Effective Deployment of Sensors in a Wireless Sensor Networks using Hebbian Machine learning Technique

Deployment of wireless sensor networks are usually found in military applications for intrusion detection, forest fire detection, environmental monitoring, civil applications etc. Deployment of sensors are application specific and poses many concerns and limitations like data redundancy, node clustering, discovery of useful information, localization and others. One of the critical issues that needs attention is the deployment of optimal number of sensor nodes in a given environment (phenomenon) based on the dynamics and redundancy in measured data to conserve energy and avoiding redundant nodes in wireless sensor network (WSNs) for cost effectiveness. To address such challenges, machine learning plays a vital role. In order to provide a solution to WSN to grasp the dynamics of the environment, we propose a novel machine learning technique i.e. Hebbian learning model that learns effectively theenvironmental changes and provides a better decision-making for optimal number of sensors deployment and energy saving in a given location. The proposed model is implemented by considering environmental parameter values with hebbian machine learning technique and is verified for its performance. We found that hebbian learning has provided a better solution by learning and adapting to the dynamics of the environment with optimal number of sensor nodes by considering redundant data flow between the nodes and intern eliminating the redundant nodes that saves energy and node cost, and also prolong the lifetime of the network without suspending the monitoring activity.