Artificial intelligence (AI)-based identification of appliances in households through NILM

The Smart Grid (SG) technologies greatly enhance the electrical grid by providing more energy security, increasing efficiency, reliability and economically benefitting the supplier which translates to reduce prices to the end customer. This is significant, as most of the current available energy comes from non-renewable sources. Economic development of nations result in increased demand for energy resources and as the world moves forward, the rate at which these resources are being spent is increasing. The SG poses numerous challenges in the fields of communication, engineering and policymaking. An aspect of the SG Technologies like Non-Intrusive Load Monitoring (NILM) enables the collection of valuable information that can be used to increase energy efficiency and gain deep insight on energy statistics. Novel software approaches like Deep Neural Networks, Markov models and Support Vector Machines play an important role for solving the NILM problem. In addition, researchers are delving into other aspects of the implementation such as privacy, price, and resource usage that coming from the paradigms of cloud and edge computing. Expert analysis on all of these topics is essential for the adoption of SG technology. In this paper, the merits of the various AI/ML techniques are expanded upon in a literature review. Then the previous and current state-of-the-art methods in the field of Deep Learning are evaluated and a conclusion is reached on their performance and viability. Furthermore, a case is made about relevant computing paradigms and how they may impact fundamental areas of the technology.