaled conjugate gradient (SCG) algorithm. The data set of 180 samples, for the training of neural network has been created. Also, the other parameters like the SNR, operable region, and energy efficient region for SoCs using the concept of duty-cycle have also been obtained for optimization.