Personal thermal comfort models with wearable sensors

Abstract

A personal comfort model is an approach to thermal comfort modeling, for thermal environmental design and control, that predicts an individual's thermal comfort response, instead of the average response of a large population. We developed personal thermal comfort models using lab grade wearable in normal daily activities. We collected physiological signals (e.g., skin temperature, heart rate) of 14 subjects (6 female and 8 male adults) and environmental parameters (e.g., air temperature, relative humidity) for 2-4 weeks (at least 20 hours per day). Then we trained 14 models for each subject with different machine-learning algorithms to predict their thermal preference. The results show that the median prediction power could be up to 24% /78% /0.79 (Cohen's kappa/accuracy/AUC) with all features considered. The median prediction power reaches 21% /71% /0.7 after 200 subjective votes. We explored the importance of different features on the prediction performance by considering all subjects in one dataset. When all features included for the entire dataset, personal comfort models can generate the highest performance of 35% /76% /0.80 by the most predictive algorithm. Personal comfort models display the highest prediction power when occupants' thermal sensations is outside thermal neutrality. Skin temperature measured at the ankle is more predictive than measured at the wrist. We suggest that Cohen's kappa or AUC should be employed to assess the performance of personal thermal comfort models for imbalanced datasets due to the capacity to exclude random success.

Acoustics for Supportive and Healthy Buildings: Emerging Themes on Indoor Soundscape Research

“…Spatial and temporal sensory variability, together with opportunities for personal adaptation and control are encouraged for offering pleasure and delight [11,[28][29][30]. New technologies from medical sciences (e.g., wearable sensors) are gradually being introduced to assess the health aspects of human responses to environmental stimuli or to improve comfort models on the basis of occupants' physiological response [31,32]. Examples of this continuous and ongoing evolution in IEQ research and practice include the development of personal comfort models [32,33] and personal comfort systems [34], the application of circadian lighting design [35], and the valuation of qualitative aspects affecting visual quality [36,37]; but what about acoustics in this "pursuit of wellbeing" [15]?…”

Measurement of Personal Experienced Temperature Variations in Rural Households Using Wearable Monitors: A Pilot Study

“…Since the same device can obviously lead to different test results when worn at different locations on the body [ 47 ], uniformly specific location was considered in order to make the test results comparable. In existing studies, wearable devices were placed on the back [ 30 ], waist [ 24 ], wrist [ 48 ], hanging on the chest [ 44 ], clipped to the shoes [ 37 ], or placed into a carry-on backpack, and kept the probe outside [ 49 ]. Moreover, ISO 28802 recommends three different heights for similar tests: ankle, abdomen, and head [ 29 ].…”

Targeted occupant surveys: A novel method to effectively relate occupant feedback with environmental conditions

“…Improvements in measurement technologies have led to sensors becoming inexpensive, increasingly accurate, easy-to-use, smaller, and more portable. This sensor revolution has driven the development of continuous IEQ monitoring systems designed to be permanently distributed throughout an indoor environment (Brager, Paliaga, and de Dear 2004;Goto et al 2007;Cheung et al 2017;de Dear, Kim, and Parkinson 2018;Kim et al 2019;Liu et al 2019;Parkinson, Parkinson, and de Dear 2019).…”

“…We calculated the two metrics for the dataset collected in our pilot study and thermal comfort datasets from Liu et al (2019), Kim et al (2019), and Cheung et al (2017). Liu et al (2019) collected environmental parameters, occupant behavior, physiological measurements, and occupant thermal satisfaction responses from 14 subjects for 2-4 weeks to develop personal comfort models. Subjects in this field study completed rightnow surveys at their own discretion but researchers required 12 surveys per day, with incentives to complete more.…”

“…The ideal datasets were generated using Halton sequences (Halton 1960), a quasi-random sampling method with applications in Monte Carlo simulations. The region of interest for the datasets are: in our pilot study, time between 8:00 and 18:00 and radiant slab surface temperatures between 19 and 27 °C; in Liu et al (2019), time between 0:00 and 23:59 and hourly mean outdoor air temperature between 3 and 32 °C; in Kim et al (2019), time between 6:00 and 18:00 and indoor operative temperature between 18 and 29 °C and; in Cheung et al (2017), time between 0:00 and 23:59 and carbon dioxide concentration between 370 and 4720 ppm. The temperature or concentration range corresponds to the observed minimum and maximum in each dataset, and the time range corresponds to the building occupancy hours in our pilot study and in Kim et al (2019).…”