Energy Efficient Active Sensor Selection Scheme for U-healthcare in WBAN

Abstract

The number of people over the age of 65 has significantly increased in all countries over the last few decades. With this current trend in population, chronic diseases are becoming the world’s leading cause of death and disability for the elderly. Chronic diseases debase quality of life because chronic diseases require frequent hospital visits. Another major reason for injury and hospital admission among elderly is accidents such as falling that occur in the home. To address these social problems, one possible solution is a Ubiquitous Healthcare (U-Health) smart home. The U-Health smart home can support remote medical care that can act as a substitute for hospital visits from the elderly. The U-Health smart home can also provide continuous health monitoring, which can support early diagnosis and treatment to reduce health care costs. Today, Wireless Body Area Network (WBAN) is considered a key element to provide health care services anywhere, anytime in the U-Health smart home and will play an important role to enhance quality of life for elderly in the future. Therefore, research about WBAN is necessary before developing the U-Health smart home.Medical sensors that organize a WBAN should be as small as possible because they must not cause any activity restriction or behavior modification to the user. A small-sized sensor implies that its energy resource such as a battery should also be small. The amount of energy available in the wireless medical sensor is limited since it is proportionate to the size of battery. This restricted energy is the most important characteristic and most important problem of WBAN. Therefore, enhancing the WBAN’s lifetime through efficient energy consumption is necessary. In a WBAN, each medical sensor monitors different vital signs such as temperature, blood pressure, or ECG. Data collected by the medical sensors is transmitted to the coordinator. However, this data is not identical to the information needed to determine whether or not the elderly person has a specific disease. Only some of the data is necessary to make a decision. However, in existing systems, even though the data is not needed, medical sensors are always activated and continuously transmit data to the coordinator. This configuration causes high energy consumption in all medical sensors and reduces seriously their operational time. For efficient energy consumption in the WBAN, it is necessary to select only the most appropriate sensors that have the highest impact on the coordinator’s decision at a particular stage.Since, it is necessary to identify the set of symptoms to monitor at a particular stage and what is the importance of each symptom in the detection of a disease, it is important to find the relationship between symptoms and diseases, and we propose to use mutual information to define the information gain. Based on this information gain, we propose a way of selecting the most appropriate sensors to activate. We validate the proposed WBAN system by implementing a prototype system based on version 2.31 of NS-2.