Personalized Thermal Comfort and Adaptive Building Control

This direction is split into other directions: subjective thermal preference. Its last version stays readable and citable here.

Version 1 of 122 papers34 conceptsfrozen 2026-10-04

Most modern buildings operate on a "one-size-fits-all" philosophy. HVAC systems typically aim for a static temperature setpoint derived from the Predicted Mean Vote (PMV) model—a decades-old standard designed for the "average" occupant. However, as Frederik Auffenberg, Sebastian Stein 0001, and Alex Rogers point out in A Personalised Thermal Comfort Model Using a Bayesian Network, this approach ignores the vast physiological and behavioral diversity of real people.

The emerging research direction in Personalized Thermal Comfort (PTC) shifts the goal from maintaining a universal temperature to learning and satisfying individual preferences. By treating comfort as a dynamic, data-driven variable rather than a static constraint, these systems can simultaneously improve occupant satisfaction and drastically reduce energy waste.

The Failure of the Average

The fundamental problem with standardized models is that they cannot distinguish between a user who is highly sensitive to a 1°C change and one who is comfortable across a 5°C range. Emil Laftchiev and colleagues argue in Personalizing Individual Comfort in the Group Setting that factors like metabolic rate, clothing insulation, and even gender create subjective comfort zones that a central thermostat simply cannot see.

When a system assumes a narrow, fixed range for everyone, it often resorts to "over-cooling" or "over-heating" to ensure no one complains. This leads to the "refrigerated office" phenomenon, where energy is expended to maintain a temperature that may actually be uncomfortable for a significant portion of the occupants.

Soft Comfort and Temperature Drift

One of the most powerful levers for energy efficiency is the concept of "soft comfort." In Supply Demand Coordination for Building Energy Saving: Explore the Soft Comfort, Zhanbo Xu suggests that occupants aren't just "comfortable" or "uncomfortable," but often have a flexible range of acceptance that fluctuates based on external factors like energy costs.

By modeling users as "frugal," "neutral," or "profligate," systems can implement controlled temperature drift. If the system knows a user has low thermal sensitivity, it can allow the temperature to rise slightly during peak energy pricing periods. This coordination between supply-side constraints (like grid load) and demand-side flexibility (individual comfort ranges) allows for significant savings without the occupant ever feeling a loss in quality of life.

Sensing Without Intrusion

To personalize comfort, a system needs data. Traditional methods rely on "feedback buttons" or invasive sensors, which suffer from user fatigue or privacy concerns. Recent breakthroughs focus on non-invasive inference:

  • Computer Vision: Xiaogang Cheng and Haibo Li developed the NIDL system, which uses Eulerian Video Magnification to detect subtle skin temperature changes from standard video feeds, allowing a deep neural network to "see" if a person is getting too warm.
  • Environmental Inference: The POSHS framework by Ali Etemad and François Rivest treats user identity as a "hidden state." By observing how temperature and humidity are adjusted over time, a Reinforcement Learning agent can infer who is in the room and what they prefer without needing biometric sensors.
  • Digital Twins: Hanqing Yang and Marie Siew propose using Large Language Models (LLMs) to simulate human behavior. These "digital twins" allow researchers to train control agents on diverse, synthetic populations, bypassing the need for months of real-world data collection.

Negotiating Shared Spaces

The most complex challenge in this field is the "multi-occupancy" problem. In a shared office or shopping mall, individual preferences inevitably conflict. Di Xie and Liang Yu 0001 address this in Distributed Real-Time HVAC Control by using "virtual temperature queues" to balance cost and collective discomfort.

Similarly, the Indoor Environmental Quality Management Ontology developed by Gabriel Jacoby-Cooper and Deborah L. McGuinness uses semantic reasoning to suggest low-energy interventions—like opening a window—that satisfy the aggregate needs of a group based on their specific clothing and activity levels.

Psychological Adaptation

Finally, research into natural ventilation highlights that comfort is partly psychological. Occupants in naturally ventilated buildings are often satisfied with a wider range of temperatures because they have "agency"—the ability to open a window or adjust a fan. By moving away from "sealed box" environments and toward adaptive thermal comfort models, buildings can slash energy use by up to 40% while maintaining high levels of perceived satisfaction.

In summary, the shift toward personalized thermal comfort replaces rigid engineering setpoints with flexible, human-centric models. By understanding the individual, these systems transform HVAC from a blunt instrument into a precise, energy-efficient service.

Papers in this version

History & citationVersion 1 · 2026-10-04

Split into other directions · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_personalized_thermal_comfort_and_adaptiv_a9e34275,
  title = {Personalized Thermal Comfort and Adaptive Building Control},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/personalized-thermal-comfort-and-adaptive-building-control/rdr_a9e34275fde24618bc18c62c81156c8b},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_personalized_thermal_comfort_and_adaptiv_a9e34275_v1,
  title = {Personalized Thermal Comfort and Adaptive Building Control},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/personalized-thermal-comfort-and-adaptive-building-control/rdr_a9e34275fde24618bc18c62c81156c8b/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}