Hierarchical Taylor quantized kernel least mean square filter for data aggregation in wireless sensor network

The advanced technology in recent years that has achieved more attention among researchers and the social community is the wireless sensor network (WSN) that includes a number of nodes that are commonly distributed in remote zones. While deploying the WSN in huge areas, WSNs produce a massive amount of data. Thus, there is a significant need to process the data through efficient models. The data aggregation technique is the common solution widely employed to obstruct congestion on large-scale WSNs. However, the demanding part of the data aggregation scheme is to mitigate the network overhead without affecting the system efficiency. Most of the data transmitted by sensor nodes are repetitious and thus result in high power consumption. Therefore, sensor nodes should utilize an efficient data aggregation model for data transmission that minimizes duplicate data. In order to maintain such complications, this article proposes a hierarchical Taylor quantized kernel least mean square (HTQKLMS) filter for aggregating data in WSN. For this purpose, WSN is initially simulated, and then data aggregation is accomplished using developed HTQKLMS filter. Additionally, the HTQKLMS is derived by amalgamating the hierarchical fractional quantized kernel least mean square (HFQKLMS) filter with the Taylor series. Here, the data prediction mechanism is done by employing HFQKLMS model that is an integration of quantized kernel least mean square (QKLMS) and hierarchical fractional bidirectional least mean square (HFBLMS). Apart from this, data redundancy is achieved by broadcasting needed data utilizing data detected at the destination. Furthermore, HTQKLMS approach has delivered a minimum energy consumption of 0.0333 J and less prediction error of 0.0326.