Ontology-based Aware for Ubiquitous Home Care for Elderly People Kurnianingsih 1, 2, Lukito Edi Nugroho 1, Widyawan 1, Lutfan Lazuardi 3, Khamla Non-alinsavath 1 1 Dept. of Electrical Engineering and Information Technology, 3 Dept. of Public Health Universitas Gadjah Mada Yogyakarta, Indonesia 2 Dept. of Electrical Engineering Politeknik Negeri Semarang Semarang, Indonesia kurnia.s3te13@mail.ugm.ac.id, {lukito, widyawan, lutfan.lazuardi}@ugm.ac.id, khamla@mail.ugm.ac.id Abstract The interaction between elderly s body states and their surrounding environment is important in many conditions. Biological data and activity data in elderly s daily life using wearable body sensor can be monitored and then can be adjusted with environment to gain the comfort condition. Seamless integration home care which is embedded with pervasive devices and awaring of the context can help residents to achieve convenience and finally improve quality of life. The objective of this paper is to analyse and characterize elderly and smart home environment contexts that affect elderly s convenience. Observation was conducted to elderly community, nursing home, and geriatry to get service activities for elderly. Ontology was used to formulate the personalization process of elderly and decision support tool for doctors and wellbeing experts. The result of this work is presented by developing an ontology-based context aware framework for ubiquitous home care for elderly people. Keywords context aware; ubiquitous home care; ontology I. INTRODUCTION Elderly people interact daily with their surrounding environment in numerous ways. They perceive the environmental conditions, react, or adjust it. If the surrounding environment at home can respond to the behaviour of elderly, it can be a smart home environment, which can give many comforts to elderly. Smart home is an environment, which is equipped with ambient intelligence and context aware, so that it can respond to the behavior of residents, provide them with various functions, and can give comfort to residents [1]. Smart home automatically or autonomously provide services to make the residential life more comfortable, secure, and more economical. Elderly who likes living autonomously is suitable with smart home because smart homes contain assisted services. Assisted service will support elderly activities in their daily life. Seamless integration home care which is embedded with pervasive devices and awaring of the context can help residents to achieve convenience and finally improve quality of life. describes heteregeneous aspects of physical characteristics, including activities and the relationships among users in a dynamic and complexity environment related to location, time, device, service, personal information, and environment. To determine what context is appropriate for the system and what context-aware services to provide in their system need an understanding of the context and how the context can be used. Dey [2,3] describes context as any information to characterize the situation of an entity comprising object, place, or person related to the interaction between user and application, including the user s emotional state, focus of attention, location and orientation, date and time in the user s environment. aware computing is a notion of having computer system being aware of the context. Being aware of the context means that system can recognize and perceive the real world using sensors, and also can act or stimulate with the real world. Giving the computers eyes and ears to act and interact more appropriately is the heart of context aware to make the use of technology easier. There are two types of context-aware computing [4], namely: Using context means to be able to detect and sense, interpret, and respond to user s environment and to provide flexibility of services. Adapting to context means to be able to change dynamically based on the user s present context. -aware ubiquitous home care is characterized by awaring technology and computation of elderly intentions, activities, preferences, and can autonomously adapts its behaviour on context changes anywhere and anytime at home. Over the past few years, the number of researchers have developed context aware system in ubiquitous home care to perform the usefulness, flexibility, and personalized service for home care residents. The design and implementation of context-aware middleware for providing home care services for elderly, which has ability to operate over multiple physical spaces, has been developed by Pung et al. [4]. Elderly can interact anytime and anywhere with care givers through 978-1-4799-9863-0/15/$31.00 c 2015 IEEE 454
Internet. This pervasive care of elderly consists of wireless sensor networks with ubiquitous computing in smart environment to provide elderly monitoring, reminding of selected activities, and assisting elderly in performing their daily activities. Another research such as the design of an agent-based component for inferring elderly s context and predicting whether they need help in performing their daily activities using ambient intelligent system has been developed by Jimenez et al. [5]. System can change the contextual condition dynamically according to different scenario in taking medication. The specific purpose of this paper is to analyse elderly and smart home environment contexts that affect elderly s convenience. This objective concerns on analysing elderly and environment contexts that affect elderly convenience. Translating elderly context, elderly s needs and preferences into environment behaviours is investigated to achieve this goal. The structure of this paper is as follows. Section 2 describes about context aware framework of ReSCUE. Section 3 presents ontology design of ReSCUE. Section 4 describes about conclusion. II. CONTEXT AWARE FRAMEWORK OF RESCUE Observation is the first phase to obtain more reliable information about people (elderly, doctor, care giver, wellbeing experts, and family member) and environment (rooms and facilities). Three activities of observation consist of screening process, location orientation, and service process in order to determine activities. The result of observation will be recorded, analysed, and interpreted in the model of context in order to determine activities. Modeling context is a process that imitate the user cognition behaviours and user s surrounding environment which is full of uncertainties. information can be acquired from sensors and other resources, which may be imprecise, unknown and fuzzy. information is all information comes in from real world about what is used to change. Fig. 1 depicts a contextaware of reflective sensing and conditioning system for ubiquitous home care for elderly people (ReSCUE). Characteristics of context information of ReSCUE consists of: Human characteristic which describes information about the surrounding s user consists of feedback, preference, medical reference, personal information, additional information, position, medical record. Environment characteristic which describe a sensor driven computing that collects the physical environmental data, namely environment temperature, room temperature, humidity, and lighting. Learning characteristic which describes learning information consists of history, date time, machine parameter. Wearable characteristic which describes information comes from body wearable sensing using sensor, namely body temperature sensor and heart rate sensor. Fig. 1 depicts context aware framework of ReSCUE which is divided into three sections, namely context information, context interpreter, and context aware. information is any kind of information which comes from multiple sources. The system builds the context model based on collected context information. The majority of information is dynamic. The information captured by different sensors and other resources is received by the system and stored as a knowledge base. Knowledge base is a knowledge repository for context that provides a means for information to be collected, organized, shared, searched and utilized. Some context information derived from body temperature sensor and heart rate sensor can be sent directly to context interpreter as a reflective sensing to be initialized and screening. The result of this context classification can be sent to appropriate users as an alert in emergency condition and can be sent to context reasoning to proceed with data from knowledge base as a conditioning system process. The context aware data fusion integrates data from multiple sources to build the overal context. The context is delivered with reference to decision scenario and is used for decision making. reasoning will manage large amount of context data in order to derive context from the existing one at semantic level. The context reasoning engine is composed of reasoner and adaptation inference rule. Reasoner allows to infer new situations from relevant contextual properties based on defined semantic relations and inference rules. Adaptation inference rule contains a set of situations. -aware is about the awareness of dynamic environment, the relationships among entities, and the relevance of parameters in decision space. -aware service of ReSCUE performs some services namely reflective service, basic service, and customized service. a. Reflective Service Reflective service is a service performed reflectively as a required action after capturing sensing data. Data sensing of elderly s body temperature and heart rate are captured using wearable sensor network to be processed in reasoning context to give alert action reflectively needed by elderly. b. Basic Service Basic service is a minimum service element that must be provided to elderly, such as notification, reporting, and environmental conditioning. Notification is a message, whether text or voice, sent to elderly notify some actions that have been done. Reporting is collecting data and drafting of the reports to any activities done. This reporting service will be sent to the doctors, wellbeing experts, care givers, and family members. Environmental conditioning is a set condition of elderly s surrounding environment based on preferences and medical references. 978-1-4799-9863-0/15/$31.00 c 2015 IEEE 455
CONTEXT INFORMATION Human Feedback Preference Medical Reference Personal Information Additional Information Position CONTEXT INTERPRETER CONTEXT AWARE SERVICE Medical Record Environment Environment Temperature Room Temperature Humidity Lighting Classification Knowledge base Reasoning Customized Service - Environmental Controlling Basic Service - Environmental Conditioning - Reporting - Notification Wearable Body Temperature Sensor Reflective Service - Alerting Heart Rate Sensor Learning Machine Parameter DateTime History Fig. 1. -aware framework of ReSCUE 978-1-4799-9863-0/15/$31.00 c 2015 IEEE 456
c. Customized Service Customized service is a service provides to personal specifications user's based on user's requirements and expectations regarding the services offered. Elderly may give feedback as response to a, unconvenience given service, both simple feedback and advanced feedback. The existing environmental services can be controlled based on request service, adjusted with their preferences and concerned to medical references. III. ONTOLOGY DESIGN OF RESCUE People, environment, and system have to communicate among themselves. Due to different needs and background contexts, a share understanding is needed to unify framework as a means to communicate and solve problems. Acquiring information from different context, performing context interpretation, and constructing context aware services are some tasks that need share understanding. A context aware model needs to be well established to provide support some tasks related to contexts. Web Ontology Language (OWL) is considered as a promising instrument to support various in contextual information modelling by defining the common upper ontology for context information in general, providing a set of low level ontologies which apply to different sub-domains, and reasoning about various contexts based on an inference engine to reason about various contexts in order to support decision making [6,7]. Ontology context is considered as a specific kind of knowledge and it can be modeled as ontology. Ontology opens a new challenge in dealing with contextaware services by representing semantics, concepts and relationships in the context data. It models the concepts of person, activity, time, and location, and describes the properties and relationships between these concepts. Several reasons for developing context model based on ontology: - Knowledge Sharing The use of context ontology enables computational entities in ubiquitous computing domain to have a share understanding of concepts about context between different systems and has capability of supporting semantic interoperability to exchange. - Logic Inference Ontology enables automated reasoning by exploiting various existing logic reasoning mechanism to deduce high-level, conceptual context from low-level, raw context, and checking and solving inconsistent context knowledge due to imperfect sensing. - Knowledge Reuse By reusing well-defined Web ontologies of different domains, we can compose large-scale context ontology without starting from scratch. The main advantage of context modeling using ontology is sharing common understanding of the structure of context information among users, devices, and services to enable semantic interoperability. The context ontology should be able to capture all the characteristics of context information. Ontologies are made up of two main components namely classes and relationship. Classes are represented by oval and relationships are represented by arrows. Each class contains multiple properties. Several relations subclass is used to link between different classes in ontology. Object properties and datatype properties are used to describe the characteristics and attributes of the concept and different types of restrictions for each property described by values or cardinality constraints. ReSCUE domain for elderly people specifies 9 (nine) contextual classes as shown in Fig. 2, namely Person, CommunicationMode, WellbeingParameter, SmartSystem, WearableDevice, Environment, MedicalReference, MedicalRecord, Preference. Fig. 2. OWL-based ontology for ReSCUE a. Person Class Person class defines five subclasses, namely: Elderly, Doctor, Wellbeing Expert, Care Giver, and Family Member. Each subclasses has data properties and object properties. Fig. 3 depicts object properties. 978-1-4799-9863-0/15/$31.00 c 2015 IEEE 457
Fig. 3. Object properties Subclass of Doctor has object properties between class and subclass, such as: analyzes MedicalRecord, gives MedicalReference. Subclass of WellbeingExpert has object properties between class and subclass, namely: determines WellbeingParameter. Subclass of Elderly has object properties between class and subclass, such as: wears WearableDevice, haspreference Preference, hasposition Room, hasmedicalreference MedicalReference, hasmedicalrecord MedicalRecord. b. CommunicationMode Class CommunicationModeClass defines three subclasses, namely: BluetoothCom, InfraRedCom, WiFiCom. Subclass of WiFiCom has object properties between class and subclass, namely: positionsearchingmethodused WearableDevice. c. WellbeingParameter Class WellbeingParameterClass has object properties between class and subclass, namely: hasbodytemperaturemeasurement BodyTemperatureSensor, hashistory History, haselderlypreference Preference, hasheartratemeasurement HeartRateSensor. d. SmartSystem Class SmartSystemClass defines four subclasses, namely: History, InputCollector, Interpreter, ServiceGenerator. Subclass of InputCollector has object properties between class and subclass, namely: hasinputvalue Environment, hasinputvalue HeartRateSensor, hasinputvalue Preference, hasinputvalue WellbeingParameter, hasinputvalue MedicalReference, hasinputvalue MedicalRecord, hasinputvalue BodyTemperatureSensor. Subclass of Interpreter has object properties between class and subclass, namely: hasreasoning ServiceGenerator. Subclass of ServiceGenerator has object properties between class and subclass, namely: controlenvironmentalconditioning LightingController, controlenvironmentalconditioning ACController, givesalerts Doctor, givesalerts FamilyMember, givesalerts CareGiver, givesnotification FamilyMember, hashistory History, performsreflectiveservice Light, performsreflectiveservice AC, reportsdecision WellbeingExpert, reportsdecision Doctor, reportsdecision CareGiver. Fig. 4. Ontology design of WearableDevice class e. WearableDevice Class WearableDeviceClass, as shown in Fig. 4, defines three subclasses, namely: BodyTemperatureSensor, HeartRateSensor, WearableDisplay. 978-1-4799-9863-0/15/$31.00 c 2015 IEEE 458
() Proc. of 2015 2 nd Int. Conference on Information Technology, Computer and Electrical Engineering (ICITACEE), Indonesia, Oct 16-18 th Subclass of BodyTemperatureSensor has object properties between class and subclass, namely: hashistory History, issensingbodytemperature Elderly. Subclass of HeartRateSensor has object properties between class and subclass, namely: hashistory History, issensingheartrate HeartRateSensor. SubclassWearableDisplay has object properties between class and subclass, namely: hascommunicationbluetooth BluetoothCom f. Environment Class Environment Class, as depicted in Fig. 5, has a subclass Room. Subclass Room defines three subclasses, namely: AC, Light, WallMountedGateway. Subclass WallMountedGateway defines four subclasses, namely: ACController, LightingController, WallMountedDisplay, WearableDeviceController. Fig. 5. Ontology design of Environment class Subclass of AC has object properties between class and subclass, namely: hascommunicationinfrared InfraRedCom. Subclass of Light has object properties, namely: hascommunicationinfrared InfraRedCom. Subclass of Light has data properties, namely: LightID, LightOnOffButton. Subclass of WallMountedGateway has object properties, namely: runs SmartSystem. Subclass of ACController has object properties, namely: controlsac AC. Subclass of LightingController has object properties, namely: controlslight Light. Subclass of Subclass of WallMountedDisplay has object properties, namely: hascommunicationbluetooth BluetoothCom, hascommunicationinfrared InfraRedCom. Subclass of WearableDeviceController has object properties, namely: controlswearabledevice WearableDevice. g. Medical Reference Class Medical Reference Class has object properties, namely: hashistory History. h. Preferences Class Preference Class has data properties, namely: HumidityPreference, LightingPreference, TemperaturePreference. Ontology-based context aware is used to support the activities in the life cycle of ReSCUE. First, there is a personalization process that adjust the contents of the ontology to the activities and health conditions observed in the home care of elderly, automatically providing a personalized ontology containing the learning process of each subsystem of ReSCUE. Secondly, the ontology is also used as the knowledge base of a decision support tool that helps doctors and wellbeing experts to detect anomalous circumstances such as exception based on experts, so that the learning system of ResCUE can be expanded to provide more accurate results. The ReSCUE ontology was implemented using Protégé [8], which is an open source ontology editor and knowledge-base framework, developed by Standford University to present the context modeling with OWL. IV. CONCLUSION The decreased of body metabolism at older ages, some medical problems, medications and the environment will affect the ability of elderly people to control and sense changes in its temperature. Monitoring of the health state s of elderly people at home will apply the concept of reflective sensing and conditioning system, which gather body temperature, heart rate data and surrounded environment data of elderly in real time to be processed and conditioned reflectively. Surrounding environment learned at this research as outputs to deliver convenience of elderly is focused on temperature, humidity and lighting. As the result of this paper, ontology-based context aware model, which was used to support the activities in the life cycle of ReSCUE, has been designed and the ReSCUE ontology was implemented using Protégé. Ontology provides a personalization process contains the learning process of each subsystem of ReSCUE. Ontology is also used as the knowledge base of a decision support tool for doctors and wellbeing experts. REFERENCES [1] A. Masaud-Wahaishi and H. Ghenniwa, Privacy-Based Multiagent Brokering Architecture for Ubiquitous Healthcare Systems, in Ubiquitous Health and Medical Informatics: The Ubiquity 2.0 Trend and Beyond, IGI Global, 2010, In-Press. [2] D. Raz, A. Juhola, J.S. Fernandez, and A. Galiz, Fast and Efficient -Aware Services, John Wiley & Sons, Ltd. 2006. [3] A. K. Dey, Understanding and using context, Journal of Personal and Ubiquitous Computing, Vol. 5, pp. 4-7, 2001. [4] H. K. Pung, T. Gu, W. Xue. P. P. Palmes, et al., -aware middleware for pervasive elderly homecare, IEEE Journal on Selected Areas In Communications, Vol. 27 (4), pp. 510-524, 2009. [5] D. S. Jimenez, M. D. Rodriguez, J. P. Garcia-Vazquez, and A. N. Espinoza, ELDer: An Ontology for Enabling Living independently of Risks, International Conference On the Move to Meaningful Internet Systems, Portugal, 2009. [6] D. Riano, F. Real, F. Campana, S. Ercolani, and R. Annicchiarico, An Ontology for the Care of the Elder at Home, Springer, pp. 235-239, 2009. [7] C. Calero, F. Ruiz, & M. Piattini, Ontologies for Software Engineering and Software Technology, Springer, 2006. [8] Protégé Ontology Editor and Knowledge Acquisition System, URL: http://protege.stanford.edu/ 978-1-4799-9863-0/15/$31.00 c 2015 IEEE 459