Edge Computing-Based Internet of Things Framework for Indoor Occupancy Estimation
Indoor occupancy estimation has become an important area of research in the recent past. Information about the number of people entering or leaving a building is useful in estimation of hourly sales, dynamic seat allocation, building climate control, etc. This work proposes a decentralized edge computing-based IoT framework in which the majority of the data analytics is performed on the edge, thus saving a lot of time and network bandwidth. For occupancy estimation, relative humidity and carbon dioxide concentration are used as inputs, and estimation models are developed using multiple linear regression, quantile regression, support vector regression, kernel ridge regression, and artificial neural networks. These estimations are compared using execution speed, power consumption, accuracy, root mean square error, and mean absolute percentage error.