A Survey on Data Prediction Techniques in Wireless Sensor Networks (WSN)

A Wireless Sensor Network (WSN) not only forms the basis of emerging smart technology but also has its own unique challenges. Wireless Sensor Network usually consists of hundreds or thousands of small sensor nodes which operate autonomously in different critical and uncontrolled environments. Since these battery operated sensor nodes are difficult to recharge, achieving energy efficient data transmission in WSN still remains a challenging task. Continuous data transmission among nodes in a WSN is found to be the main reason for earlier discharge of the sensor nodes’ battery power. Data collection and prediction is a crucial and effective operation to achieve longer network lifetime by reducing the frequent communication among nodes. A detailed survey on various data prediction techniques and algorithms is presented in this paper. The data samples collected during the data transmission and the prediction scheme are analyzed and it is proved that data prediction can drastically reduce energy consumption of sensor nodes. Data prediction enables each node to predict its own sensor data and to compare the predicted values with the actually observed data thereby generating a confidence value. To boot our work, we have analyzed various prediction approaches and have proposed a new approach that reduces communication and computational cost without compromising data quality.