Stepwise regression analysis based decision tree classifier for target tracking in WSN

Abstract

Target tracking is a key problem to be resolved in Wireless Sensor Network (WSN). In addition, energy consumption during the tracking process is the key concern to prolong the network lifetime. In existing works, many techniques designed for target tracking in wireless network. However, target location and trajectory tracking accuracy was not sufficient. Further, energy utilization during target tracking was not minimized. In order to overcome such limitations, Stepwise Regression Analysis based Decision Tree Classifier (SRA-DTC) Model is proposed. In SRA-DTC, three types of nodes, namely reference node, sensor node and target node are initialized. Reference node transmits the beacon message to all sensor nodes in order to discover target node location. When the target node enters into the network, sensor node senses the target node and sends the sensed information to the base station. After that, base station performs the stepwise regression analysis of sensed data to track the target with lesser energy consumption. After that, target trajectories are identified through Decision Tree Classifier in SRA-DTC model. Base station uses Decision Tree Classifier to identify the target trajectories based on the collected information. By this way, tracking accuracy of target location and trajectory is gets improved. The simulation process of SRA-DTC model is carried out on factors such as target tracking accuracy, target tracking time, energy consumption, and network lifetime with respect to number of sensor nodes. The simulation result shows that the SRA-DTC model is able to increases the target tracking accuracy and also reduces the energy consumption in WSN as compared to state-of-the-art works.