Conference PaperPDF Available

Differences in thermal comfort state transitional time among comfort preference groups

Authors:

Abstract

As the individual difference in thermal comfort transition time is yet to be thoroughly explored, this paper studies time adaptability amongst individuals by collecting high frequency subjective comfort feedback using micro ecological momentary assessments on a smart-watch. The individuals were grouped based on their thermal preference responses and had their transition times analyzed. Each group has various transition time from arriving at and leaving from their comfort state. On average people who are more sensitive to cold temperatures, Group 1, take 8.9 minutes from being uncomfortably cold and 25.0 minutes from being uncomfortably hot to reach comfort zones. On the other hand, people who are generally comfortable, Group 2, take 22.4 minutes from uncomfortably cold and 27.1 minutes from uncomfortably hot to be thermally comfortable. The average transition time within a cluster matches the thermal comfort trend of said cluster. Ultimately, the transient time of preference groups raises the possibility to improve individualized thermal comfort models and machine learning in the future.
The 16th Conference of the International Society
of Indoor Air Quality & Climate ONLINE |
From November 1, 2020
Paper ID ABS-1162
Differences In Thermal Comfort State Transitional Time Among Comfort
Preference Groups
Pimpatsohn Sae-Zhang, Matias Quintana and Clayton Miller*
School of Design and Environment, National University of Singapore
*Corresponding email: clayton@nus.edu.sg
SUMMARY
As the individual difference in thermal comfort transition time is yet to be thoroughly explored,
this paper studies time adaptability amongst individuals by collecting high frequency subjective
comfort feedback using micro ecological momentary assessments on a smart-watch. The
individuals were grouped based on their thermal preference responses and had their transition
times analyzed. Each group has various transition time from arriving at and leaving from their
comfort state. On average people who are more sensitive to cold temperatures, Group 1, take
8.9 minutes from being uncomfortably cold and 25.0 minutes from being uncomfortably hot to
reach comfort zones. On the other hand, people who are generally comfortable, Group 2, take
22.4 minutes from uncomfortably cold and 27.1 minutes from uncomfortably hot to be
thermally comfortable. The average transition time within a cluster matches the thermal comfort
trend of said cluster. Ultimately, the transient time of preference groups raises the possibility to
improve individualized thermal comfort models and machine learning in the future.
KEYWORDS
Thermal adaptation, Adaptability, Transient time, Preference clustering, EMA
1 INTRODUCTION
For data collection purposes, many fields of study have relied upon Ecological Momentary
Assessment (EMA) to acquire subjective feedback at a higher frequency to adequately analyse
the environment. EMA is the collection of assessments, over time, of the subject’s current or
recent states in their natural environment (Shiffman et al., 2008). A variant known as
microinteraction-based EMA, or micro-EMA, is performed via a smartwatch where all
prompted surveys are reduced to fast and succinct interactions that can be answered within a
few seconds (Intille et al., 2016). As it is less disruptive to a participant’s ongoing activity,
micro-EMA achieves a higher sampling frequency i.e., 8 times more than traditional EMA
(Intille et al., 2016).
On the other hand, temperature change is known to be an important factor that affects human
thermal comfort (Mihara et al., 2019). Plenty of thermal comfort research has primarily focused
on steady-state conditions, yet the actual thermal environment is naturally transient and
dynamic over time. While most of the existing thermal comfort models are only applicable to
steady-state thermal environments, many researchers have acknowledged transient
environments and deviated to anticipate non-uniform human thermal response to better building
design (Zhang & Zhao, 2009; Mihara et al., 2019). However, there is a lack of studies on the
transition rate a person takes to arrive at a stable thermal comfort level. As individual
preferences correlate to thermal comfort and such thermal comfort is hardly stable throughout
the day (Kim et al., 2019), it is questionable whether the average time taken for an individual
also varies according to their differences or whether once a person arrives at their thermal
comfort condition, they will experience a change on their thermal comfort status. If so, how
long does one take to become thermally uncomfortable and how long will it be to arrive at the
thermal comfort condition once again, remain as questions.
This work aims to quantify individual differences and their time taken for the transition between
comfort and uncomfortable state on occupants during their normal day-to-day activities. We
employ a micro-EMA field study strategy to capture in-situ real-world data in a fully operating
building. Then, we classify the subjects into three different groups or clusters,, based on the
ratio of their preference responses to one’s overall votes. The three groups are Group 1 who
prefers a cooler environment, Group 2 who feels comfortable and prefers no change and Group
3 who prefers a warmer environment. We found that subjects based on comfort preferences
experience a similar transition time behaviour amongst respective groups.
2 METHODS
A micro-EMA data collection experiment took place in an educational building in Singapore
between October 2019 to September 2019. Subjects who either study or work at the building
were recruited for the experiment with a balanced gender mix and aged ranging from 20 to 30
years old. They were tasked to wear a Fitbit smartwatch with the cozie (Jayathissa et al., 2019)
application to provide feedback on their thermal comfort upon arriving and leaving the building
or whenever they sensed a change in their perceived thermal comfort. Each participant provided
a minimum of 100 data points with at least 50 given in spaces within the building. The thermal
comfort feedback was collected using a 3-point scale where ‘9’ denotes that the subjects prefer
the environment to be warmer, ‘10’ denotes that the subjects are thermally comfortable (no
change is needed) in such surroundings and ‘11’ denotes that the subjects prefer the
environment to be cooler. When the feedback is given, occupants’ location and timestamp are
captured and processed via the subject’s smartphone that is paired with the smartwatch and
Bluetooth beacons for indoor localisation inside the building This resulted in 3,638 data points
from 30 subjects. More details about the data collection experiment can be found in Jayathissa
et al., 2020.
The time difference between two data points for each subject was calculated if the timestamp
of a data point does not exceed 15 minutes with the previous data point and if they were given
in the same location. Figure 1 Left shows an example of a participant’s transition with its data
points that occurred within 15 minutes of each other across three different locations. The
computation of the transient duration was done in reference to the transition of the thermal
comfort vote from one value to another. If there is no change in the subjective vote, the time is
accumulated until a change occurs. Once there is a change, the accumulated time is recorded
for such a transition. The accumulated time then reset to 0 for the calculation of the next
transition.
Figure 1 (Left) Comfort transition graph of one participant at different locations. A total of 8
thermal transitions are shown as increasing or decreasing slopes. The yellow nodes represent
the time interval between the two data points which are accumulated for each transition. E.g,,
in Location A, the first transition is from ‘comfortable’ to ‘prefer cooler’ (10-11) and it takes
~16 minutes. Similarly, the second transition of ‘prefer cooler’ to ‘comfortable’ (11-10) takes
21minutes. However, the third transition, from ‘comfortable’ to ‘prefer cooler’ (10 -11),
happened in two different spaces so the time was not recorded. Figure 1 (Right) Percentage
distribution of thermal votes of all 30 subjects. The thermal comfort votes of each subject are
normalized to ratios with respect to each subject's total votes.
Later, participants were grouped based on the percentage of the highest vote they provided. A
participant with 80% of its votes being ‘prefer cooler’ is then treated as part of Group 1 whereas
a participant with 51% of its votes being ‘no change’ will be part of Group 2. Group 3 is then
formed by participants with the majority of their votes being ‘prefer warm’, Figure 1 (Right)
shows the participants, the group they belong to, and their ratio of responses.
Based on the resulting clusters and the values of the subjective thermal comfort and their ratios,
we see that the groups can be interpreted as groups where subjects 1) prefer cooler spaces, 2)
are mostly comfortable (require no change) and 3) prefer warmer spaces. Figure 1 Left shows
the subjects’ responses and despite being in the same tropical climate and the same building
throughout the whole experiment, they vary for each subject. This showcases the potential and
opportunity for personalized thermal comfort. The majority of the subjects, i.e., 28 subjects
(80%) belong to Group 2 (tend to be more comfortable), while the minority would usually seek
a change in their environment. Out of these remaining occupants, 6 subjects (17%) seem to
prefer a cooler environment (Group 1) and only one subject (3%) seem to prefer a warmer
environment (Group 3).
3 RESULTS
The changes in the subject’s thermal sensation can be broken down into various distinct groups
based on the initial and last self-reported thermal votes. This paper focuses on the changes from
being thermally uncomfortable to being thermally comfortable and vice versa. A breakdown of
the types and number of transitions is shown in Table 1. Based on these thermal behaviour
arrangements, even though it is expected that each preference group would have voted
according to their preference such as Group 1 who has higher ratio on prefer-cooler-option (11)
and Group 2 voted more for comfortable-option (10), the transient behaviour of the subject is
another aspect to remark i.e., how long the subject takes to reach or leave the comfortable state.
Table 1. Transition types and durations across all 30 participants
Transition Types
Number of Transition (No)
Mean (Minutes)
Uncomfortable to No Change State
‘Prefer Warmer’ to ‘Comfortable’ (9-10)
‘Prefer Cooler’ to ‘Comfortable’ (11-10)
Comfortable to Uncomfortable State
‘Comfortable’ to ‘Prefer Warmer’ (10-9)
‘Comfortable’ to ‘Prefer Cooler’ (10-11)
32
52
37
49
20.74
27.79
32.14
20.03
Looking further into the transitions times provide intuition about the behaviour of the groups
and allows us to better understand the relationship of thermal preference ratios, i.e., how the
groups were initially formed, and the transient time of their members.
4 DISCUSSION
Due to the high-frequency sampling of EMA methods like cozie (Jayathissa et al., 2019),
grouping the subjects according to their thermal comfort preferences provides a better
understanding of individual comfort differences. The discovered characteristics of the clustered
groups, such as statistics of the transient time, could serve as input features to improve a more
detailed individual or group thermal comfort model.
In the transition from uncomfortably cold to comfortable (9-10) state, subjects from Group 1
took less time on average than Group 2, 8.9 minutes and 22.4 minutes respectively: Subjects
who preferred to stay in a cooler environment were able to reach their comfortable state faster.
The absence of Group 3 could imply that the air supplied was not cool enough to be thermally
comfortable for this group. The participant in this group might either genuinely be in a constant
state of uncomfortably cold or possibly take a long time to reach their comfortable state.
Contradicting the ‘prefer warmer’ to ‘comfortable’ (9-10) transition where Group 1 was found
to change faster than Group 2, Group 1 group took lesser time for changing from comfortable
to uncomfortably hot (10-9), 18 minutes and 33.5 minutes respectively. Both Group 1 and 2 do
not have members that have experienced going from uncomfortably hot to comfortable (11-10)
state, but shared similar average transition time, 25 minutes and 27.1 minutes respectively,
shown in Figure 2(b). These results can be understood as the effectiveness of the existing
cooling systems in the buildings; regardless of the predominant preference in Group 1 and 2,
subjects in both groups adapt and reach a comfortable state in roughly the same time on average.
This may imply that, even if Group 1 is formed by users who voted ‘prefer cooler’ the most,
subjects in this group do not really rather a cooler environment but are just more sensitive to
colder environments. The underlying reasons, however, require further investigation.
Nevertheless, it is arguable that this type of transition is not common since there were only 3
data points from Group 1 by only 2 subjects out of 30 subjects. Contrastly, Group 2 has a much
wider range in transition times as high as 100 minutes to feel uncomfortable, highlighting the
characteristics of the group as they stayed in their comfort for a much longer time.
(a) ‘Prefer Warmer’ to ‘Comfortable’ (9-10)
shows the PC takes a shorter time than the
NC.
(b) ‘Prefer Cooler’ to ‘Comfortable’ (11-10)
shows that both preference groups have
similar transition times.
(c) ‘Comfortable’ to ‘Prefer Warmer’ (10-9)
shows that two drastic adaptability when
both groups were thermally comfortable.
(d) ‘Comfortable’ to ‘Prefer Cooler (10-11)
shows a long tail distribution for the NC and
bimodal distribution for the PC.
Figure 2 Distribution graphs of thermal transition times that each cluster takes for each
transition behaviour. Note that there is insufficient data for Group 3 subjects who generally
prefer a warmer environment, the distribution graphs thus only evaluate transitional behavioural
patterns of only Group 1 who prefer cooler and Group 2 who generally is comfortable with the
environment.
Given the definition of Group 1, it is not surprising to find a high frequency of transition from
being ‘comfortable’ to ‘prefer cooler’ (10-11). While some in Group 1 almost immediately felt
hot after reaching the premise, some took a while to feel the same way. The two distinct peaks
in Figure 2(d) show that there are people with different characteristics within the same Group.
In addition, in the same figure, Group 2 who seems to be thermally comfortable most of the
time has a high frequency of this type of transition too.
As for the limitations of this work, an experiment with a bigger cohort across different buildings
with different types of air-conditioning could provide deeper insights on the individual and
group transient time behaviour. While the data collection exercise imposes measures to mitigate
user gaming the voting system, refinements on the smartwatch application like the ability to
redo a feedback vote are still needed for more robust and bigger deployment.
5 CONCLUSIONS
The work sought to analyse transient behaviours of individuals and better comprehend the
correlation of such behaviour with respect to the characteristic of their thermal preference
group. Micro-EMA was used to achieve a high frequency sample of individual thermal
satisfaction over time and across different settings. The results suggest that thermal transient
behaviours are much more dynamic and significantly influenced by its immediate ambient
context but at the same time, the trend of subjective responses is reflected in the transition time.
Overall, Group 1, with the average of 8.9, 18 and 25 minutes, took shorter durations than Group
2 which experienced a change at 22.4, 33.5 and 27.1 minutes, upon reaching and leaving their
comfortable states.
Thermal satisfaction is not easy to quantify or sampled as it tends to vary continuously
according to the personal context and surrounding condition. A time-based sampling EMA
where the subjective measurement is done as per schedule i.e., a fixed period duration with
random interval could be a potential methodology to continue this research. Such fixed
frequency could provide a more representative overview of the subject's behaviour. Overall,
the understanding and calculation of transient behaviour, or time taken to adapt to a comfortable
state, yield explanations of groups of occupants formed solely on their ratio of thermal comfort
votes. This serves as a basis for further research into the usability of these features beyond
occupant’s behaviour description, e.g., input features for data-driven thermal comfort models.
6 ACKNOWLEDGEMENT
The Dept. of Building at the National University of Singapore and the Building and Urban Data
Science (BUDS) Lab provided support for the development and implementation of this work.
The data collection experiment was supported by Dr. Prageeth Jayathissa and Mahmoud M.
Abdelrahman.
7 REFERENCES
Intille, S., Haynes, C., Maniar, D., Ponnada, A., & Manjourides, J. (2016). muEMA:
Microinteraction-based Ecological Momentary Assessment (EMA) Using a Smartwatch.
Proc ACM Int Conf Ubiquitous Comput, 2016, 1124-1128. doi:10.1145/2971648.2971717
Jayathissa, P., Quintana, M., Sood, T., Nazarian, N., & Miller, C. (2019). Is your clock-face
cozie? A smartwatch methodology for the in-situ collection of occupant comfort data.
Jayathissa, P., Quintana, M., Abdelrahman, M., & Miller, C. (2020). Indoor Comfort
Personalities: Scalable Occupant Preference Capture Using Micro Ecological Momentary
Assessments. (Preprint)
Kim, J., Schiavon, S., & Brager, G. (2018). Personal comfort models A new paradigm in
thermal comfort for occupant-centric environmental control. Building and Environment,
132, 114-124. doi:https://doi.org/10.1016/j.buildenv.2018.01.023
Shiffman, S., Stone, A. A., & Hufford, M. R. (2008). Ecological momentary assessment. Annu
Rev Clin Psychol, 4, 1-32. doi:10.1146/annurev.clinpsy.3.022806.091415
Mihara, K., Sekhar, C., Tham, K. W., Takemasa, Y., & Lasternas, B. (2019). Effects of
temperature, air movement and initial metabolic rate on thermal sensation during transient
state in the tropics. Building and Environment, 155, 70-82.
doi:10.1016/j.buildenv.2019.03.030
Zhang, Y., & Zhao, R. (2009). Relationship between thermal sensation and comfort in non-
uniform and dynamic environments. Building and Environment, 44(7), 1386-1391.
doi:10.1016/j.buildenv.2008.04.006
... Accordingly, the integration of the momentary assessment with IoT techniques could revolutionize how the data in the built environment is tagged with information generated by occupants. Smartphone and smartwatch apps can be leveraged to create ecological momentary assessments that prompt a user to provide instantaneous feedback in an elegant and least-intrusive way (Moskowitz and Young, 2006;Jayathissa et al., 2020;Sae-Zhang et al., 2020;Quintana et al., 2021;Abdelrahman et al., 2022). An example of wearable devices that can be leveraged to create ecological momentary assessments that prompt a user to Frontiers in Built Environment frontiersin.org ...
... An example of wearable devices that can be leveraged to create ecological momentary assessments that prompt a user to Frontiers in Built Environment frontiersin.org (Moskowitz and Young, 2006;Sae-Zhang et al., 2020;Quintana et al., 2021;Abdelrahman et al., 2022). An example of using smartphone apps is discussed by Liu et al. (2018) where subjects are prompted by their cell phone (through a text reminder) to take an online survey and report their "right now" thermal comfort as often as possible. ...
Article
Full-text available
Cities today encounter significant challenges pertaining to urbanization and population growth, resource availability, and climate change. Concurrently, unparalleled datasets are generated through Internet of Things (IoT) sensing implemented at urban, building, and personal scales that serve as a potential tool for understanding and overcoming these issues. Focusing on air pollution and thermal exposure challenges in cities, we reviewed and summarized the literature on IoT environmental sensing on urban, building, and human scales, presenting the first integrated assessment of IoT solutions from the data convergence perspective on all three scales. We identified that there is a lack of guidance on what to measure, where to measure, how frequently to measure, and standards for the acceptable measurement quality on all scales of application. The current literature review identified a significant disconnect between applications on each scale. Currently, the research primarily considers urban, building, and personal scale in isolation, leading to significant data underutilization. We addressed the scientific and technological challenges and opportunities related to data convergence across scales and detailed future directions of IoT sensing along with short- and long-term research and engineering needs. IoT application on a personal scale and integration of information on all scales opens up the possibility of developing personal thermal comfort and exposure models. The development of personal models is a vital promising area that offers significant advancements in understanding the relationship between environment and people that requires significant further research.
... Recent literature employs machine learning models in order to contextualize environmental data [32][33][34][35] as well as thermoregulation from human skin through video [36], and physiological data [27,37,38]. In addition, some work has also analyzed the transition time it takes for an occupant to change its thermal preferences in the same indoor environment [39]. At the individual level, [19] has shown that data-driven PCM perform better than conventional models, such as the PMV, on a cohort of 38 occupants. ...
Article
Cohort Comfort Models (CCM) are introduced as a technique for creating a personalized thermal prediction for a new building occupant without the need to collect large amounts of individual comfort-related data. This approach leverages historical data collected from a sample population, who have some underlying preference similarity to the new occupant. The method uses background information such as physical and demographic characteristics and one-time onboarding surveys (satisfaction with life scale, highly sensitive person scale, personality traits) from the new occupant, as well as physiological and environmental sensor measurements paired with a few thermal preference responses. The framework was implemented using two personal comfort datasets containing longitudinal data from 55 people. The datasets comprise more than 6,000 unique right-here-right-now thermal comfort surveys. The results show that a CCM that uses only the one-time onboarding survey information of an individual occupant has generally as good or better performance as compared to conventional general-purpose models, but uses no historical longitudinal data as compared to personalized models. If up to ten historical personal preference data points are used, CCM increased the thermal preference prediction by 8% on average and up to 36% for half of the occupants in the first of the tested datasets. In the second dataset, one-third of the occupants increased their thermal preference prediction by 5% on average and up to 46%. CCM can be an important step toward the development of personalized thermal comfort models without the need to collect a large number of datapoints per person.
... Recent literature employs machine learning models in order to contextualize environmental data [32,33,34,35] as well as thermoregulation from human skin through video [36], and physiological data [28,37,38]. In addition, some work has also analyzed the transition time it takes for an occupant to change its thermal preferences in the same indoor environment [39]. At the individual level, [19] has shown that data-driven PCM perform better than conventional models, such as the PMV, on a cohort of 38 occupants. ...
Preprint
Full-text available
We introduce Cohort Comfort Models, a new framework for predicting how new occupants would perceive their thermal environment. Cohort Comfort Models leverage historical data collected from a sample population, who have some underlying preference similarity, to predict thermal preference responses of new occupants. Our framework is capable of exploiting available background information such as physical characteristics and one-time on-boarding surveys (satisfaction with life scale, highly sensitive person scale, the Big Five personality traits) from the new occupant as well as physiological and environmental sensor measurements paired with thermal preference responses. We implemented our framework in two publicly available datasets containing longitudinal data from 55 people, comprising more than 6,000 individual thermal comfort surveys. We observed that, a Cohort Comfort Model that uses background information provided very little change in thermal preference prediction performance but uses none historical data. On the other hand, for half and one third of each dataset occupant population, using Cohort Comfort Models, with less historical data from target occupants, Cohort Comfort Models increased their thermal preference prediction by 8~\% and 5~\% on average, and up to 36~\% and 46~\% for some occupants, when compared to general-purpose models trained on the whole population of occupants. The framework is presented in a data and site agnostic manner, with its different components easily tailored to the data availability of the occupants and the buildings. Cohort Comfort Models can be an important step towards personalization without the need of developing a personalized model for each new occupant.
... Through most of these studies, personal thermal comfort data from individuals were collected using computer and smartphone interfaces. To decrease survey fatigue and increase spatial and temporal diversity, smartwatches were used to collect data using micro-ecological momentary assessments (EMA) from building occupants about their satisfaction perception momentarily within certain environmental conditions [24,25,26]. ...
Article
Full-text available
Collecting intensive longitudinal thermal preference data from building occupants is emerging as an innovative means of characterizing the performance of buildings and the people who use them. These techniques have occupants giving subjective feedback using smartphones or smartwatches frequently over the course of days or weeks. The intention is that the data will be collected with high spatial and temporal diversity to best characterize a building and the occupant’s preferences. But in reality, leaving the occupant to respond in an ad-hoc or fixed interval way creates unneeded survey fatigue and redundant data. This paper outlines a scenario-based (virtual experiment) method for optimizing data sampling using a smartwatch to achieve comparable accuracy in a personal thermal preference model with fewer data. This method uses BIM-extracted spatial data and Graph Neural Network-based (GNN) modeling to find regions of similar comfort preference to identify the best scenarios for triggering the occupant to give feedback. This method is compared to two baseline scenarios that use conventional zoning and a generic 4x4 square meter grid method from two field-based data sets. The results show that the proposed Build2Vec method has an 18%–23% higher overall sampling quality than the spaces-based and square-grid-based sampling methods. The Build2Vec method also performs similar to the baselines when removing redundant occupant feedback points but with better scalability potential.
... The platform proved to be an efficient way to collect subjective feedback as compared to other post-occupancy feedback methods such as smartphone-based surveys. Subsequent field studies evaluated the platform's ability to capture a large number of thermal, aural, and visual comfort feedback to create personalized comfort models (Jayathissa et al. 2020), the ability to create personal comfort models using spatial proximity (Abdelrahman et al. 2022) and understand the thermal comfort transition behavior of occupants between different spaces (Sae-Zhang et al. 2020). Cozie is open source and can be easily modified to include different sets of questions beyond thermal comfort. ...
Preprint
Full-text available
This paper describes the adaptation of an open-source ecological momentary assessment smart-watch platform with three sets of micro-survey wellness-related questions focused on i) infectious disease (COVID-19) risk perception, ii) privacy and distraction in an office context, and iii) triggers of various movement-related behaviors in buildings. This platform was previously used to collect data for thermal comfort, and this work extends its use to other domains. Several research participants took part in a proof-of-concept experiment by wearing a smartwatch to collect their micro-survey question preferences and perception responses for two of the question sets. Participants were also asked to install an indoor localization app on their phone to detect where precisely in the building they completed the survey. The experiment identified occupant information such as the tendencies for the research participants to prefer privacy in certain spaces and the difference between infectious disease risk perception in naturally versus mechanically ventilated spaces.
... Through most of these studies, personal thermal comfort data from individuals were collected using computer and smartphone interfaces. To decrease survey fatigue and increase spatial and temporal diversity, smartwatches were used to collect data using micro-ecological momentary assessments (EMA) from building occupants about their satisfaction perception momentarily within certain environmental conditions [24,25,26]. ...
Preprint
Full-text available
Collecting intensive longitudinal thermal preference data from building occupants is emerging as an innovative means of characterizing the performance of buildings and the people who use them. These techniques have occupants giving subjective feedback using smartphones or smartwatches frequently over the course of days or weeks. The intention is that the data will be collected with high spatial and temporal diversity to best characterize a building and the occupant's preferences. But in reality, leaving the occupant to respond in an ad-hoc or fixed interval way creates unneeded survey fatigue and redundant data. This paper outlines a scenario-based (virtual experiment) method for optimizing data sampling using a smartwatch to achieve comparable accuracy in a personal thermal preference model with less data. This method uses BIM-extracted spatial data, and Graph Neural Network (GNN) based modeling to find regions of similar comfort preference to identify the best scenarios for triggering the occupant to give feedback. This method is compared to two baseline scenarios based on the spatial context of specific spaces and 4 x 4 m grid squares in the building using a theoretical implementation on two field-collected data sets. The results show that the proposed Build2Vec method is 18-23% more in the overall sampling quality than the spaces-based and the square-grid-based sampling methods. The Build2Vec method also has similar performance to the baselines when removing redundant occupant feedback points but with better scalability potential.
... For six weeks in early 2021, an experiment was deployed in the SDE1, SDE2, and SDE4 buildings with the use of the Cozie Fitbit smartwatch platform (https://github.com/cozie-app) to collect subjective thermal preference feedback. This method builds upon previous work [12,14], but with adaptations to increase the number of ecological momentary assessment questions and the diversity of spaces in the case study buildings. The test participants were asked about their thermal preference, clothing level, and activity level on the Cozie watch-face several times per day while they were in the case study buildings. ...
Article
Full-text available
Internet-of-Things (IoT) devices in buildings and wearable technologies for occupants are quickly becoming widespread. These technologies provide copious amounts of high-quality temporal data pertaining to indoor and outdoor environmental quality, comfort, and energy consumption. However, a barrier to their use in many applications is the lack of spatial context in the built environment. Adding Building Information Models (BIM) and Geographic Information Systems (GIS) to these temporal sources unleashes potential. We call this data convergence the Internet-of-Buildings or IoB. In this paper, a digital twin case study of data intersection from various systems is outlined. Initial insights are discussed for an experiment with 17 participants that focused on the collection of occupant subjective feedback to characterize indoor comfort. The results illustrate the ability to capture data from wearables in the context of a BIM data environment.
... Each survey comprised questions about thermal preference, clothing level, metabolic rate, perceived air velocity, and location. This application has been used to collect similar preferences responses in previous studies [1,6,11,13] with very little impact on the participants' activities and it can be installed for free on both Android 1 and Apple 2 platforms. Participants were not required to respond at specific times of the day, as long as they completed two responses more than 15 minutes apart and while within three educational buildings. ...
Conference Paper
Full-text available
Thermal comfort affects the well-being, productivity, and overall satisfaction of building occupants. However, due to economical and practical limitations, the number of longitudinal studies that have been conducted is limited, and only a few of these studies have shared their data publicly. Longitudinal datasets collected indoors are a valuable resource to better understand how people perceive their environment. Moreover, they provide a more realistic scenario to those conducted in thermal chambers. Our objective was to share publicly a longitudinal dataset comprising data collected over a 4-week long experiment. A total of 17 participants completed thermal preferences surveys which accounted for a total of approximately 1400 unique responses across indoor and outdoor 17 spaces. For the whole duration of the study, we monitored environmental variables (e.g., temperature and relative humidity) throughout 3 buildings. Participants completed comfort surveys from the screen of their smartwatches using an open-source application named Cozie. Their indoor location was continuously monitored using a custom-designed smartphone application. Location data were used to time and spatially align environmental measurements to thermal preference responses provided by the participants. Background information of participants, such as physical characteristics and personality traits (satisfaction with life scale, highly sensitive person scale, the Big Five personality traits), was collected using an on-boarding survey administered at the beginning of the experiment. The dataset is available at: https://zenodo.org/record/5502441\#.YT7xyaARUTs}.
... This study also collected temporal data from fixed IEQ sensors in the indoor environment (temperature, humidity, noise, and carbon dioxide) and physiological data from the smartwatch (heart rate, near body temperature). The data from this study was used to build personal comfort models [28], analyze the comfort behavior for the transition period between spaces [46], and understand the value of class balancing for thermal comfort prediction [47]. This paper builds upon this work by adding the building's spatial data (from the BIM model) and the spatialtemporal data of occupants related to their environment (location of people as compared to the sensors). ...
Article
Full-text available
Conventional thermal preference prediction in buildings has limitations due to the difficulty in capturing all environmental and personal factors. New model features can improve the ability of a machine learning model to classify a person’s thermal preference. The spatial context of a building can provide information to models about the windows, walls, heating and cooling sources, air diffusers, and other factors that create micro-environments that influence thermal comfort. Due to spatial heterogeneity, it is impractical to position sensors at a high enough resolution to capture all conditions. This research aims to build upon an existing vector-based spatial model, called Build2Vec, for predicting spatial–temporal occupants’ indoor environmental preferences. Build2Vec utilizes the spatial data from the Building Information Model (BIM) and indoor localization in a real-world setting. This framework uses longitudinal intensive thermal comfort subjective feedback from smart watch-based ecological momentary assessments (EMA). The aggregation of these data is combined into a graph network structure (i.e., objects and relations) and used as input for a classification model to predict occupant thermal preference. The results of a test implementation show 14%–28% accuracy improvement over a set of baselines that use conventional thermal preference prediction input variables.
... This study also collected temporal data from fixed IEQ sensors in the indoor environment (temperature, humidity, noise, and carbon dioxide) and physiological data from the smartwatch (heart rate, near body temperature). The data from this study was used to build personal comfort models [28], analyze the comfort behavior for the transition period between spaces [46], and understand the value of class balancing for thermal comfort prediction [47]. This paper builds upon this work by adding the building's spatial data (from the BIM model) and the spatialtemporal data of occupants related to their environment (location of people as compared to the sensors). ...
Preprint
Full-text available
Conventional thermal preference prediction in buildings has limitations due to the difficulty in capturing all environmental and personal factors. New model features can improve the ability of a machine learning model to classify a person's thermal preference. The spatial context of a building can provide information to models about the windows, walls, heating and cooling sources, air diffusers, and other factors that create micro-environments that influence thermal comfort. Due to spatial heterogeneity, it is impractical to position sensors at a high enough resolution to capture all conditions. This research aims to build upon an existing vector-based spatial model, called Build2Vec, for predicting spatial-temporal occupants' indoor environmental preferences. Build2Vec utilizes the spatial data from the Building Information Model (BIM) and indoor localization in a real-world setting. This framework uses longitudinal intensive thermal comfort subjective feedback from smart watch-based ecological momentary assessments (EMA). The aggregation of these data is combined into a graph network structure (i.e., objects and relations) and used as input for a classification model to predict occupant thermal preference. The results of a test implementation show 14-28% accuracy improvement over a set of baselines that use conventional thermal preference prediction input variables.
Conference Paper
Full-text available
Ecological Momentary Assessment (EMA) is a method of in situ data collection for assessment of behaviors, states, and contexts. Questions are prompted during everyday life using an individual's mobile device, thereby reducing recall bias and increasing validity over other self-report methods such as retrospective recall. We describe a microinteraction-based EMA method ("micro" EMA, or μEMA) using smartwatches, where all EMA questions can be answered with a quick glance and a tap -- nearly as quickly as checking the time on a watch. A between-subjects, 4-week pilot study was conducted where μEMA on a smartwatch (n=19) was compared with EMA on a phone (n=14). Despite an =8 times increase in the number of interruptions, μEMA had a significantly higher compliance rate, completion rate, and first prompt response rate, and μEMA was perceived as less distracting. The temporal density of data collection possible with μEMA could prove useful in ubiquitous computing studies.
Article
Full-text available
Assessment in clinical psychology typically relies on global retrospective self-reports collected at research or clinic visits, which are limited by recall bias and are not well suited to address how behavior changes over time and across contexts. Ecological momentary assessment (EMA) involves repeated sampling of subjects' current behaviors and experiences in real time, in subjects' natural environments. EMA aims to minimize recall bias, maximize ecological validity, and allow study of microprocesses that influence behavior in real-world contexts. EMA studies assess particular events in subjects' lives or assess subjects at periodic intervals, often by random time sampling, using technologies ranging from written diaries and telephones to electronic diaries and physiological sensors. We discuss the rationale for EMA, EMA designs, methodological and practical issues, and comparisons of EMA and recall data. EMA holds unique promise to advance the science and practice of clinical psychology by shedding light on the dynamics of behavior in real-world settings.
Article
A personal comfort model is a new approach to thermal comfort modeling that predicts an individual's thermal comfort response, instead of the average response of a large population. It leverages the Internet of Things and machine learning to learn individuals' comfort requirements directly from the data collected in their everyday environment. Its results could be aggregated to predict comfort of a population. To provide guidance on future efforts in this emerging research area, this paper presents a unified framework for personal comfort models. We first define the problem by providing a brief discussion of existing thermal comfort models and their limitations for real-world applications, and then review the current state of research on personal comfort models including a summary of key advances and gaps. We then describe a modeling framework to establish fundamental concepts and methodologies for developing and evaluating personal comfort models, followed by a discussion of how such models can be integrated into indoor environmental controls. Lastly, we discuss the challenges and opportunities for applications of personal comfort models for building design, control, standards, and future research.
Is your clock-face cozie? A smartwatch methodology for the in-situ collection of occupant comfort data
  • P Jayathissa
  • M Quintana
  • T Sood
  • N Nazarian
  • C Miller
Jayathissa, P., Quintana, M., Sood, T., Nazarian, N., & Miller, C. (2019). Is your clock-face cozie? A smartwatch methodology for the in-situ collection of occupant comfort data.
Indoor Comfort Personalities: Scalable Occupant Preference Capture Using Micro Ecological Momentary Assessments
  • P Jayathissa
  • M Quintana
  • M Abdelrahman
  • C Miller
Jayathissa, P., Quintana, M., Abdelrahman, M., & Miller, C. (2020). Indoor Comfort Personalities: Scalable Occupant Preference Capture Using Micro Ecological Momentary Assessments. (Preprint)