Energy Efficient Optimum Cluster Head Estimation for Body Area Networks

Abstract

Wireless sensor networks (WSNs) are becoming increasingly familiar since they are an inevitable part of human-centric applications. The WSNs are also used in health applications. It has small, low energy sensing devices that can sense the data as required and send it to the collecting base station. Since the sensor nodes have limited energy and in most of the applications sensors are not replaceable and rechargeable, energy conservation of sensor nodes is the primary goal over the design of WSNs. Although a lot of techniques are available for energy conservation, clustering is the important method used for preserving energy. Body area networks (BANs) are one of the specific applications of WSNs. This type of network is normally used to monitor the health of the human and functionalities of various organs of the human body. The sensors are generally implanted or positioned in the human body. Therefore, the BANs are helping the medical attendants to access the patient’s conditions regularly and keep track of the medical data of the person. The BANs consist of different type of sensors, they can sense patients pulse rate, sugar level, blood pressure, etc. Similarly, wearable devices are also having sensors to watch the behaviour of the organs of the body. These sensors sense the data and send them to the base station, where they will be analyzed and processed by the medical experts at any time. If any emergency occurs, the system will alarm the medical assistants 424and rapid actions will be taken. As the sensors are implanted, it is not possible to replace them. Since they are energy limited, the conservation of energy is very important in this type of applications. In this work, the sensors implanted and used in wearable are considered as reactive sensors. The reactive sensors trigger to send the sensed data when it is beyond the hard threshold value or the difference between the two consecutive sensed values is greater than the soft threshold value. These sensors are grouped under several clusters so that they can send the data to the cluster heads instead of the base stations to preserve energy. The cluster heads in turn send the received data to the base station after aggregation is performed. The triggered nodes or otherwise called as active nodes are the only busy nodes at the corresponding round and other sensors are kept under inoperative state. Hence, the active sensors are considered for the calculation of the optimum number of cluster heads. The optimum number of clusters plays a crucial role in deciding the clusters in each round of operation as they can also save the energy consumption of sensor nodes (cluster heads). The numbers of active nodes present in the earlier round and current round are determined and the change in the total network energy between two consecutive rounds is also computed. The ratio of active nodes among two consecutive rounds and the ratio of total network energy between two consecutive rounds are calculated. These values play a significant role in computing the suitable cluster heads for each round. After the cluster heads count is computed, the LEACH method is used to elect the cluster heads. If the elected cluster heads are below the optimum number, the balance cluster heads are selected among the remaining eligible nodes. If the cluster heads elected are greater than the optimum number, the cluster heads above the optimum numbers with least energy are converted as normal nodes. Simulations are performed and the results are compared with the existing protocols. The experimental results show that the proposed protocol outperforms the existing protocols in terms of network lifetime and throughput.