A Cluster-Based Data Fusion Technique to Analyze Big Data inWireless Multi-Sensor System
With the development of the latest technologies and changes in market demand, the wireless
multi-sensor system is widely used. These multi-sensors are integrated in a way that produces an overwhelming
amount of data, termed as big data. The multi-sensor system creates several challenges, which include
getting actual information from big data with high accuracy, increasing processing efficiency, reducing power
consumption, providing a reliable route toward destination using minimum bandwidth, and so on. Such
shortcomings can be overcome by exploiting some novel techniques, such as clustering, data fusion, and
coding schemes. Moreover, data fusion and clustering techniques are proven architectures that are used for
efficient data processing; resultant data have less uncertainty, providing energy-aware routing protocols.
Because of the limited resources of the multi-sensor system, it is a challenging task to reduce the energy
consumption to survive a network for a longer period. Keeping challenges above in view, this paper presents
a novel technique by using a hybrid algorithm for clustering and cluster member selection in the wireless
multi-sensor system. After the selection of cluster heads and member nodes, the proposed data fusion
technique is used for partitioning and processing the data. The proposed scheme efficiently reduces the blind
broadcast messages but also decreases the signal overhead as the result of cluster formation. Afterward, the
routing technique is provided based on the layered architecture. The proposed layered architecture efficiently
minimizes the routing paths toward the base station. Comprehensive analysis is performed on the proposed
scheme with state-of-the-art centralized clustering and distributed clustering techniques. From the results,
it is shown that the proposed scheme outperforms competitive algorithms in terms of energy consumption,
packet loss, and cluster formation.
KeyWords
Data fusion, big data, clustering, multi-sensors, layered architecture.
|