Occupancy estimation using environmental parameters
Date of Issue2017
School of Electrical and Electronic Engineering
Heating, ventilation, and air conditioning (HVAC) systems are the biggest energy consumer in office building. This has become a major problem as large amount of energy is wasted which contributes to global warming and greenhouse gas emission. Thus, saving energy has become very important, especially for countries like Singapore which have quite limited resources. Although image cameras and wearable sensors were demonstrated to be successful in detecting occupancy accurately, they are intrusive to the privacy of occupants. Motion sensors are limited to only binary detection. In this work, we use environmental sensors which are non-intrusive. To determine occupancy information, it is necessary to select a good feature set from the environmental parameters while irrelevant features are to be discarded. Filter methods such as Mutual Information and Pearson’s Correlation Coefficient have shown to be fast and effective in removing irrelevant features. In this work, Correlation Based Filter Method is used for feature selection as it is a popular method for real world problems. The selected features are then used to train three classifiers, namely K-Nearest Neighbour (KNN), Naïve Bayes and Neural Network. The respective accuracies are compared to identify the classifier that gives the highest accuracy. The Naïve Bayes classifier has the highest accuracy among all.
DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Control engineering
Final Year Project (FYP)
Nanyang Technological University