Definition and Principles

Occupant-centric building controls (OCC) refer to indoor climate control systems that prioritize the real-time preferences, behaviors, and comfort needs of building occupants over fixed environmental setpoints, dynamically adjusting parameters such as temperature, lighting, ventilation, and air quality using data from sensors and occupant inputs.[3] OCC strategies are broadly categorized into occupancy-centric methods, which respond to presence or density (e.g., demand-controlled ventilation via CO₂ sensors), and occupant behavior-centric methods, which infer or learn preferences from interactions (e.g., thermostat adjustments) or explicit feedback (e.g., mobile apps). This approach integrates occupant presence, interactions with building elements (e.g., thermostat adjustments or window openings), and indoor environmental quality (IEQ) metrics to enable personalized environmental conditions while optimizing energy efficiency.[1] Unlike passive assumptions of uniform occupant needs, OCC treats individuals as active participants in the control loop, fostering adaptive responses that enhance satisfaction and well-being.[4]

The core principles of OCC emphasize personalization, adaptability, and sustainability. Personalization tailors environmental controls to individual or group-specific needs, such as varying thermal preferences based on activity levels or demographics, achieved through explicit feedback (e.g., via mobile apps) or implicit inference from behaviors.[3] Adaptability involves real-time responsiveness to occupancy patterns, such as scaling ventilation to occupant density or adjusting lighting to presence, using data-driven models to predict and react to changes without rigid schedules. Key technologies include explicit sensors (e.g., motion detectors, Wi-Fi counters, wearables) for direct detection and implicit methods (e.g., inferring from CO₂ levels or energy patterns).[1] Sustainability is embedded by balancing comfort with energy optimization, reducing waste from over-conditioning unoccupied spaces or mismatched setpoints, often yielding 20-50% energy savings in simulated scenarios while maintaining or improving IEQ.[3]

In contrast to traditional building controls, which rely on rule-based or scheduled operations with static setpoints (e.g., fixed HVAC timetables assuming constant occupancy), OCC incorporates continuous feedback loops to incorporate occupant agency.[5] Traditional systems often treat occupants as uniform heat sources, leading to inefficiencies like simultaneous heating and cooling due to overrides, whereas OCC uses sensors for IEQ monitoring, occupant inputs for preferences, and actuators for adjustments, creating a closed-loop system.[3] A basic representation of this feedback loop is as follows:

This occupant-in-the-loop dynamic reduces performance gaps from behavioral mismatches, which can account for up to 80% variance in energy use.[3]

The principles of OCC trace their origins to responses to the 1970s energy crisis, when research highlighted the critical role of human behavior in building energy consumption, shifting focus from purely technical solutions to occupant-inclusive strategies.[3] Practical emergence accelerated in the 2000s with advancements in Internet of Things (IoT) technologies, enabling affordable sensors and real-time data integration for occupant-focused automation.[5]

Importance and Benefits

Occupant-centric building controls offer substantial energy savings by aligning HVAC operations with real-time occupancy patterns rather than fixed schedules, potentially reducing energy use by 19% to 45% in office buildings across various U.S. climate zones.[6] Field studies, numbering over 35 worldwide as of 2023, report 24–58% savings in HVAC for presence-based controls and up to 37% overall through integrated systems, with 65% reductions in thermal discomfort.[1] This approach, informed by ASHRAE guidelines, optimizes ventilation, airflow, and temperature setpoints, with higher savings in heating-dominated and marine climates where resets for unoccupied spaces yield greater impacts.[7] Such reductions contribute to lower operational costs, with studies indicating a return on investment within two years for occupancy-based systems in commercial settings.[8]

Beyond efficiency, these controls enhance occupant well-being by improving thermal comfort and indoor environmental quality, thereby reducing instances of sick building syndrome and associated health issues.[9] Research from Lawrence Berkeley National Laboratory estimates that better indoor environments, achievable through responsive controls, could yield annual productivity gains valued at $20 to $160 billion nationwide by minimizing health-related absences and boosting direct performance.[9] This translates to tangible benefits like fewer sick days and higher cognitive function, fostering healthier work environments. Post-COVID-19 adaptations, such as variable ventilation for health compliance in hybrid work settings using implicit occupancy data (e.g., plug loads), further underscore OCC's role.[1]

On a sustainability level, occupant-centric controls lower carbon footprints via demand-responsive strategies that curtail unnecessary energy consumption, supporting global efforts to decarbonize buildings.[7] Economically, the associated cost savings and extended equipment life further amplify their value, often achieving payback periods of two years or less. Challenges include privacy concerns with occupant data (addressed via pseudonymization under regulations like GDPR), the need for building-specific sensor calibration, and training for operators to integrate quantitative (e.g., energy metrics) and qualitative (e.g., satisfaction surveys) assessments.[1][3]

Broader societal impacts include stronger alignment with green building standards, where these controls help achieve LEED certification through energy optimization and WELL certification via enhanced occupant health metrics, such as 39% higher satisfaction rates in WELL buildings compared to LEED.[10] By prioritizing human-centered design, they promote equitable, resilient built environments that balance efficiency with user needs.[11]

Early Development

The origins of occupant-centric building controls can be traced to the 1970s, amid the global oil crises that spurred efforts to optimize energy use in buildings. Early developments focused on demand-controlled ventilation (DCV) systems, which adjusted airflow based on occupancy to reduce energy consumption without compromising indoor air quality. These systems emerged as a direct response to the 1973-1974 energy crisis, leveraging CO2 sensors to modulate ventilation rates in real-time, marking an initial shift toward incorporating occupant presence into building operations.[12]

A foundational contribution during this period was P.O. Fanger's Predicted Mean Vote (PMV) model, introduced in 1970, which quantified thermal comfort based on occupant physiological responses to environmental factors like temperature, humidity, and air velocity. The PMV model provided a predictive framework for assessing how building controls could align with human thermal sensations, influencing subsequent standards for occupant-focused HVAC design.

In the 1990s, progress accelerated with the integration of occupant feedback mechanisms in emerging smart building concepts, facilitated by basic occupancy sensors and environmental monitors. European initiatives, such as the 1992 European Indoor Air Quality Audit Project, employed trained sensory panels alongside instrumental measurements to evaluate occupant perceptions in 56 office buildings, laying groundwork for feedback-driven control adjustments. These efforts highlighted the value of direct occupant input in refining ventilation and thermal systems, though prototypes remained limited to research settings.[13]

The 2000s marked a breakthrough through the adoption of wireless sensors and nascent Internet of Things (IoT) technologies, enabling more dynamic occupant modeling in building controls. A pivotal publication, Bourgeois et al.'s 2005 paper, analyzed the energy impacts of occupant behavioral responses to lighting systems, demonstrating how detailed stochastic models of occupancy patterns could improve simulation accuracy and control efficacy. This work underscored the need for probabilistic approaches to capture variability in occupant actions, paving the way for integrated systems.[14]

A key milestone came with the 2005 revision of ISO 7730, which established ergonomics standards for moderate thermal environments by incorporating occupant data into PMV-based assessments, including provisions for personalized comfort zones. This standard formalized the use of occupant-centric metrics in design and operation, influencing global practices for energy-efficient buildings.

Key Theoretical Concepts

Occupant-centric building controls are grounded in human-centric models that prioritize individual thermal comfort over uniform environmental conditions. A foundational approach is Fanger's Predicted Mean Vote (PMV) model, which quantifies thermal sensation on a seven-point scale from -3 (cold) to +3 (hot), with zero representing neutral comfort. The PMV is calculated using the equation:

where MMM is the metabolic rate and LLL is the thermal load on the body, incorporating factors such as air temperature, mean radiant temperature, relative humidity, air velocity, clothing insulation, and activity level.[15] This model underpins the Predicted Percentage of Dissatisfied (PPD) index, which estimates the proportion of occupants likely to be dissatisfied with thermal conditions, typically aiming for PPD below 10% in controlled environments.[15] Complementing this, the adaptive comfort model by de Dear and Brager accounts for behavioral adaptation and contextual influences, positing that occupants in naturally ventilated buildings adjust expectations based on outdoor climate, allowing wider indoor temperature ranges (e.g., 23–26°C in temperate zones) while maintaining satisfaction.[16]

Behavioral modeling in occupant-centric controls addresses variability in occupancy patterns and individual preferences, essential for dynamic system responses. Concepts such as occupancy patterns capture presence and movement rhythms, often modeled stochastically to predict absences or relocations with probabilistic accuracy. Preference elicitation involves gathering occupant feedback on comfort parameters to personalize controls, enabling systems to adapt to heterogeneous needs. A key method employs Markov chains to simulate occupant presence as a sequence of states (e.g., occupied or vacant), where transition probabilities reflect empirical data from sensor observations, achieving prediction errors as low as 5–10% in office settings.[17]

Feedback theory frames occupant-centric controls as closed-loop systems, where occupant satisfaction serves as the dynamic setpoint rather than fixed thresholds. In these systems, sensors and algorithms continuously monitor environmental conditions and user inputs, adjusting actuators (e.g., HVAC) to minimize deviation from comfort targets, akin to proportional-integral-derivative (PID) controllers but augmented with occupant data. This approach integrates the PMV equation as a core metric for real-time feedback, ensuring iterative refinements toward optimal satisfaction.[15]

The integration of psychology enhances these models by incorporating cognitive factors that influence perceived comfort, such as expectations, control over the environment, and emotional states. Drawing from environmental psychology, research highlights how perceived controllability—e.g., ability to adjust personal settings—can increase tolerance for discomfort by up to 2–3°C deviations from neutral temperatures. Cognitive appraisals, including past experiences and social context, modulate thermal sensations, underscoring the need for holistic models beyond physiological metrics.

Occupant-Centric Control Inputs

Occupant-centric building controls rely on a variety of sensor-based inputs to capture real-time data about environmental conditions and occupant physiology, enabling responsive adjustments to indoor environments. Environmental sensors, such as those measuring CO2 levels, humidity, and motion, are commonly deployed to detect occupancy patterns and indoor air quality, informing ventilation and lighting controls. For instance, CO2 sensors help identify occupant density by correlating elevated levels with presence, while motion detectors track movement to optimize energy use in unoccupied zones. Wearable devices provide physiological data, like heart rate variability, to infer individual thermal preferences, allowing for personalized comfort adjustments beyond standard models. Bluetooth beacons facilitate location tracking within buildings, enabling systems to associate specific occupants with zones for targeted environmental control.[18][19]

Direct user inputs empower occupants to actively communicate preferences, enhancing control personalization through intuitive interfaces. Mobile applications often incorporate voting mechanisms where users rate thermal or lighting conditions on scales like "too hot/cold" or "too bright/dim," aggregating feedback to guide system responses. Voice commands and gesture-based interfaces, integrated into smart devices, allow hands-free adjustments, such as requesting temperature changes or dimming lights, making controls accessible in diverse settings. Privacy considerations are paramount in collecting these inputs, with techniques like data anonymization—such as pseudonymization or aggregation—employed to protect user identities while retaining utility for control decisions. These methods ensure compliance with regulations and mitigate risks of surveillance perception.

Indirect inputs derive from observed occupant behaviors, providing implicit signals of comfort needs without explicit feedback. Actions such as opening windows or manually adjusting thermostats reveal preferences for fresh air or temperature shifts, often captured via integrated sensors on fixtures. These behavioral traces inform models of habitual patterns, allowing controls to anticipate and automate similar adjustments. Data fusion techniques combine these indirect signals with other sources for robust inference; for example, basic Kalman filtering smooths noisy behavioral data by estimating states from sequential observations, improving reliability without complex derivations.

Ensuring input accuracy poses significant challenges, particularly in handling data noise and accounting for occupant heterogeneity. Sensor readings can be affected by environmental interference, such as drafts skewing humidity measurements or transient motions triggering false positives in detectors, necessitating robust filtering to maintain precision. Occupant variability further complicates this, as metabolic rates—key to thermal comfort—differ by factors like age and gender; for example, older adults and females often exhibit 5% lower rates, leading to divergent comfort thresholds that standard controls may overlook. These issues underscore the need for adaptive strategies that interpret inputs within theoretical models of human physiology.[20]

Predictive Controls

Predictive controls in occupant-centric building systems anticipate future occupant behaviors and environmental needs by leveraging forecasting models, enabling proactive adjustments to systems like HVAC and lighting rather than reacting to real-time conditions. These methods rely on predictive analytics to forecast variables such as occupancy levels, thermal comfort preferences, and energy demands, optimizing building performance while maintaining occupant satisfaction. By integrating data from sensors, historical records, and external sources, predictive controls can pre-emptively modulate indoor environments, reducing energy waste and enhancing responsiveness.

Forecasting techniques form the core of predictive controls, employing statistical and machine learning models tailored to building data. For occupancy prediction, autoregressive integrated moving average (ARIMA) models are widely used, capturing temporal patterns in occupant presence. An ARIMA(p,d,q) model involves applying the ARMA(p,q) structure to the d-th differenced series for stationarity and is expressed as:

where ∇dyt\nabla^d y_t∇dyt​ is the d-th difference of the series at time ttt, ppp denotes the autoregressive order, ddd the degree of differencing, qqq the moving average order, βi\beta_iβi​ are autoregressive coefficients, θj\theta_jθj​ are moving average coefficients, and ϵt\epsilon_tϵt​ is white noise. This formulation allows the model to extrapolate future occupancy based on past observations, with applications in commercial buildings showing improved prediction accuracy over baseline methods. Neural networks, such as long short-term memory (LSTM) architectures, extend this to comfort forecasting by processing multivariate time-series data, including occupant feedback and physiological signals, to predict personalized thermal preferences. These deep learning approaches handle non-linear relationships effectively, outperforming traditional models in dynamic environments.

Data-driven prediction enhances these techniques by incorporating historical patterns alongside external factors like weather forecasts and occupancy schedules to pre-adjust building systems. For instance, models trained on past usage data can predict peak occupancy periods, enabling HVAC pre-cooling or heating to maintain comfort without overconsumption. Studies evaluating such systems report mean absolute error (MAE) values as low as 0.5 occupants for short-term forecasts, leading to approximately 20% greater energy savings compared to reactive controls in office settings. This predictive capability is particularly valuable in variable climates, where integrating real-time weather APIs allows models to adjust ventilation proactively.

Hybrid approaches combine rule-based predictions with artificial intelligence to balance interpretability and accuracy, often applied in lighting and HVAC systems. Rule-based elements might use predefined thresholds from building codes or schedules to initiate predictions, while AI refines them with learned patterns from occupant interactions. In lighting applications, hybrids forecast natural daylight availability to dim artificial sources preemptively, reducing energy use by up to 15% in testbed facilities. For HVAC, these methods enable zone-specific pre-heating based on anticipated arrivals, as demonstrated in smart office implementations where occupant comfort scores improved alongside efficiency gains.

The evolution of predictive controls has progressed from simple rule-based predictors in the early 2010s, which relied on static schedules for basic anticipatory adjustments, to sophisticated deep learning frameworks in the 2020s that incorporate real-time learning and multi-modal data fusion. This shift has been driven by advances in computational power and sensor affordability, enabling scalable deployment in diverse building types.

Control Algorithms

Control algorithms in occupant-centric building controls form the core computational logic that processes occupant data—such as preferences, behaviors, and real-time feedback—into actionable decisions for systems like HVAC, lighting, and shading. These algorithms aim to balance multiple objectives, including energy efficiency, thermal comfort, and occupant satisfaction, by optimizing control variables over time horizons. Unlike traditional rule-based controls, they incorporate dynamic models of occupant needs and building dynamics to enable adaptive, personalized responses.[21]

Optimization-based algorithms are widely used for multi-objective tradeoffs in occupant-centric controls, often employing techniques like genetic algorithms to search for solutions that minimize conflicts between energy use and comfort metrics. For instance, genetic algorithms evolve populations of potential control strategies through selection, crossover, and mutation processes, evaluating fitness based on occupant-reported satisfaction and simulated energy consumption. This approach has been applied to optimize shading and ventilation in office settings, achieving up to 20% energy savings while maintaining thermal comfort within personalized bounds.[22] A key formulation in such optimizations is the cost function minimization:

where JJJ is the total cost, EEE represents energy consumption, DDD quantifies discomfort (e.g., deviation from preferred temperature setpoints), and w1,w2w_1, w_2w1​,w2​ are weighting factors tuned to occupant priorities. This objective is typically solved using linear programming when constraints are linear, such as bounds on temperature and airflow rates. The solution process involves: (1) defining decision variables (e.g., setpoint adjustments), (2) formulating inequality constraints (e.g., Tmin⁡≤T≤Tmax⁡T_{\min} \leq T \leq T_{\max}Tmin​≤T≤Tmax​), (3) minimizing JJJ subject to these constraints via a solver like simplex method, and (4) iterating over prediction horizons. Pseudocode for a basic linear programming implementation in occupant control might resemble:

This method ensures feasible, occupant-aligned solutions but requires accurate linear approximations of nonlinear building behaviors.[23]

Reinforcement learning (RL) algorithms treat the building as an environment where an agent learns optimal policies through trial-and-error interactions, using rewards derived from occupant satisfaction scores, such as reduced thermal discomfort or improved air quality preferences. In occupant-centric applications, deep RL variants like Q-learning or actor-critic methods model states including occupancy patterns and external weather, enabling the agent to adapt controls like window opening or lighting dimming without explicit programming. Experimental deployments in residential buildings have demonstrated RL reducing energy use by 15-30% compared to static controls, while learning individualized behaviors over weeks of operation. Challenges include the need for safe exploration during learning to avoid discomfort spikes.[24]

Integration with Building Systems

Occupant-centric building controls (OCC) integrate with existing building infrastructure, such as heating, ventilation, and air conditioning (HVAC) systems and building management systems (BMS), to enable real-time adaptation to occupant needs while optimizing energy use and indoor environmental quality.[3] This integration relies on standardized communication protocols like BACnet and Modbus, which facilitate interoperability between OCC components and legacy BMS, allowing data exchange for occupancy patterns, comfort preferences, and control signals without requiring full system overhauls.[27] For instance, BACnet (ASHRAE Standard 135) supports peer-to-peer communication across HVAC devices, enabling OCC to adjust setpoints based on occupant data while maintaining syntactic compatibility with diverse vendors.[27]

Retrofitting legacy systems for OCC often involves edge computing devices that bridge older infrastructure with modern IoT sensors, processing occupant data locally to reduce latency and bandwidth demands on central BMS.[28] These devices, such as interoperable supervisory controllers, connect via Modbus or BACnet to existing meters and actuators, supporting occupant-driven adjustments like demand-controlled ventilation in multi-zone setups without extensive rewiring.[29]

System architecture for OCC typically employs a hierarchical structure, with local controllers handling zone-level decisions—such as HVAC setpoint adjustments based on predicted occupancy—and cloud analytics aggregating data for building-wide optimization.[30] In multi-zone buildings, this setup ensures scalability by distributing computational loads; for example, artificial neural networks at the edge predict occupancy per zone, feeding into model predictive control (MPC) layers that coordinate via cloud platforms for adaptive strategies across floors or wings, achieving up to 20% energy savings while preserving thermal comfort.[30][31]

Interoperability challenges arise from heterogeneous data formats and proprietary systems, addressed through standardization efforts like those by the European Committee for Standardization (CEN), including EN 16798-1, which defines parameters for indoor environmental quality and energy performance to support OCC integration in ventilation and thermal controls.[3] API-based integrations with smart thermostats exemplify solutions; ecobee's HTTP API allows BMS to poll and update thermostat data for occupant preferences, while Google's Smart Device Management (SDM) API enables similar control of Nest devices in building networks, facilitating seamless embedding of personal comfort inputs.[32][33]

Security in these integrated networks prioritizes occupant data protection through measures like encryption standards (e.g., for data in transit via VPNs) and Zero Trust architectures, which verify all access to prevent breaches in BMS handling sensitive behavioral patterns.[34] Compliance with frameworks such as ISO/IEC 27001 ensures pseudonymization and risk assessments, mitigating inference risks from occupancy data shared across hierarchical layers.[34][3]

Real-World Case Studies

One prominent example of occupant-centric building controls in an office setting is the implementation in commercial office buildings where sensor-based systems adjust HVAC and lighting based on real-time occupancy and interactions. This approach utilizes presence detection to minimize energy waste from overrides, incorporating operator training for effective deployment. Post-occupancy evaluations (POE) revealed 15-20% savings in HVAC energy consumption alongside 20-25% improvements in thermal comfort satisfaction, while maintaining indoor air quality during periods of variable occupancy, such as hybrid work arrangements influenced by the COVID-19 pandemic.[3]

In residential applications, the Swedish Energy Agency-funded DOCENT project (2024–2027) aims to develop occupant-centric controls in energy-efficient buildings, including apartment testbeds. It focuses on integrating occupant preferences for heating, ventilation, and lighting through intuitive interfaces, behavioral data, mobile apps, and sensors, prioritizing well-being alongside efficiency. The project seeks to demonstrate energy savings in ventilation and heating systems, with planned evaluations of thermal comfort through occupant feedback.[35][3]

Educational facilities have adopted these controls, as seen in university buildings where wearable devices and environmental sensors inform personalized HVAC adjustments. Literature reviews of personal comfort models highlight applications in North American settings, using wearables to predict thermal preferences in shared spaces like classrooms and labs, enabling predictive zoning of air conditioning. These approaches have demonstrated 20-30% reductions in cooling energy use in various studies without compromising occupant comfort, as measured by POE and thermal satisfaction indices.[3][36]

Across these implementations, key lessons emerge from post-occupancy evaluations, which serve as critical metrics for assessing success by quantifying comfort, energy performance, and behavioral adaptation. Scalability challenges persist in diverse climates, where models trained in temperate zones may underperform in extreme conditions without localized data adjustments, underscoring the need for robust, context-aware predictive controls to ensure consistent outcomes.[3][37]

Technologies and Tools

Occupant-centric building controls rely on a suite of hardware, software, and integrated platforms to sense occupant preferences, automate responses, and optimize building environments. These technologies enable real-time data collection from users and systems, facilitating personalized adjustments to lighting, temperature, and ventilation while minimizing energy use. Key components include interconnected IoT devices, algorithmic software, and vendor-supported ecosystems that ensure scalability and interoperability.[3]

Hardware components form the foundational layer for capturing occupant data and effecting control actions. IoT sensors, such as occupancy detectors and environmental monitors, gather information on presence, activity levels, and comfort metrics like temperature and humidity. For instance, Zigbee-enabled thermostats operate on low-power mesh networks, allowing wireless communication for precise HVAC adjustments based on individual preferences. Actuators, including motorized blinds and dampers, enable automated responses to occupant needs, such as shading adjustments to reduce glare or thermal loads. A prominent example is the Philips Hue system, which uses smart lighting bulbs and hubs to personalize illumination based on user schedules and motion detection, integrating with broader building networks.[3][38]

Software platforms provide the orchestration and simulation capabilities essential for implementing these controls. Open-source solutions like OpenHAB offer flexible, rule-based automation that integrates diverse devices into a unified interface, supporting occupant-driven scenarios such as adaptive setpoints for lighting and climate. Commercial platforms, such as Honeywell's Niagara Framework, enable supervisory control and data acquisition (SCADA) for large-scale buildings, allowing real-time analytics on occupant behavior to optimize resource allocation. Simulation software like EnergyPlus, extended with occupant behavior plugins (e.g., the OpenStudio Occupant Behavior Model), models human interactions with building systems to predict and test control strategies before deployment.[3][39][40][41]

AI and machine learning frameworks enhance decision-making by processing occupant data for predictive personalization. Integrations with TensorFlow allow developers to build custom models that learn from historical patterns, such as predicting preferred thermal comfort zones based on user feedback. Cloud services like AWS IoT facilitate scalable data ingestion and edge computing, enabling real-time anomaly detection and control optimization across distributed building sensors. These tools support advanced features like reinforcement learning for dynamic HVAC tuning.[3][42][43]

The vendor ecosystem drives adoption through comprehensive solutions and market expansion. Major players like Siemens, with its Desigo CC platform, and Johnson Controls, via its Metasys system, provide end-to-end occupant-centric controls that integrate sensors, analytics, and actuation for commercial buildings. The global smart building market, including these technologies, was valued at approximately $130 billion as of 2025 projections.[3][44][45][46]

Current Limitations

Occupant-centric building controls face significant technical challenges, particularly in data privacy and model accuracy across diverse contexts. The collection of occupant data through sensors and IoT devices raises substantial privacy risks, as spatiotemporal data on presence, activities, and preferences can enable re-identification or profiling without explicit consent, potentially violating regulations like the EU's General Data Protection Regulation (GDPR). Compliance requires anonymization and pseudonymization techniques, but implementation remains inconsistent, especially in residential settings where only 45% of studies include occupant data due to heightened sensitivities. Additionally, machine learning models for predicting occupant behavior and indoor environmental quality (IEQ) exhibit biases stemming from geographic and demographic limitations in training data, with research predominantly focused on North America, Europe, and China in temperate climates, leading to inaccuracies in non-Western or tropical regions. These biases can result in suboptimal control strategies, such as unreliable thermal comfort predictions for diverse populations, including older adults or those in high-density urban low-income housing.

Implementation barriers further hinder widespread adoption, including elevated upfront costs and interoperability issues. Deploying occupant-centric systems demands denser sensor networks and advanced infrastructure for high-resolution data collection, which increases installation and maintenance expenses compared to traditional building controls, often prioritizing quantitative metrics like energy savings over qualitative occupant needs. In legacy buildings, interoperability gaps arise from incompatible communication protocols and outdated building automation systems (BAS), complicating integration with modern occupant data streams and limiting scalability in retrofits. These challenges are exacerbated in older structures, where standardized guidelines for sequences of operation are lacking, leading to operator resistance and incomplete system fusion.

Performance gaps emerge from over-reliance on predictive models, particularly during unforeseen anomalies. Systems tuned to historical occupancy patterns falter when events like pandemics disrupt routines, as seen during COVID-19 when irregular presence and reduced densities caused prediction errors of up to 15-20% in occupancy forecasts, resulting in either excessive ventilation energy use or elevated infection risks. For instance, model predictive controls assuming steady-state behaviors underestimated stochastic changes, leading to 28-44% variations in airflow rates and coil loads, highlighting vulnerabilities in non-adaptive algorithms. Such failures underscore the need for robust handling of uncertainties, yet current implementations often yield minimal energy savings—despite 24-58% potential—due to occupant overrides and rebound effects.

Ethical concerns, including surveillance implications and equity disparities, pose profound hurdles. Constant monitoring via occupancy sensors and wearables creates a pervasive surveillance environment, where aggregated data infers personal routines and vulnerabilities, eroding occupant autonomy and raising risks of misuse for non-consensual profiling. This "privacy paradox" persists as occupants trade data for comfort benefits but resent implicit tracking, with ethical frameworks demanding transparent consent and opt-out mechanisms that are rarely enforced. Equity issues are evident in low-income buildings, where occupant-centric designs overlook socio-cultural and behavioral diversity, such as adaptive habits constrained by affordability and gender norms, resulting in up to 37% variations in energy use and thermal discomfort in dense urban housing. Access remains limited for these populations, as high costs and rigid infrastructures perpetuate energy poverty without tailored interventions.

Emerging Trends

Recent advancements in artificial intelligence are poised to enhance the transparency and privacy of occupant-centric building controls. Explainable AI (XAI) techniques are being developed to provide interpretable decision-making processes, allowing users to understand how control systems adjust environmental parameters based on occupant preferences, thereby building trust and facilitating regulatory compliance.[47] Federated learning emerges as a key method to address data privacy concerns, enabling models to be trained across multiple buildings without centralizing sensitive occupant data, thus preserving individual privacy while improving control accuracy through collaborative learning.[48]

Integration with smart city infrastructures represents a growing trend, where occupant-centric controls connect to urban energy grids for optimized demand response. By linking building systems to city-wide networks, these controls can dynamically adjust loads based on real-time occupant behavior and grid signals, reducing peak energy use and enhancing grid stability. The advent of 5G technology further enables seamless, low-latency data sharing between buildings and urban systems, supporting instantaneous adjustments to lighting, HVAC, and ventilation in response to occupant movements and city-level events.[49]

Sustainability efforts are increasingly incorporating occupant data into net-zero building designs, with controls tracking individual carbon footprints to inform adaptive strategies that minimize emissions. For instance, systems can use occupant patterns to optimize energy use in renewable-integrated structures, aligning with goals for carbon-neutral operations. IEA research indicates that building standards aligned with net-zero emissions by 2050 will increasingly emphasize data-driven approaches integrating behavioral insights with energy modeling.[50]

Novel input modalities are expanding the scope of occupant-centric controls, particularly through biometric sensors. Advancements in electroencephalography (EEG) devices allow for stress detection, enabling automated adjustments to indoor environments—such as altering lighting or temperature—to mitigate occupant discomfort in real time. Additionally, virtual reality (VR) simulations are gaining traction for pre-design testing, where architects and engineers model occupant interactions in virtual building prototypes to refine control strategies before physical implementation, improving efficiency and satisfaction.

Control algorithms are evolving alongside these trends, incorporating hybrid machine learning models to handle diverse inputs from biometrics and smart city data.

Definition and Principles

Occupant-centric building controls (OCC) refer to indoor climate control systems that prioritize the real-time preferences, behaviors, and comfort needs of building occupants over fixed environmental setpoints, dynamically adjusting parameters such as temperature, lighting, ventilation, and air quality using data from sensors and occupant inputs.[3] OCC strategies are broadly categorized into occupancy-centric methods, which respond to presence or density (e.g., demand-controlled ventilation via CO₂ sensors), and occupant behavior-centric methods, which infer or learn preferences from interactions (e.g., thermostat adjustments) or explicit feedback (e.g., mobile apps). This approach integrates occupant presence, interactions with building elements (e.g., thermostat adjustments or window openings), and indoor environmental quality (IEQ) metrics to enable personalized environmental conditions while optimizing energy efficiency.[1] Unlike passive assumptions of uniform occupant needs, OCC treats individuals as active participants in the control loop, fostering adaptive responses that enhance satisfaction and well-being.[4]

The core principles of OCC emphasize personalization, adaptability, and sustainability. Personalization tailors environmental controls to individual or group-specific needs, such as varying thermal preferences based on activity levels or demographics, achieved through explicit feedback (e.g., via mobile apps) or implicit inference from behaviors.[3] Adaptability involves real-time responsiveness to occupancy patterns, such as scaling ventilation to occupant density or adjusting lighting to presence, using data-driven models to predict and react to changes without rigid schedules. Key technologies include explicit sensors (e.g., motion detectors, Wi-Fi counters, wearables) for direct detection and implicit methods (e.g., inferring from CO₂ levels or energy patterns).[1] Sustainability is embedded by balancing comfort with energy optimization, reducing waste from over-conditioning unoccupied spaces or mismatched setpoints, often yielding 20-50% energy savings in simulated scenarios while maintaining or improving IEQ.[3]

In contrast to traditional building controls, which rely on rule-based or scheduled operations with static setpoints (e.g., fixed HVAC timetables assuming constant occupancy), OCC incorporates continuous feedback loops to incorporate occupant agency.[5] Traditional systems often treat occupants as uniform heat sources, leading to inefficiencies like simultaneous heating and cooling due to overrides, whereas OCC uses sensors for IEQ monitoring, occupant inputs for preferences, and actuators for adjustments, creating a closed-loop system.[3] A basic representation of this feedback loop is as follows:

This occupant-in-the-loop dynamic reduces performance gaps from behavioral mismatches, which can account for up to 80% variance in energy use.[3]

The principles of OCC trace their origins to responses to the 1970s energy crisis, when research highlighted the critical role of human behavior in building energy consumption, shifting focus from purely technical solutions to occupant-inclusive strategies.[3] Practical emergence accelerated in the 2000s with advancements in Internet of Things (IoT) technologies, enabling affordable sensors and real-time data integration for occupant-focused automation.[5]

Importance and Benefits

Occupant-centric building controls offer substantial energy savings by aligning HVAC operations with real-time occupancy patterns rather than fixed schedules, potentially reducing energy use by 19% to 45% in office buildings across various U.S. climate zones.[6] Field studies, numbering over 35 worldwide as of 2023, report 24–58% savings in HVAC for presence-based controls and up to 37% overall through integrated systems, with 65% reductions in thermal discomfort.[1] This approach, informed by ASHRAE guidelines, optimizes ventilation, airflow, and temperature setpoints, with higher savings in heating-dominated and marine climates where resets for unoccupied spaces yield greater impacts.[7] Such reductions contribute to lower operational costs, with studies indicating a return on investment within two years for occupancy-based systems in commercial settings.[8]

Beyond efficiency, these controls enhance occupant well-being by improving thermal comfort and indoor environmental quality, thereby reducing instances of sick building syndrome and associated health issues.[9] Research from Lawrence Berkeley National Laboratory estimates that better indoor environments, achievable through responsive controls, could yield annual productivity gains valued at $20 to $160 billion nationwide by minimizing health-related absences and boosting direct performance.[9] This translates to tangible benefits like fewer sick days and higher cognitive function, fostering healthier work environments. Post-COVID-19 adaptations, such as variable ventilation for health compliance in hybrid work settings using implicit occupancy data (e.g., plug loads), further underscore OCC's role.[1]

On a sustainability level, occupant-centric controls lower carbon footprints via demand-responsive strategies that curtail unnecessary energy consumption, supporting global efforts to decarbonize buildings.[7] Economically, the associated cost savings and extended equipment life further amplify their value, often achieving payback periods of two years or less. Challenges include privacy concerns with occupant data (addressed via pseudonymization under regulations like GDPR), the need for building-specific sensor calibration, and training for operators to integrate quantitative (e.g., energy metrics) and qualitative (e.g., satisfaction surveys) assessments.[1][3]

Broader societal impacts include stronger alignment with green building standards, where these controls help achieve LEED certification through energy optimization and WELL certification via enhanced occupant health metrics, such as 39% higher satisfaction rates in WELL buildings compared to LEED.[10] By prioritizing human-centered design, they promote equitable, resilient built environments that balance efficiency with user needs.[11]

Early Development

The origins of occupant-centric building controls can be traced to the 1970s, amid the global oil crises that spurred efforts to optimize energy use in buildings. Early developments focused on demand-controlled ventilation (DCV) systems, which adjusted airflow based on occupancy to reduce energy consumption without compromising indoor air quality. These systems emerged as a direct response to the 1973-1974 energy crisis, leveraging CO2 sensors to modulate ventilation rates in real-time, marking an initial shift toward incorporating occupant presence into building operations.[12]

A foundational contribution during this period was P.O. Fanger's Predicted Mean Vote (PMV) model, introduced in 1970, which quantified thermal comfort based on occupant physiological responses to environmental factors like temperature, humidity, and air velocity. The PMV model provided a predictive framework for assessing how building controls could align with human thermal sensations, influencing subsequent standards for occupant-focused HVAC design.

In the 1990s, progress accelerated with the integration of occupant feedback mechanisms in emerging smart building concepts, facilitated by basic occupancy sensors and environmental monitors. European initiatives, such as the 1992 European Indoor Air Quality Audit Project, employed trained sensory panels alongside instrumental measurements to evaluate occupant perceptions in 56 office buildings, laying groundwork for feedback-driven control adjustments. These efforts highlighted the value of direct occupant input in refining ventilation and thermal systems, though prototypes remained limited to research settings.[13]

The 2000s marked a breakthrough through the adoption of wireless sensors and nascent Internet of Things (IoT) technologies, enabling more dynamic occupant modeling in building controls. A pivotal publication, Bourgeois et al.'s 2005 paper, analyzed the energy impacts of occupant behavioral responses to lighting systems, demonstrating how detailed stochastic models of occupancy patterns could improve simulation accuracy and control efficacy. This work underscored the need for probabilistic approaches to capture variability in occupant actions, paving the way for integrated systems.[14]

A key milestone came with the 2005 revision of ISO 7730, which established ergonomics standards for moderate thermal environments by incorporating occupant data into PMV-based assessments, including provisions for personalized comfort zones. This standard formalized the use of occupant-centric metrics in design and operation, influencing global practices for energy-efficient buildings.

Key Theoretical Concepts

Occupant-centric building controls are grounded in human-centric models that prioritize individual thermal comfort over uniform environmental conditions. A foundational approach is Fanger's Predicted Mean Vote (PMV) model, which quantifies thermal sensation on a seven-point scale from -3 (cold) to +3 (hot), with zero representing neutral comfort. The PMV is calculated using the equation:

where MMM is the metabolic rate and LLL is the thermal load on the body, incorporating factors such as air temperature, mean radiant temperature, relative humidity, air velocity, clothing insulation, and activity level.[15] This model underpins the Predicted Percentage of Dissatisfied (PPD) index, which estimates the proportion of occupants likely to be dissatisfied with thermal conditions, typically aiming for PPD below 10% in controlled environments.[15] Complementing this, the adaptive comfort model by de Dear and Brager accounts for behavioral adaptation and contextual influences, positing that occupants in naturally ventilated buildings adjust expectations based on outdoor climate, allowing wider indoor temperature ranges (e.g., 23–26°C in temperate zones) while maintaining satisfaction.[16]

Behavioral modeling in occupant-centric controls addresses variability in occupancy patterns and individual preferences, essential for dynamic system responses. Concepts such as occupancy patterns capture presence and movement rhythms, often modeled stochastically to predict absences or relocations with probabilistic accuracy. Preference elicitation involves gathering occupant feedback on comfort parameters to personalize controls, enabling systems to adapt to heterogeneous needs. A key method employs Markov chains to simulate occupant presence as a sequence of states (e.g., occupied or vacant), where transition probabilities reflect empirical data from sensor observations, achieving prediction errors as low as 5–10% in office settings.[17]

Feedback theory frames occupant-centric controls as closed-loop systems, where occupant satisfaction serves as the dynamic setpoint rather than fixed thresholds. In these systems, sensors and algorithms continuously monitor environmental conditions and user inputs, adjusting actuators (e.g., HVAC) to minimize deviation from comfort targets, akin to proportional-integral-derivative (PID) controllers but augmented with occupant data. This approach integrates the PMV equation as a core metric for real-time feedback, ensuring iterative refinements toward optimal satisfaction.[15]

The integration of psychology enhances these models by incorporating cognitive factors that influence perceived comfort, such as expectations, control over the environment, and emotional states. Drawing from environmental psychology, research highlights how perceived controllability—e.g., ability to adjust personal settings—can increase tolerance for discomfort by up to 2–3°C deviations from neutral temperatures. Cognitive appraisals, including past experiences and social context, modulate thermal sensations, underscoring the need for holistic models beyond physiological metrics.

Occupant-Centric Control Inputs

Occupant-centric building controls rely on a variety of sensor-based inputs to capture real-time data about environmental conditions and occupant physiology, enabling responsive adjustments to indoor environments. Environmental sensors, such as those measuring CO2 levels, humidity, and motion, are commonly deployed to detect occupancy patterns and indoor air quality, informing ventilation and lighting controls. For instance, CO2 sensors help identify occupant density by correlating elevated levels with presence, while motion detectors track movement to optimize energy use in unoccupied zones. Wearable devices provide physiological data, like heart rate variability, to infer individual thermal preferences, allowing for personalized comfort adjustments beyond standard models. Bluetooth beacons facilitate location tracking within buildings, enabling systems to associate specific occupants with zones for targeted environmental control.[18][19]

Direct user inputs empower occupants to actively communicate preferences, enhancing control personalization through intuitive interfaces. Mobile applications often incorporate voting mechanisms where users rate thermal or lighting conditions on scales like "too hot/cold" or "too bright/dim," aggregating feedback to guide system responses. Voice commands and gesture-based interfaces, integrated into smart devices, allow hands-free adjustments, such as requesting temperature changes or dimming lights, making controls accessible in diverse settings. Privacy considerations are paramount in collecting these inputs, with techniques like data anonymization—such as pseudonymization or aggregation—employed to protect user identities while retaining utility for control decisions. These methods ensure compliance with regulations and mitigate risks of surveillance perception.

Indirect inputs derive from observed occupant behaviors, providing implicit signals of comfort needs without explicit feedback. Actions such as opening windows or manually adjusting thermostats reveal preferences for fresh air or temperature shifts, often captured via integrated sensors on fixtures. These behavioral traces inform models of habitual patterns, allowing controls to anticipate and automate similar adjustments. Data fusion techniques combine these indirect signals with other sources for robust inference; for example, basic Kalman filtering smooths noisy behavioral data by estimating states from sequential observations, improving reliability without complex derivations.

Ensuring input accuracy poses significant challenges, particularly in handling data noise and accounting for occupant heterogeneity. Sensor readings can be affected by environmental interference, such as drafts skewing humidity measurements or transient motions triggering false positives in detectors, necessitating robust filtering to maintain precision. Occupant variability further complicates this, as metabolic rates—key to thermal comfort—differ by factors like age and gender; for example, older adults and females often exhibit 5% lower rates, leading to divergent comfort thresholds that standard controls may overlook. These issues underscore the need for adaptive strategies that interpret inputs within theoretical models of human physiology.[20]

Predictive Controls

Predictive controls in occupant-centric building systems anticipate future occupant behaviors and environmental needs by leveraging forecasting models, enabling proactive adjustments to systems like HVAC and lighting rather than reacting to real-time conditions. These methods rely on predictive analytics to forecast variables such as occupancy levels, thermal comfort preferences, and energy demands, optimizing building performance while maintaining occupant satisfaction. By integrating data from sensors, historical records, and external sources, predictive controls can pre-emptively modulate indoor environments, reducing energy waste and enhancing responsiveness.

Forecasting techniques form the core of predictive controls, employing statistical and machine learning models tailored to building data. For occupancy prediction, autoregressive integrated moving average (ARIMA) models are widely used, capturing temporal patterns in occupant presence. An ARIMA(p,d,q) model involves applying the ARMA(p,q) structure to the d-th differenced series for stationarity and is expressed as:

where ∇dyt\nabla^d y_t∇dyt​ is the d-th difference of the series at time ttt, ppp denotes the autoregressive order, ddd the degree of differencing, qqq the moving average order, βi\beta_iβi​ are autoregressive coefficients, θj\theta_jθj​ are moving average coefficients, and ϵt\epsilon_tϵt​ is white noise. This formulation allows the model to extrapolate future occupancy based on past observations, with applications in commercial buildings showing improved prediction accuracy over baseline methods. Neural networks, such as long short-term memory (LSTM) architectures, extend this to comfort forecasting by processing multivariate time-series data, including occupant feedback and physiological signals, to predict personalized thermal preferences. These deep learning approaches handle non-linear relationships effectively, outperforming traditional models in dynamic environments.

Data-driven prediction enhances these techniques by incorporating historical patterns alongside external factors like weather forecasts and occupancy schedules to pre-adjust building systems. For instance, models trained on past usage data can predict peak occupancy periods, enabling HVAC pre-cooling or heating to maintain comfort without overconsumption. Studies evaluating such systems report mean absolute error (MAE) values as low as 0.5 occupants for short-term forecasts, leading to approximately 20% greater energy savings compared to reactive controls in office settings. This predictive capability is particularly valuable in variable climates, where integrating real-time weather APIs allows models to adjust ventilation proactively.

Hybrid approaches combine rule-based predictions with artificial intelligence to balance interpretability and accuracy, often applied in lighting and HVAC systems. Rule-based elements might use predefined thresholds from building codes or schedules to initiate predictions, while AI refines them with learned patterns from occupant interactions. In lighting applications, hybrids forecast natural daylight availability to dim artificial sources preemptively, reducing energy use by up to 15% in testbed facilities. For HVAC, these methods enable zone-specific pre-heating based on anticipated arrivals, as demonstrated in smart office implementations where occupant comfort scores improved alongside efficiency gains.

The evolution of predictive controls has progressed from simple rule-based predictors in the early 2010s, which relied on static schedules for basic anticipatory adjustments, to sophisticated deep learning frameworks in the 2020s that incorporate real-time learning and multi-modal data fusion. This shift has been driven by advances in computational power and sensor affordability, enabling scalable deployment in diverse building types.

Control Algorithms

Control algorithms in occupant-centric building controls form the core computational logic that processes occupant data—such as preferences, behaviors, and real-time feedback—into actionable decisions for systems like HVAC, lighting, and shading. These algorithms aim to balance multiple objectives, including energy efficiency, thermal comfort, and occupant satisfaction, by optimizing control variables over time horizons. Unlike traditional rule-based controls, they incorporate dynamic models of occupant needs and building dynamics to enable adaptive, personalized responses.[21]

Optimization-based algorithms are widely used for multi-objective tradeoffs in occupant-centric controls, often employing techniques like genetic algorithms to search for solutions that minimize conflicts between energy use and comfort metrics. For instance, genetic algorithms evolve populations of potential control strategies through selection, crossover, and mutation processes, evaluating fitness based on occupant-reported satisfaction and simulated energy consumption. This approach has been applied to optimize shading and ventilation in office settings, achieving up to 20% energy savings while maintaining thermal comfort within personalized bounds.[22] A key formulation in such optimizations is the cost function minimization:

where JJJ is the total cost, EEE represents energy consumption, DDD quantifies discomfort (e.g., deviation from preferred temperature setpoints), and w1,w2w_1, w_2w1​,w2​ are weighting factors tuned to occupant priorities. This objective is typically solved using linear programming when constraints are linear, such as bounds on temperature and airflow rates. The solution process involves: (1) defining decision variables (e.g., setpoint adjustments), (2) formulating inequality constraints (e.g., Tmin⁡≤T≤Tmax⁡T_{\min} \leq T \leq T_{\max}Tmin​≤T≤Tmax​), (3) minimizing JJJ subject to these constraints via a solver like simplex method, and (4) iterating over prediction horizons. Pseudocode for a basic linear programming implementation in occupant control might resemble:

This method ensures feasible, occupant-aligned solutions but requires accurate linear approximations of nonlinear building behaviors.[23]

Reinforcement learning (RL) algorithms treat the building as an environment where an agent learns optimal policies through trial-and-error interactions, using rewards derived from occupant satisfaction scores, such as reduced thermal discomfort or improved air quality preferences. In occupant-centric applications, deep RL variants like Q-learning or actor-critic methods model states including occupancy patterns and external weather, enabling the agent to adapt controls like window opening or lighting dimming without explicit programming. Experimental deployments in residential buildings have demonstrated RL reducing energy use by 15-30% compared to static controls, while learning individualized behaviors over weeks of operation. Challenges include the need for safe exploration during learning to avoid discomfort spikes.[24]

Integration with Building Systems

Occupant-centric building controls (OCC) integrate with existing building infrastructure, such as heating, ventilation, and air conditioning (HVAC) systems and building management systems (BMS), to enable real-time adaptation to occupant needs while optimizing energy use and indoor environmental quality.[3] This integration relies on standardized communication protocols like BACnet and Modbus, which facilitate interoperability between OCC components and legacy BMS, allowing data exchange for occupancy patterns, comfort preferences, and control signals without requiring full system overhauls.[27] For instance, BACnet (ASHRAE Standard 135) supports peer-to-peer communication across HVAC devices, enabling OCC to adjust setpoints based on occupant data while maintaining syntactic compatibility with diverse vendors.[27]

Retrofitting legacy systems for OCC often involves edge computing devices that bridge older infrastructure with modern IoT sensors, processing occupant data locally to reduce latency and bandwidth demands on central BMS.[28] These devices, such as interoperable supervisory controllers, connect via Modbus or BACnet to existing meters and actuators, supporting occupant-driven adjustments like demand-controlled ventilation in multi-zone setups without extensive rewiring.[29]

System architecture for OCC typically employs a hierarchical structure, with local controllers handling zone-level decisions—such as HVAC setpoint adjustments based on predicted occupancy—and cloud analytics aggregating data for building-wide optimization.[30] In multi-zone buildings, this setup ensures scalability by distributing computational loads; for example, artificial neural networks at the edge predict occupancy per zone, feeding into model predictive control (MPC) layers that coordinate via cloud platforms for adaptive strategies across floors or wings, achieving up to 20% energy savings while preserving thermal comfort.[30][31]

Interoperability challenges arise from heterogeneous data formats and proprietary systems, addressed through standardization efforts like those by the European Committee for Standardization (CEN), including EN 16798-1, which defines parameters for indoor environmental quality and energy performance to support OCC integration in ventilation and thermal controls.[3] API-based integrations with smart thermostats exemplify solutions; ecobee's HTTP API allows BMS to poll and update thermostat data for occupant preferences, while Google's Smart Device Management (SDM) API enables similar control of Nest devices in building networks, facilitating seamless embedding of personal comfort inputs.[32][33]

Security in these integrated networks prioritizes occupant data protection through measures like encryption standards (e.g., for data in transit via VPNs) and Zero Trust architectures, which verify all access to prevent breaches in BMS handling sensitive behavioral patterns.[34] Compliance with frameworks such as ISO/IEC 27001 ensures pseudonymization and risk assessments, mitigating inference risks from occupancy data shared across hierarchical layers.[34][3]

Real-World Case Studies

One prominent example of occupant-centric building controls in an office setting is the implementation in commercial office buildings where sensor-based systems adjust HVAC and lighting based on real-time occupancy and interactions. This approach utilizes presence detection to minimize energy waste from overrides, incorporating operator training for effective deployment. Post-occupancy evaluations (POE) revealed 15-20% savings in HVAC energy consumption alongside 20-25% improvements in thermal comfort satisfaction, while maintaining indoor air quality during periods of variable occupancy, such as hybrid work arrangements influenced by the COVID-19 pandemic.[3]

In residential applications, the Swedish Energy Agency-funded DOCENT project (2024–2027) aims to develop occupant-centric controls in energy-efficient buildings, including apartment testbeds. It focuses on integrating occupant preferences for heating, ventilation, and lighting through intuitive interfaces, behavioral data, mobile apps, and sensors, prioritizing well-being alongside efficiency. The project seeks to demonstrate energy savings in ventilation and heating systems, with planned evaluations of thermal comfort through occupant feedback.[35][3]

Educational facilities have adopted these controls, as seen in university buildings where wearable devices and environmental sensors inform personalized HVAC adjustments. Literature reviews of personal comfort models highlight applications in North American settings, using wearables to predict thermal preferences in shared spaces like classrooms and labs, enabling predictive zoning of air conditioning. These approaches have demonstrated 20-30% reductions in cooling energy use in various studies without compromising occupant comfort, as measured by POE and thermal satisfaction indices.[3][36]

Across these implementations, key lessons emerge from post-occupancy evaluations, which serve as critical metrics for assessing success by quantifying comfort, energy performance, and behavioral adaptation. Scalability challenges persist in diverse climates, where models trained in temperate zones may underperform in extreme conditions without localized data adjustments, underscoring the need for robust, context-aware predictive controls to ensure consistent outcomes.[3][37]

Technologies and Tools

Occupant-centric building controls rely on a suite of hardware, software, and integrated platforms to sense occupant preferences, automate responses, and optimize building environments. These technologies enable real-time data collection from users and systems, facilitating personalized adjustments to lighting, temperature, and ventilation while minimizing energy use. Key components include interconnected IoT devices, algorithmic software, and vendor-supported ecosystems that ensure scalability and interoperability.[3]

Hardware components form the foundational layer for capturing occupant data and effecting control actions. IoT sensors, such as occupancy detectors and environmental monitors, gather information on presence, activity levels, and comfort metrics like temperature and humidity. For instance, Zigbee-enabled thermostats operate on low-power mesh networks, allowing wireless communication for precise HVAC adjustments based on individual preferences. Actuators, including motorized blinds and dampers, enable automated responses to occupant needs, such as shading adjustments to reduce glare or thermal loads. A prominent example is the Philips Hue system, which uses smart lighting bulbs and hubs to personalize illumination based on user schedules and motion detection, integrating with broader building networks.[3][38]

Software platforms provide the orchestration and simulation capabilities essential for implementing these controls. Open-source solutions like OpenHAB offer flexible, rule-based automation that integrates diverse devices into a unified interface, supporting occupant-driven scenarios such as adaptive setpoints for lighting and climate. Commercial platforms, such as Honeywell's Niagara Framework, enable supervisory control and data acquisition (SCADA) for large-scale buildings, allowing real-time analytics on occupant behavior to optimize resource allocation. Simulation software like EnergyPlus, extended with occupant behavior plugins (e.g., the OpenStudio Occupant Behavior Model), models human interactions with building systems to predict and test control strategies before deployment.[3][39][40][41]

AI and machine learning frameworks enhance decision-making by processing occupant data for predictive personalization. Integrations with TensorFlow allow developers to build custom models that learn from historical patterns, such as predicting preferred thermal comfort zones based on user feedback. Cloud services like AWS IoT facilitate scalable data ingestion and edge computing, enabling real-time anomaly detection and control optimization across distributed building sensors. These tools support advanced features like reinforcement learning for dynamic HVAC tuning.[3][42][43]

The vendor ecosystem drives adoption through comprehensive solutions and market expansion. Major players like Siemens, with its Desigo CC platform, and Johnson Controls, via its Metasys system, provide end-to-end occupant-centric controls that integrate sensors, analytics, and actuation for commercial buildings. The global smart building market, including these technologies, was valued at approximately $130 billion as of 2025 projections.[3][44][45][46]

Current Limitations

Occupant-centric building controls face significant technical challenges, particularly in data privacy and model accuracy across diverse contexts. The collection of occupant data through sensors and IoT devices raises substantial privacy risks, as spatiotemporal data on presence, activities, and preferences can enable re-identification or profiling without explicit consent, potentially violating regulations like the EU's General Data Protection Regulation (GDPR). Compliance requires anonymization and pseudonymization techniques, but implementation remains inconsistent, especially in residential settings where only 45% of studies include occupant data due to heightened sensitivities. Additionally, machine learning models for predicting occupant behavior and indoor environmental quality (IEQ) exhibit biases stemming from geographic and demographic limitations in training data, with research predominantly focused on North America, Europe, and China in temperate climates, leading to inaccuracies in non-Western or tropical regions. These biases can result in suboptimal control strategies, such as unreliable thermal comfort predictions for diverse populations, including older adults or those in high-density urban low-income housing.

Implementation barriers further hinder widespread adoption, including elevated upfront costs and interoperability issues. Deploying occupant-centric systems demands denser sensor networks and advanced infrastructure for high-resolution data collection, which increases installation and maintenance expenses compared to traditional building controls, often prioritizing quantitative metrics like energy savings over qualitative occupant needs. In legacy buildings, interoperability gaps arise from incompatible communication protocols and outdated building automation systems (BAS), complicating integration with modern occupant data streams and limiting scalability in retrofits. These challenges are exacerbated in older structures, where standardized guidelines for sequences of operation are lacking, leading to operator resistance and incomplete system fusion.

Performance gaps emerge from over-reliance on predictive models, particularly during unforeseen anomalies. Systems tuned to historical occupancy patterns falter when events like pandemics disrupt routines, as seen during COVID-19 when irregular presence and reduced densities caused prediction errors of up to 15-20% in occupancy forecasts, resulting in either excessive ventilation energy use or elevated infection risks. For instance, model predictive controls assuming steady-state behaviors underestimated stochastic changes, leading to 28-44% variations in airflow rates and coil loads, highlighting vulnerabilities in non-adaptive algorithms. Such failures underscore the need for robust handling of uncertainties, yet current implementations often yield minimal energy savings—despite 24-58% potential—due to occupant overrides and rebound effects.

Ethical concerns, including surveillance implications and equity disparities, pose profound hurdles. Constant monitoring via occupancy sensors and wearables creates a pervasive surveillance environment, where aggregated data infers personal routines and vulnerabilities, eroding occupant autonomy and raising risks of misuse for non-consensual profiling. This "privacy paradox" persists as occupants trade data for comfort benefits but resent implicit tracking, with ethical frameworks demanding transparent consent and opt-out mechanisms that are rarely enforced. Equity issues are evident in low-income buildings, where occupant-centric designs overlook socio-cultural and behavioral diversity, such as adaptive habits constrained by affordability and gender norms, resulting in up to 37% variations in energy use and thermal discomfort in dense urban housing. Access remains limited for these populations, as high costs and rigid infrastructures perpetuate energy poverty without tailored interventions.

Emerging Trends

Recent advancements in artificial intelligence are poised to enhance the transparency and privacy of occupant-centric building controls. Explainable AI (XAI) techniques are being developed to provide interpretable decision-making processes, allowing users to understand how control systems adjust environmental parameters based on occupant preferences, thereby building trust and facilitating regulatory compliance.[47] Federated learning emerges as a key method to address data privacy concerns, enabling models to be trained across multiple buildings without centralizing sensitive occupant data, thus preserving individual privacy while improving control accuracy through collaborative learning.[48]

Integration with smart city infrastructures represents a growing trend, where occupant-centric controls connect to urban energy grids for optimized demand response. By linking building systems to city-wide networks, these controls can dynamically adjust loads based on real-time occupant behavior and grid signals, reducing peak energy use and enhancing grid stability. The advent of 5G technology further enables seamless, low-latency data sharing between buildings and urban systems, supporting instantaneous adjustments to lighting, HVAC, and ventilation in response to occupant movements and city-level events.[49]

Sustainability efforts are increasingly incorporating occupant data into net-zero building designs, with controls tracking individual carbon footprints to inform adaptive strategies that minimize emissions. For instance, systems can use occupant patterns to optimize energy use in renewable-integrated structures, aligning with goals for carbon-neutral operations. IEA research indicates that building standards aligned with net-zero emissions by 2050 will increasingly emphasize data-driven approaches integrating behavioral insights with energy modeling.[50]

Novel input modalities are expanding the scope of occupant-centric controls, particularly through biometric sensors. Advancements in electroencephalography (EEG) devices allow for stress detection, enabling automated adjustments to indoor environments—such as altering lighting or temperature—to mitigate occupant discomfort in real time. Additionally, virtual reality (VR) simulations are gaining traction for pre-design testing, where architects and engineers model occupant interactions in virtual building prototypes to refine control strategies before physical implementation, improving efficiency and satisfaction.

Control algorithms are evolving alongside these trends, incorporating hybrid machine learning models to handle diverse inputs from biometrics and smart city data.