A Semantic Web-of-Things Architecture for Monitoring the Risk of Bedsores Rita Zgheib ∗ , R´ emi Bastide ∗ , Emmanuel Conchon † ∗ University of Toulouse, IRIT/ISIS, F-81100 Castres {rita.zgheib,remi.bastide}@irit.fr † University of Limoges, XLIM, F-87060 Limoges cedex emmanuel.conchon@xlim.fr Abstract—Bedsores are a common injury that mainly plagues elders and frail persons, and are a major cause of concerns in medical institutions. We present a system based on the Internet- of-Things technologies, aiming at detecting the risk of bedsores using sensor fusion. This paper mainly focuses on the software architecture of the proposed system, based on the principles of weak coupling and of semantic data exchange. We present a model of the application in terms of the Semantic Sensor Network (SSN) ontology. Index Terms—Keywords: Bedsore detection, Home care, In- ternet Of Things, Semantic middleware. I. I NTRODUCTION Elders, whether they are staying at home, hospitals or retirement homes, often incur the risk of health symptoms and problems. In many cases, some form of monitoring is helpful to help the healthcare personnel preventing the degradation of the patient’s health status. In [1], a study has been established on the trends in disease and injury incidence, prevalence, and years lived with disability (YLDs) which it is considered as an essential input into health policies. This study estimated these quantities for acute and chronic diseases and injuries for 188 countries between 1990 and 2013. Based on the authors interpretation, ageing of the world’s population is leading to a substantial increase in the numbers of individuals with disease after effects and injuries. The non-fatal dimensions of disease and injury will require more and more attention from health systems. Bedsores (also called pressure sores or pressure ulcers) [2], are one of the dangerous diseases that an elder can face. Bedsores are a localized injury resulting from prolonged pressure on the skin. They plague persons who have a reduced ability to move and change positions, and who stay in bed or wheelchair most of the time. Bedsores are dangerous and can have important consequences, leading to long-term hospitalization. At more severe stages, bedsores become very painful, the patient is at risk of surgery and even of death. Prevention techniques in hospitals and retirement homes today are still traditional, where the personnel spends a con- siderable amount of time regularly checking (usually every 15 minutes) the status of their patients and their changes in body position. The development of a pressure ulcer in a patient is considered a serious fault from the healthcare team. Many research projects like openIoT [3], OM2M [4] are fo- cused on smart home, smart cities and many IoT applications. These applications are based on intelligent sensors deployed on a smart space in order to gather real time information that can be treated and correlated to infer useful information. In the context of home care for dependent elderly people, and due to the risks of the bedsores that can damage the daily life of elderly, it is important to have an accurate bedsore detection system based on real time sensors deployed in the patient’s environment e.g. her bed or wheelchair. We present in this paper an innovative e-Health system dedicated to monitoring the risk of development of bedsores for elders staying in a hospital or retirement home. This paper mainly focuses oh the description of the software architecture of this system, which is based on a combination of Internet- of-Things and semantic technologies. In the proposed system, data is gathered from many ambient sensors. A semantic technique based on the use of ontologies is described in order to overcome the interoperability challenges introduced by the variety of sensors potentially used. Sensor’s data is handled by a knowledge-based, semantic middleware which routes this data to the appropriate decision modules. The paper is organized as follow: In Section 2 we describe the general approach of Braden Scale. The architecture with ontology and middleware processes are described in Section 3. We present then, OpenIoT project in Section 4. We discuss the related work in Section 5 before concluding in Section 6. II. BEDSORE DETECTION Several assessment scales [5] have been studied in the literature in order to quantify the risk of bedsores, among which are the Norton scale, the Waterlow scale and the Braden scale. The Braden scale is the most used method in clinical settings, since it results from a simple calculation based on 6 risk factors: sensitivity, mobility, nutrition, activity, moisture and friction. For each patient, a nurse creates a dashboard collecting the following information: • Sensitivity: to what extend the patient is able to respond to pressure-related discomfort; • Mobility: to what extend the patient is able to change and control her body position; • Nutrition: the patient’s usual diet, and it’s adequacy to her state of frailty; • Activity: the degree of the patient’s physical activity; 2015 International Conference on Computational Science and Computational Intelligence 978-1-4673-9795-7/15 $31.00 © 2015 IEEE DOI 10.1109/CSCI.2015.128 319 2015 International Conference on Computational Science and Computational Intelligence 978-1-4673-9795-7/15 $31.00 © 2015 IEEE DOI 10.1109/CSCI.2015.128 318