Energy Disaggregation and Appliance Identification in Smart Homes

日期 : Sep 2020 - 现在

<div><ul> <li>Non-Intrusive Load Management (NILM) is used, operating all equipment from a single interface, and making it smart enough to disaggregate power from the aggregated power signal to identify and manage appliances smartly. </li> <li>For disaggregation and identification, different machine learning (ML) algorithms and neural networks (deep learning) are utilized such as <strong>k-Nearest Neighbors</strong> (kNN), <strong>Artificial Neural Network</strong> (ANN), <strong>Long Short-Term Memory</strong> (LSTM), <strong>Random Forest Regression</strong>, <strong>Decision Tree</strong>, <strong>Gaussian Naïve Bayes</strong>, <strong>Logistic regression</strong>,<strong> Linear regression</strong>, and <strong>Multi-Layer Perceptron (MLP)</strong>. These algorithms are implemented with the help of python's libraries, ie., Keras, TensorFlow, pandas, Scikit Learn, Seaborn, Keras tuner and NumPy</li> <li>Testing and training of these algorithms are done on our own collected data of different appliances, i.e, laptop, bulb, mobile and more. </li> </ul></div>