(Back to the menu of AmI projects)
Universal Ontology for Smart Environments (UnOvi)
Develop an Universal Ontology for Smart Environments.
The UnOvi ontology is developed to overcome some issues related to smart environments.
- Describe smart environments.
- Describe domains as well as applications.
- Overcome the limitations of the existing ontologies.
- Enrich our ontology by introducing new concepts like the referentiality and environmental change.
The existing ontologies focus mainly on very specific domain aspects. For example, in the domain of smart environments, there exist ontologies which treat only the context modeling or QoS related service properties, other ontologies focus on the smart meeting space modeling, and others emphasize the human home interaction. Nevertheless, numerous problems remain unresolved and have not been addressed by the existing ontologies.
The design of UnOvi ontology is based on the general idea which states that, a being lives and interacts in an environment with a certain dynamic. From here, we can extract the major concepts of our top level ontology, which are: Being, Environment and Dynamic.
The environment modeling is achieved in many ontologies found in the literature, but in our ontology, this concept is differently formalized to overcome the limitations of the existing modeling works. The concept of Environment refers to the environment in which a Being evolves. This environment includes all what is necessary for the evolution of a Being, in terms of habitation, food, travel, leisure, care, etc. In our ontology, the concept of Environment is subdivided into two main sub-concepts. The first sub-concept is Tangible, which refers to all things that can be able to be touched or capable of being given a physical existence, such as utensils, ingredients, devices, appliances, furniture, etc. and the second sub-concept is Intangible, which refers to something that cannot have a physical existence like geometric shapes and computational entities. For example, a computer memory has a storage capacity which is intangible, but the memory as an electronic component is tangible. Figure 5 shows the hierarchy of the Environment concept and its related sub-concepts.
Location-awareness is highly relevant subject in ubiquitous environments, as many applications exploit location information to provide adequate services or adapt to a changing physical environment. To deal with the localization problem in smart environments, and to overcome the limitations of the existing ontologies, we define two types of localization in smart environments as described below.
The Being plays a vital role in smart environments. In UnOvi ontology, the concept of Being is divided into two important components. A Physical component that reflects the physical existence of the Being. This component encompassing Animal, Plant, Insect and Person. The second component focuses mainly on the profile of the Being that allows to efficiently characterize the Being’s state, in terms of behavior, preferences, disabilities and so on. The profile which is a NonPhysical component, that according to the international classification of functioning, disability and health is composed of three main categories: personal profile, health profile and organic profile. These categories are mainly applied to characterize human being. However, they are implemented to be applied for other beings such as animals and plants.
The concept of Dynamic refers to all kinds of activities that might exist in an Environment. In UnOvi ontology, the concept of Dynamic is not limited only to activities related to a Being like daily living activities (preparing, eating, cleaning, etc.) or entertainment, but rather, it may expand to include other kinds of activities such as mechanical activities, and computational activities provided by computational entities disseminated in the environment, in order to help beings and ensure their well being. Therefore, the concept of Dynamic can be broadly classified into two sub-concepts such as Virtual, which refers to all computational activities, and Non_Virtual which refers to all activities that can be achieved by a human being or a machine such as a robot.
The notion of referentiality is an important notion that should be taken into account during the development of smart environments ontologies. In our UnOvi, the referentiality encompasses two main categories, referential localization and referential time.
Several researchs use the location as a unique concept (absolute localization). However, a significant weakness of this location remains in the imprecision, insufficiency and ambiguity. For example in a home, the fridge is located in the kitchen, and the oven is located in the kitchen, but there is no a clear idea about the physical disposition of both the fridge and the oven. None have taken advantage of the information that describes what surrounds a particular object and its relationship to other objects in the environment. Thus, unlike the absolute localization, the referential localization takes into account the localization of the object in relation to the other objects in the environment. This localization takes the form of (upTo, downTo, behindTo, inFrontTo, nearTo, inTo, Xpositive, Xnegative). The Xpositive and Xnegative are used respectively instead of the right and left side of the object according the angle with which the object is perceived. This kind of localization allows us to know the physical disposition and arrangement of objects in the environment, and how they are interlinked between them. By knowing the physical arrangement of objects in the environment and how they are interlinked constitute additional information of great importance, which can be used thereafter among others for detecting and recognizing activities, security aspects in the environment, detecting and avoiding obstacles for a robot, etc.
In addition to the referential localization, we define the absolute location. Each object in the smart environment has an absolute location, which represents the object’s coordinates in the space, taking into account the rotations upon the three axis, yaw, pitch and roll in order to detect thereafter the volume occupied by the object. Therefore, the absolute localization of each object in the smart environment is defined as follows (x, y, z, yaw, pitch, roll).
The notion of referential time is inspired from the work of James F. Allen about his theory of action and time. The temporal relations defined in his theory are applied for action. These temporal relations are integrated in our ontology to describe temporal relations among activities. The referential time takes into account the time of an activity in relation to other activities either scheduled or deduced. It allows us to ensure that activities are realized in the good order which increases considerably the monitoring process. The referential time takes the form of (before, after, parallel to, since, from, Overlaps). Moreover, each activity has an absolute time that refers to the activity start time, and a duration that refers to the activity realization time.
- Abdulrazak, Bessam.
|Chikhaoui, Belkacem.||Meta-antologie for for Smart Environments.||Summer
|Gouin-Vallerand, Charles||Dynamic Contextual Assistance for Ambient Intelligent Environments.||Spring
|Roy, Patrice.||Context-Awareness for Open Smart Spaces.||Spring
|Benazzouz, Yazid.||Automatic Learning of contextual adaptation of services.
PhD at the École des Mines de Saint-Étienne, France.
Undergraduate Students: Bachelor projects at the UdeS
|Brunelle, Michael.||Context description for intelligent environments.||Fall
|Lhommet, Margaux.||Description and Taxonomy of Smart Spaces.||Summer