WSN-IoT enabled Fractional Eel and Grouper Optimizer based Siamese Convolutional Neural Network for aqua quality prediction in distributed environment

Abstract

Efficient water pollution monitoring solutions are increasingly important due to advancements in Wireless Sensor Network-based Internet of Things (WSN-IoT), communication technology, and sensors. In this research, Fractional Eel and Grouper Optimizer (FEGO) for Cluster Head (CH) selection and routing, as well as FEGO based Siamese Convolutional Neural Network (FEGO_SCNN) for aqua quality prediction, are designed. Initially, the WSN-IoT system model is simulated, where IoT nodes are employed for sensing data from water. Thereafter, the selection of CH and routing is executed by FEGO, which is modeled by integrating Eel and Grouper Optimizer (EGO) with Fractional Calculus (FC). The considered multi-objectives are delay, energy consumption, throughput and Link Life Time (LLT). At Base Station (BS), aqua quality prediction is performed by considering input data from the water quality dataset. Then, data partitioning is conducted employing Deep Embedded Clustering (DEC). Afterwards, partitioned data is given to MapReduce framework comprising of mapper as well as reducer phases. In a mapper phase, the min-max normalization technique is employed for data normalization. Lastly, aqua quality is predicted in the reducer phase by the Siamese Convolutional Neural Network (SCNN), and it is trained by FEGO. In addition, FEGO attained a minimal delay of about 0.712 ms as well as maximal energy and throughput of about 0.286 J and 88.798 Mbps for dataset 1. Also, the FEGO attained delay, energy, and throughput of 0.698 ms, 0.280 J and 87.022 Mbps for dataset 2. Moreover, for dataset 1, FEGO_SCNN obtained a maximal value of accuracy, about 90.797%. Similarly, FEGO_SCNN achieved 88.981% of accuracy for dataset 2.