An IoT Framework for Modeling and Controlling Thermal Comfort in Buildings

Abstract

Humans spend more than 90% of their day in buildings, where their health and productivity are demonstrably linked to thermal comfort. Building thermal comfort systems account for the largest share of U.S energy consumption. Despite this high-energy cost, due to building design complexity and the variety of building occupant needs, addressing thermal comfort in buildings remains a difficult problem. To overcome this challenge, this paper presents an Internet of Things (IoT) approach to efficiently model and control comfort in buildings. In the model phase, a method to access and exploit wearable device data to build a personal thermal comfort model has been presented. Various supervised machine-learning algorithms are evaluated to produce accurate personal thermal comfort models for each building occupant that exhibit superior performance compared to a general model for all occupants. The developed comfort models were used to simulate an intelligent comfort controller that uses the particle swarm optimization(PSO) method to search for optimal control parameter values to achieve maximum comfort. Finally, a framework for experimental validation of the new proposed comfort controller that interactively works with the HVAC element has been introduced.

Comparison of machine learning algorithms applied to symptoms to determine infectious causes of death in children: national survey of 18,000 verbal autopsies in the Million Death Study in India

“…The advantages of GBM include its high predictive accuracy and ability to predict multiclass data. The disadvantages of GBM include overfitting of data, sensitivity to noisy data, and requirement of high processing time [ 28 , 34 , 35 ].…”

Machine Learning for Smart and Energy-Efficient Buildings

“…Kernel-based approaches have been widely used for thermal sensation/preference prediction. The list of kernel-based methods popular for thermal comfort prediction includes SVM (Chaudhuri et al, 2018; Alsaleem et al, 2020; Chai et al, 2020; Liu et al, 2020; Zhou et al, 2020), K-nearest neighbors (KNN) (Liu et al, 2019; Lu et al, 2019; Pigliautile et al, 2020; Lee and Ham, 2021; Cheung et al, 2022), and ensemble learning algorithms, such as random forest (RF) (Kim et al, 2018; Liu et al, 2019) and AdaBoost (Ab). Recently, feed-forward neural networks (Zhai and Soh, 2017; Lu et al, 2019; Das et al, 2021), and time-series based networks (Chennapragada et al, 2022) have surpassed state-of-the-art kernel-based models in thermal comfort prediction.…”

A Review of Thermal Comfort in Primary Schools and Future Challenges in Machine Learning Based Prediction for Children

“…The most commonly used traditional algorithms include: (a) the family of linear and non-linear regression algorithms [87][88][89][90][91][92][93], (b) Decision Trees (DT) and Random Forests (RF) [87,89,[94][95][96], (c) Support Vector Machine (SVM) [23,87,95,[97][98][99][100][101], and (d) K-Nearest Neighbors (KNN) [87,91,[102][103][104][105][106]. The less commonly used ML techniques include Bayesian Models (BM) [87,107], Ensemble Learning (ENL) [87,88,97], Gaussian Models (GM) [108], Markov Models (MM) [88], Fuzzy classification [109,110], and Genetic Programming [111].…”