Hierarchical Fractional Quantized Kernel Least mean Square Filter in Wireless Sensor Network for Data Aggregation

Abstract

In modern days, Wireless Sensor Network (WSN) is an emerging research area in which various resources constrained sensor nodes are associated by wireless radio with minimum bandwidth. Generally, WSN includes more devices, which are microsensors, wireless radios, and microprocessors. The sensor node in WSN has capable of sensing, self-controlling, offering wireless communication and estimation processing. The information attained by sensors is accumulated resourceful devices, named actuator nodes or central unit, termed as Sink node or base station. The sink node assists in transferring data gathered from a network and vice versa. Therefore, several WSN applications need data collection from sensor nodes based on sink nodes. A productive approach is needed for obtaining data efficiency through decreasing nodal energy consumption. In this research, a Hierarchical Fractional quantized kernel least mean square (HFQKLMS) filter was devised for data aggregation in WSN. Moreover, the HFQKLMS technique was devised by combining Kernel Least Mean Square and Hierarchical Fractional Bidirectional Least-Mean-Square (HFBLMS) approach. Besides, data redundancy is attained by broadcasting the required data using data predicted at the sink node. Besides, the performance of the developed HFQKLMS technique for data aggregation obtained less energy consumption of 0.021 J, and a prediction error of 7.45 based on 100 nodes in the localization database.