Ontology-Based Data Representation Prototype for Indoor Air Quality, Building Energy Performance, and Health Data Computation

Abstract

This study investigates the data integration of IoT-enabled sensor networks, emphasizing energy performance and personalized indoor air quality (IAQ) solutions to improve indoor environments, energy efficiency, and sustainability. Ontologies—structured frameworks that standardize data representation and enable interoperability—are the tools for interpreting complex IAQ data for optimal energy rate plans and health situations. Our methodology follows the well-established three-phase engineering approach. We present the design of a prototype with essential classes, which is proposed to integrate IAQ data with health conditions, enhancing real-time monitoring and automated decision making for optimal energy performance for smart buildings. Our research goal is to define the most essential classes, arranging them hierarchically to create a prototype for data computing covering IAQ, energy performance, and health aspects. This ontological framework, covering all three aspects, addresses a current research gap. Results demonstrate the minimum viable product with 78 classes for a smart home IoT system, providing tailored indoor climate control based on user health profiles and energy performance. This prototype represents a significant advancement in sustainable building and IAQ management, promising improved building energy performance, occupant health, and comfort. Future research will validate this framework through extensive testing in real-world environments.

1. Introduction

The digitalization of buildings and the integration of Internet of Things (IoT) devices are rapidly advancing, enabling more autonomous, efficient, and flexible monitoring solutions. Real-time indoor air quality (IAQ) monitoring solutions offer individuals valuable insights into the air they breathe within their homes and workplaces. IoT integration optimizes energy efficiency by controlling connected objects such as thermostats, air handling systems, air-conditioning devices, windows, and household appliances. This empowers dwellers to make informed and greener lifestyle choices and take preventive measures to create healthier and more sustainable living environments. The future of indoor air quality monitoring is expected to move toward personalized and energy-efficient solutions. Solutions utilizing data from and performing actuation with IoT devices will be tailored to individual preferences and health requirements and appliances’ energy preferences, allowing for a more customized and effective approach to energy and air quality management. The integration of IoT devices in smart homes and commercial buildings has seen a surge, driven by the desire for increased connectivity and automation. We foresee that IoT-enabled sensor networks will form the backbone of monitoring systems solutions. These sensors can detect a myriad of indoor air pollutants, energy consumption, and more. IoT devices are designed for seamless connectivity and interoperability. This allows for the integration of diverse sensors and devices into a unified system, creating a holistic approach to real-time IAQ management while meeting the building energy performance goals. In buildings with existing automation, the novel IoT sensing systems can be integrated with existing automation, augmenting their sensing capability, resulting in an integrated system that merges the positive features of building automation and flexible IoT sensing and energy management control.

Owing to their robust expressiveness and capacity for supporting reasoning, ontologies emerge as a pre-requisite, being a viable and advantageous solution for the modeling and interpretation of sensor data—an imperative undertaking within the data life cycle of IoT solutions [1,2]. Ontologies, in the context of IoT, energy performance, and IAQ, refer to structured frameworks that define the relationships between different entities and concepts in the domain. They provide a common understanding and vocabulary for interpreting data. Ontologies standardize the representation of knowledge, facilitating communication between different IoT devices and systems or static datasets that could also potentially estimate IAQ-related health hazards. This standardized representation enhances system reliability and efficiency. Ontologies bridge the gap between diverse data sources, like devices’ energy consumption, toxicology, health pre-conditions, and IAQ levels, by providing a common framework for interpretation. This is particularly valuable in complex environments where multiple sensors measure various parameters, complex algorithms are executed for computing control decisions, and static data needs to be considered for automated decision making.

Ontologies have become viable and effective tools for practitioners to develop applications requiring data and process interoperability, big data management, and automated reasoning on knowledge [3]. Research on ontologies capturing energy performance is well documented, and validated prototypes have been presented [4,5]. The intersection of IAQ monitoring and health effects ontologies has just recently garnered significant attention [6]. While engineers in companies are combining IAQ with personalized health settings and green building management in an ad hoc manner, systematic research is needed to cover energy efficiency and IAQ. This can only be done effectively if comprehensive ontologies exist to organize the source data and the outputs of individual algorithms. Advances in sensor technologies have revolutionized ambient air quality assessment, generating vast datasets that must be organized for analysis using a suitable ontology. Ontologies, which are semantic frameworks that capture complex relationships, simplify the integration and analysis of diverse data sources. An ontology defines a common vocabulary for researchers who need to share information in a domain. It includes machine-readable and interpretable definitions of basic concepts in the domain and relations among them. These ontologies not only enhance data interoperability but also enable complex data analytics, supporting the identification of correlations between air quality parameters, energy performance, and health preferences. Ideal prototypes avoid unnecessary complexity and, nevertheless, are capable of covering all data collection needs for healthier and sustainable living choices [7]. The evolving landscape emphasizes the current challenge for standardized representations, spurring the development of ontologies to bridge the semantic gap [8]. Our research goal is to combine the scattered research in the domain and to present a prototype with fundamental classes to capture IAQ, energy performance, and health aspects. To bridge the research gap, our research question is the following:

What are the essential classes and their properties required to construct a comprehensive ontology framework for prototyping the integration of building energy performance, indoor air quality, and occupant health data?

Section 2 explores the current trends, developments, and challenges in ontology-driven research, emphasizing the already existing and well-documented frameworks for energy performance but the growing demand for personalized ontological frameworks, explaining the complex connections between air quality and human health. Asthma, allergies, and respiratory diseases, for instance, can be tackled by intervening directly in air quality, temperature, and humidity (among other parameters in indoor spaces). As biomedical research continues to unfold, the integration of advanced ontologies stands as a key driver in shaping the future of ambient air quality assessment and its implications for public health [9]. This research provides an overview of the state of the art in the dynamic IAQ, energy performance, and health field in the next chapter and introduces a semantic-based framework approach in Section 3. Section 5 presents the application web interface and suggests future research. The main contribution is the ontological structure proposed in Section 4. The prototype is a possible toolkit for combining energy performance and IAQ data with health pre-conditions and beyond. A novel ontological structure for IAQ customization inside a living environment developed for the Evidence Driven Indoor Air Quality Improvement (EDIAQI) project is summarized in the last chapter.

2. State of the Art

As the artificial intelligence literature contains many definitions of ontology [10], especially in the IAQ and health domain, many of these contradict one another [6,11,12]. This study reviewed research focusing on building energy performance, IAQ parameters (chemical, biological, and environmental), and secondly on dwellers (health situations and/or activities).

Research on ontologies for capturing building energy performance data in smart buildings and controlling for indoor comfort metrics aims to create standardized frameworks to improve data integration, interoperability, and advanced analytics. These well-documented frameworks facilitate efficient energy management and enhance building sustainability by focusing on sensor data representation, energy consumption patterns, indoor comfort metrics, and heating and ventilation system optimization [4,5].

In the fields of environmental health, toxicology, and IAQ, scattered endeavors have been undertaken to formulate ontologies tailored to distinct facets of sensor technology and health considerations. Notably, a review of IoT ontologies conducted in 2021, encompassing a total of 43 papers, revealed a prevalent focus on air quality sensors. Specifically, 11 contributions were identified in the form of ontologies designed to establish a cohesive bridge between health-related ontologies and those pertinent to the IoT [9,13]. There is a documented demand for the integration of environmental health into interoperable data resources using ontologies [14,15], as such standardized frameworks related to building and indoor space management do not exist so far. The Open Biomedical Ontologies Foundry is a collaborative effort to develop ontologies for the life sciences that include terms relevant to toxicology but have no connection to pollutant exposure. The National Center for Biomedical Ontology in the United States has also been involved in projects related to ontologies in the biomedical domain, but this work does not cover the IAQ exposure classes either. The most advanced semantics research that has emerged recently has made advances in connecting environmental exposure (unfortunately not exclusive to IAQ) to toxicology and human health [16]. The Environmental Conditions, Treatments, and Exposures Ontology connects toxicology and exposure to human health and beyond and is designed for interoperability, reuse, and axiomatization with other ontologies and holds promise for future research for automated building management systems. Its terms have been utilized in axioms within the Mondo Disease Ontology to represent diseases caused or influenced by environmental factors, as well as for survey encoding for the Personalized Environment and Genes Study. Nevertheless, the recent advancements in ontology engineering, connecting toxicology and environmental conditions and possible health effects, and controlling for lifestyle factors are not specific to IAQ and/or green building management systems.

Semantic representations for real-time air quality data monitoring [17,18], as well as specific health outcomes related to air quality, e.g., human cognition and physiology [19], have been documented. However, in the context of IAQ, there is a lack of guidance on what to measure, where to measure, how frequently to measure, and standards for acceptable measurement quality on all scales of application [20]. There is a need to develop a robust, scalable, and consistent IAQ ontology domain model that plays an imminent role not only in the advancement of IAQ monitoring systems but also serves as an element in the related health risks, as just scattered research exists, but no standard approach has been developed so far [21]. Currently, efforts for standard semantics and guidelines are distributed across various research institutions, environmental agencies, and public health organizations. These initiatives have been focusing on creating various, not fully harmonized frameworks to represent different aspects of IAQ, including pollutant levels [22], ventilation rates, and health-related parameters [23]. Recent research in the IAQ semantics domain includes cases of using type-2 fuzzy logic to express ambiguity and describe vague data in an ontological manner. The proposed type-2 fuzzy theory for IAQ ontologies by Ghorbani and Zamanifar [24] can improve semantic reasonability in query information retrieval thanks to using semantic technology for IAQ assessment. A relevant advancement for considering the occupant profile was made by Adeleke and Moodley [25]. The ontology was applied in a real-world case study in Durban, South Africa, and focused on IAQ in homes where there are pregnant mothers and infants. Pregnant individuals, infants, unborn babies, and children in general are considered the most vulnerable population to negative health effects related to poor air quality [26]. More recent research by Chung et al. [27] advances the IAQ semantics by combining several sets of knowledge: indoor environmental conditions, outdoor environmental conditions, and occupant profiles. Occupant profiles were also covered by Tzouvaras’s [28] research with an extra focus on IAQ and energy efficiency. Spoladore et al. [29] described a decision support system that leverages ontological representation of individuals and their health condition—relying on a World Health Organization standard, the International Classification of Functioning, Disability and Health (ICF) [30]. An ontology designed to serve as a prototypical smart home toolkit, integrating various technological solutions to enhance inhabitants’ comfort and safety, was later developed by the same researcher for the Domestic Ontology Managed Ubiquitous System [31]. None of the previous research covers IAQ, related health, and energy management in unified ontological systems, addressing the minimum necessary classes for basic data computing and processing. There is an immediate need for a standard ontological structure capturing these various aspects [21].

A comprehensive ontology encompassing IAQ, building energy performance, and dwellers’ health is still lacking. Specific ontologies that fully address energy performance, IAQ, and related health impacts do not exist. Over the last decade, several fragmented attempts to address aspects of IAQ, building energy performance, and dwellers’ health have been documented, as mentioned above. Therefore, we present a prototype built on previous fragmented research to fill this gap.

3. Methods

In the context of this study, an ontology is defined as a rigorously formalized and explicit representation detailing the conceptual framework within the IAQ, energy performance, and health domain. We apply the well-documented ontology engineering method “Methontology” [32] for the prototype creation. The method is displayed step-by-step in Figure 1.

The knowledge acquisition phase was carried out by desktop research and via numerous EDIAQI project meetings spanning from March 2023 until January 2024. The information collection process involved experts with diverse experience. Extensive negotiations took place during the summer of 2023. After each meeting, the framework was reviewed, leading to either consensus and implementation of changes or the scheduling of additional meetings for further discussion.

The evaluation phase started with IAQ ontologies domain analysis. Scattered existing classifications were identified. Existing classifications and frameworks were presented in Section 2. The conceptualization of a novel framework was presented at the Air Protection 2023 conference, where a hypothesis of a minimum viable product covering three aspects: (1) indoor air quality, (2) energy performance, and (3) health condition(s) was presented. The conceptualization framework includes well-defined classes from previous research, which embody fundamental entities. Furthermore, the properties associated with each class serve to articulate a comprehensive set of features and attributes pertaining to the concept, denoted as slots or roles. Slots describe properties. The ontology also incorporates restrictions on these slots, termed facets or role restrictions, which delineate the permissible characteristics of the associated properties. It is essential to recognize that an ontology, when coupled with a collection of individual instances corresponding to the defined classes, collectively forms a knowledge base. Despite the practical challenge of precisely distinguishing the boundary between ontology and knowledge base, an ontology essentially functions as a conceptual model representing the intricacies of the real world and its components [32]. Consequently, the concepts captured within the ontology must accurately mirror the components of reality they represent [33,34].

The implementation phase, with a goal of capturing reality, is carried out in the Tallinn pilot and Vilnius campaign within the EDIAQI project until 2026. Within the implementation phase, ontologies, which are explicit, shared, and descriptive logic-based conceptualizations of knowledge and relationships within our domain, are utilized to model IAQ, energy performance, and health-related concepts. This modeling is conducted using W3C-endorsed languages such as the Resource Description Framework (RDF), Ontology Web Language (OWL), and Semantic Web Rule Language (SWRL). An interesting aspect of employing ontologies is the capacity to infer new facts that are not explicitly stated in the ontology through reasoning. This capability facilitates the discovery of additional knowledge, thereby enhancing the value proposition for various knowledge-based enterprises and domains.

In practical terms, the implementation phase of an ontology includes the following:

• Defining classes in the ontology;
• Arranging the classes in a taxonomic (subclass–superclass) hierarchy;
• Defining slots and describing allowed properties for these slots;
• Filling in the properties/values for slots, for instance.

It is then possible to create a knowledge base by defining individual instances of these classes and filling in specific slot value information and additional slot restrictions [31]. For example, for the EDIAQI project, the research goal was to arrange an ontology for indoor environments that manage indoor air quality, energy performance, and dweller health pre-conditions. We begin by defining the most important classes for indoor air quality data management and interoperability (point 1 above). These classes should be arranged hierarchically to reflect their relationships (point 2 above). We can define key classes as follows:

• Thing: The root class from which all other classes inherit;
• Device: A subclass of Thing, representing any physical device in the smart home;

  ○

  Sensor: A subclass of Device, representing sensors that collect data.

  ▪

  AirQualitySensor: Monitors indoor air quality (e.g., pollution levels, humidity);

  ▪

  EnergyConsumptionSensor: Tracks energy usage of various devices.

• Room: Represents different rooms in the home;
• Dweller: Represents inhabitants of the home, including health pre-conditions.

  ○

  HealthCondition: Subclass for different health conditions (e.g., asthma, allergies).

It is then possible to answer our research question by creating a hierarchy of the classes and giving slots values (properties). SWRL rules can be used to infer new knowledge based on the data collected by the sensors. For instance, a rule should be implemented to adjust ventilation based on air quality and dweller health conditions. As various relevant ontologies already exist, all possible previous work should be merged. Namely, the World Wide Web Consortium (W3C) description of date and time structured with separate values for the various elements of a calendar-clock system is applied. SAREF, the Smart Applications REFerence ontology, is incorporated. The SAREF ontology focuses on the concept of a device, which is defined as “a tangible object designed to accomplish a particular task in households, common public buildings or offices. In order to accomplish this task, the device performs one or more functions”. Examples of devices are a ventilation switch, a carbon dioxide (CO₂) sensor, an energy meter, and a washing machine. Following the work of Spoladore et al. [30], the ontological representation of individuals and their health condition—relying on a World Health Organization standard, the International Classification of Functioning, Disability and Health (ICF)—is implemented in the EDIAQI ontological structure.

Finally, for the evaluation’s concluding step, the ontology adaption is an ongoing process, accommodating new data interoperability needs as these emerge as the project evolves. If the scientists working on the EDIAQI project scientific pilots and campaigns who will maintain the ontology will describe the domain in a language that is different from the language of the ontology users, it is necessary to provide the mapping between the languages.

Documentation has been carried out and will continue until 2026 via the project’s official delivery reports and the EDIAQI GitHub account.

4. Results

The main contribution of the paper consists of an ontology and the methodological description of its design process that is provided for replicability for the EDIAQI project pilots and campaigns and for scientific purposes beyond the project lifetime. It provides a common vocabulary and structured representation of building energy performance, IAQ, and dweller health conditions, enabling a shared and standardized understanding of the information collected and its relationships with the data of the IoT system. This promotes more effective aggregation and integration of the energy performance data and IAQ management in the decision-making process of the IoT system. In the knowledge acquisition phase, essential classes to be defined with their properties to cover the three above-mentioned domains were agreed upon in the summer of 2023 in a 2-day workshop at Thinnect offices in Tallinn. Namely, as follows:

• Contextual and metadata classes, a minimum of two subclasses with time and location properties;
• Data and communication classes, a minimum of three subclasses for measurement recorded with timestamp, value, and sensor ID; data exchange format type as JSON, XML; and the communication protocol used for data transmission as MQTT, HTTP;
• Control and management classes, a minimum of two subclasses for the user role and the strategy for controlling building systems with properties covering user preferences, management strategy type, parameters, and objectives.

Well-documented existing ontologies, like SAREF, W3C description of date and time, and ICF, were incorporated into the ontological structure, and the utilization of previous work is explained in detail in the next chapters. We present the findings in more detail in three different categories: (1) indoor air quality, (2) energy performance, and (3) health. We present the culmination of the merged work in the EDIAQI ontology prototype sub-chapter.

4.1. Indoor Air Quality

The inaugural iteration of the ontology developed for the EDIAQI research project draws upon the foundational framework of Thinnect initial IAQ sensor technology, which initially focused on temperature, humidity, particulate matter (PM2.5, PM10), and CO₂ measurement. For capturing the IAQ data, a minimum of two classes are necessary. The environmental performance class represents the quality of air inside the building, incorporating subclasses of indoor air quality with various properties, such as CO₂ levels, PM2.5 levels, PM10 levels, humidity, and temperature. A second class within this domain for sensor devices is necessary, covering sensors used for monitoring indoor air quality and energy performance with its property types (e.g., TemperatureSensor, HumiditySensor, CO₂Sensor, ParticulateMatterSensor).

Measurements acquired by a sensor are semantically annotated, providing for each a unique ID, the value of the performed measurement, a DateTime stamp of the measurement, the unit of measurement involved, and the comfort metrics it is referred to as described in the last section. Each measurement acquired by a sensor is then classified thanks to a set of SWRL rules. The characterization of devices relies on the utilization of SAREF as developed in Spoladore et al. [35], which serves to delineate certain attributes of a device and complement the provided device location within the space as established by DogOnt. Commencing with the foundational concept of “Device”, this reference ontology furnishes the framework for describing a device’s properties, spatial positioning, and classification. Additionally, SAREF encapsulates the operational functionality of a device, including its role as a sensor and the parameters it measures. For instance, a PM sensor (classified under the saref:Sensor category) performs a saref:SensingFunction. Designed specifically for the characterization of indoor environment devices, SAREF is constructed using RDF/OWL and has been employed in various studies focusing on device interoperability [30]. Within the context of pilots in the EDIAQI project spanning from 2023 until 2026 aiming at developing novel solutions for indoor climate management with 19 partners, SAREF was adopted to delineate the devices deployed within the indoor spaces, specifying their respective types (sensor, actuator, appliance, etc.) and unique identifiers (ID).

4.2. Building and Energy Performance

The European Energy Performance of Buildings Directive adopted in 2023 [35] mandates that while the indoor climate parameters should be maintained in the optimal range, the building energy performance parameters should also be optimized. Thus, building characteristics classes are also presented in this section. A comprehensive ontology is required that enables the handling of all of the building and energy-related data. As each building is unique, four class categories for mapping energy performance were defined for the prototype design. Firstly, building and space classes, incorporating properties for the entire building structure and type (residential, commercial, etc.); floor number(s) or area(s); room type (office, kitchen, etc.), area, volume and zone ID, area, controlled temperature/IAQ parameters need to be arranged. The location of the building and its characteristics have a direct effect on energy performance (sunny area, cold area, etc.), and therefore, the above-mentioned classes and their properties were identified in the conceptualization phase as essential for the ontology design. Secondly, the appliances, systems, and equipment usage have a direct effect on energy performance. Essential classes for capturing device energy usage include the following:

• Actuator—devices that can control environmental conditions (e.g., thermostats, smart windows, or balcony doors) with their property types (actuator type, control range, location);
• Sensor—various sensors used for monitoring energy performance with its property types (e.g., EnergyMeter, CurrentSensor, VoltageSensor);
• HVAC system—the heating, ventilation, and air-conditioning systems with their properties (system type, capacity, efficiency).

As the performance, condition, and operation of the HVAC systems have a direct effect on the energy consumption of the building, rules for turning devices on and off automatically (threshold values) must be handled, collected, communicated, and processed in tandem for advanced analytics and control. As described above for sensor devices, the characterization of all devices within our work relies on the utilization of the SAREF. Therefore, each actuator, energy consumption sensor, and HVAC system is assigned a unique ID. With the use of SWRL rules, it is possible to describe the conditions under which a specific action is activated. For instance, to set the proper air-conditioning system’s program during periods of warm weather when heating is not required, the following conditions must be met:

->provide IndoorClimateSettings (?clcs) EnvironmentalMeasurement (?m), NotAccepta-

bleSummerDomesticTemperature (?m), hasMeasur- ementValue (?m, ?value), greaterThanOrEqual

(?value, 27), EnvironmentalMeasurement(?n), NotAcceptableDomesticHumidityRate (?n), has-

MeasurementValue (?n, ?value2), great- erThanOrEqual (?value2, 60), Appliance

(air_conditioning_system), AirConditioning- SystemProgram (CoolMode) -> hasProgram

(air_conditioning_system, CoolMode)saref:SensingFunction.

As new building types emerge in pilots, additions to the code may have to be made to capture air indoor–outdoor airflow elements. Data interoperability is crucial for managing air flows through windows, balconies, doors, and ventilators. It ensures seamless communication between various devices and systems, facilitating efficient air quality control and energy management. Interoperability enables data from different sensors and actuators to be integrated, analyzed, and acted upon in real time, optimizing ventilation and saving energy. This is essential for buildings where coordinated control of air flows can significantly enhance comfort, reduce energy consumption, and improve indoor air quality simultaneously.

4.3. Health

The elucidation of dwellers’ health status is underpinned by the utilization of a recognized standard endorsed by the WHO, namely the International Classification of Functioning, Disability and Health [30]. Class for user was created for our project purposes, with subclasses of roles (e.g., facility manager, dweller) and type (e.g., has a health condition or not). Our proposed class for user integrates the WHO ontology for diseases and disabilities (ICF). This classification framework conceptualizes an individual’s functioning as an interplay between their health condition and the surrounding environment. The ICF framework comprises two primary components: “Functioning and Disability”, which delineates the constituents of body functions, body structures, and activities and participation, and “Contextual Factors”, which enables the depiction of the influence of environmental factors and personal factors. Each component within the ICF framework can be further defined into specific chapters, which represent distinct health-related domains. Through a progressive specification of these chapters, the ICF facilitates the comprehensive assessment of an individual’s functioning and disability domains. Every slot within the ICF framework is denoted by a letter (e.g., b for body functions, s for body structures, e for environmental factors, d for activities and participation) and can be further elaborated by appending digits, as presented in Figure 2.

As with every prototype, this framework cannot address all the physiological and poor IAQ or health-related problems, nor can it provide a permanent solution; instead, the classification can make the dwellers feel more comfortable and possibly more productive in indoor environments. Similarly, the proposed system set-up can provide personalized IAQ thresholds related to persons with respiratory issues, which can be the cause of poor IAQ (described with ICF range of codes related to the functions of the respiratory system b440–b449). In these cases, a dweller can upload (or ask his/her building owner to upload) his/her physiological status (formalized in an RDF/OWL ontology) for his/her home or office space and a set of SWRL rules to adapt IAQ metrics of the customized settings by increasing or decreasing the specific relevant IAQ parameter levels indoors. The SWRL acquires the following form:

Dweller_with_mild_b445_imp(?g), selectsActivity(?g, ?a),Activity (sleep), CustomizedCO2Setting (?clcs)->provideComfortSetting (?clcs)

4.4. The EDIAQI Ontology Prototype

The ontology is based on cutting-edge, rigorously validated ontologies such as those employed in the Brick Schema, Project Haystack, iZeb, and CasAware [31,36] projects, among others, delineated in the comprehensive Section 2. This amalgamation culminates in a pioneering ontology designed for the holistic monitoring of IAQ, health parameters, and energy planning. The structural framework of this novel ontology is outlined (representation of the ontology in Protégé) in Figure 3.

To answer the research question, a minimum viable product needs to cover contextual and metadata classes, data and communication classes, control and management classes, and energy performance, IAQ, and health condition classes to capture energy performance, IAQ, and health data. By applying the step-by-step design approach described in this study, a prototype of 78 essential classes was identified. Furthermore, the prototype included 25 object properties, 21 data properties, six annotation properties, 68 individuals, and nine various data types. The “Building Data” class encompasses properties such as year of construction and type (residential, commercial). A subclass of “Floor” specifies floor number and area, while “Room/Space” defines attributes such as room type (office, kitchen), area, and volume. A second subclass of “Zone” represents HVAC areas with properties like zone ID, area, and controlled temperature. “Equipment and Systems” classes include “HVAC System” with properties of system type, capacity, and efficiency, and “Sensor” subclasses (e.g., EnergySensor, TemperatureSensor) with properties like sensor type, accuracy, and location. “Actuator” manages devices (e.g., thermostats) with properties such as actuator type, control range, and location. Environmental and Performance classes cover “IndoorAirQuality” with properties such as CO₂ levels, PM levels, temperature, humidity, etc., and “EnergyPerformance” details metrics like energy consumption and efficiency rating. Data classes like “Measurement” capture sensor data (timestamp, value, sensor ID), “DataInterchangeFormat” specifies data exchange formats (JSON, XML), and “Protocol” defines communication protocols (MQTT, HTTP). “Control and Management” classes include “ControlStrategy” for system strategies (type, parameters, and objectives) and “User” for user roles and health preferences. Contextual and Metadata classes encompass “Location” with properties latitude, longitude, and altitude, and “Time” specifies timestamp, duration, and time zone. These are the essential classes needed to be covered for a framework covering IAQ, energy performance, and health aspects within our study. Additional classes are needed to tailor the framework for each research endeavor and goal. In Figure 4 is a model of the version 2 prototype applied in the Tallinn pilot in 2023, which incorporates all essential classes mentioned above covering ICF for health, IAQ, and energy performance domains.

The study also recommends leveraging existing ontologies and standards for the effective implementation of these classes. This prototype is documented in the project deliverables, available to the public, and in the project GitHub account. It serves as a toolkit to set snippets of code for individual pilots and/or campaign needs within the EDIAQI project and beyond.

5. Discussion

This research proposed an ontology prototype for residential buildings, schools, universities, kindergartens, and office buildings. Our proposed ontology is intended for implementation in private homes, universities, schools, daycare centers, gyms, laboratories, and office environments. The prototype, based on previous research in the IAQ ontologies domain, integrates classes with properties to define building dynamics (Building data), classes for roles and profiles for health conditions (Users), and a framework for data communication and control and management classes (Advanced), contextual and metadata (Data) offering a comprehensive approach to IAQ and energy management. A front-end representation of the prototype, with a traffic light color code system, based on Thinnect technology, is displayed in Figure 5.

The use of ontologies, such as the WHO’s International Classification of Functioning, Disability and Health, enables customized environmental settings for “users” that cater to individual health conditions, demonstrating the potential of ontology-driven solutions in enhancing indoor air quality and personalized occupant well-being. In these cases, a dweller can upload (or request the building owner to upload) their physiological status (formatted in an RDF/OWL ontology) and use SWRL rules to the preferred indoor environment settings, as displayed in Figure 6.

Evaluation of the prototype is an ongoing progress until 2026 and beyond. After an initial version of the relevant reviews of IAQ management ontologies was defined and presented at the Air Protection 2023 conference, a step to evaluate and debug it by using it in applications or problem-solving methods and by discussing it with partners in the pilots was taken after the conference. As a result, it was necessary to revise the initial ontology. The SOSA light-weight ontology [37] was merged with the preliminary version to capture the interaction between the entities involved in the acts of observation, actuation, and sampling. This evaluation process of iterative design will continue throughout the entire lifecycle of the ontology throughout the years as the results from pilots and scientific working packages emerge.

Future research and evaluations should focus on refining these ontological frameworks and expanding their applicability across different environments and domains. The main limitation of our work is the lack of sufficient personal IAQ health risk data. If research in toxicology and IAQ advances and extensive IAQ measurement campaigns controlling for lifestyle factors (e.g., diet, sleep, activity) emerge, we suggest modeling the health framework accordingly. Insights into indoor air pollution exposure would help identify personalized health risks.

Evaluating the effectiveness of these systems in real-world settings will provide valuable insights for further development and optimization. The user study will be published in a complementary manuscript once the pilots have successfully finished. The ongoing evolution of ontology-based IAQ solutions holds promise for advancing sustainability in public health and creating healthier indoor environments.

6. Conclusions

The significant contributions of an ontology designed for the EDIAQI project pilots and campaigns are aimed at enhancing building energy performance, indoor air quality (IAQ), and dweller health conditions within one framework, thus filling a research gap. This ontology promotes sustainability by enabling more efficient and integrated monitoring solutions, which is essential for green building initiatives. For this purpose, a novel ontology prototype was suggested based on the state of the art and the project requirements. The ontology prototype consists of 78 classes, 25 object properties, 21 data properties, six annotation properties, 68 individuals, and nine various data types that are all subcategories of a design approach starting with defining contextual and metadata classes, data and communication classes, control and management classes followed up by arranging classes for energy performance, IAQ and health condition data. The ontology prototype is saved in the EDIAQI project GitHub account, which is available for the general public to provide a standardized framework for representing building energy performance, IAQ, and health data. This standardized understanding facilitates effective data aggregation and integration, which is crucial for optimizing energy usage and improving air quality.

Essential classes were defined by applying the well-established ontology engineering method “Methontology” to cover contextual metadata, data communication, and control management. These classes include properties for time, location, sensor IDs, data exchange formats (JSON, XML), communication protocols (MQTT, HTTP), user roles, and management strategies. Such a structured approach ensures comprehensive data handling and interoperability suggested beyond the EDIAQI project. The ontology includes classes for environmental performance and sensor devices, covering various properties like CO₂ levels, PM2.5, PM10, humidity, and temperature. This detailed categorization aids in the precise monitoring and management of IAQ, contributing to healthier indoor environments. Classes for building characteristics, appliances, systems, and equipment usage are defined to optimize energy performance. These include actuators (e.g., thermostats, smart windows), sensors (e.g., energy meters), and HVAC systems. Rules for automated control of these devices enhance energy efficiency, aligning with sustainability goals. As the future holds promise for personalized smart systems, the ontology incorporates the WHO’s International Classification of Functioning, Disability and Health to tailor environmental settings to individual health conditions. This customization improves occupant well-being and demonstrates the potential of ontology-driven solutions in creating healthier indoor environments.

The ontology is designed for diverse settings, including residential buildings, schools, universities, kindergartens, gyms, laboratories, and office environments. The integration of health conditions, comfort metrics, and environmental data offers a comprehensive approach to IAQ and energy management. The ongoing evaluation and iterative design process ensure that the ontology remains effective and relevant. Future research should focus on refining these frameworks and expanding their applicability. Evaluating their effectiveness in real-world settings will provide valuable insights for further development, ultimately advancing public health and promoting sustainable building practices.