Identificação de dispositivos elétricos em ambientes domésticos

Abstract

The Internet of Things (IoT) has emerged as a key area in computing, driving global technological advancement. Smart cities and buildings have become part of our daily lives, and the idea that everything is connected to the Internet is becoming increasingly common. Task automation, traffic control, and energy consumption monitoring are just some of the areas impacted by this trend. With the increase in the number of electrical devices, managing energy resources in large buildings has become more complex. Poor management can lead to energy waste, unnecessary financial costs, and negative environmental impacts. In short, the difficulty in identifying which devices are operating can result in energy waste and reduced appliance lifespan. In this context, this paper proposes a solution to identify household electrical devices with a view to implementation in an IoT environment. The proposal involves analyzing the current signal in the frequency domain, which allows reducing the volume of processed data. Machine learning-based classification models were used to identify the device responsible for the signal, taking advantage of previously known patterns for this identification. The analysis of the electrical signals of the devices and the comparison with known patterns demonstrated that the investigated algorithms are effective in identifying the devices. Among the tested algorithms, the kNN classifier stood out, achieving approximately 99% accuracy and F1-score. These results indicate the high accuracy of the proposed approach and its effectiveness in identifying electrical devices in IoT environments.