E2IA-HWSN: Energy Efficient Dual Intelligent Agents based Data Gathering and Emergency Event Delivery in Heterogeneous WSN Enabled IoT

• Published in: Wireless personal communications Vol. 122; no. 1; pp. 379 - 408
• Main Authors: Susan Shiny, G., Muthu Kumar, B.
• Format: Journal Article
• Language: English
• Published: New York Springer US 2022; Springer Nature B.V
• Subjects: Algorithms; Clustering; Communications Engineering; Computer Communication Networks; Energy consumption; Engineering; Intelligent agents; Markov chains; Multiple objective analysis; Networks; Optimization; Residual energy; Scheduling; Sensors; Signal,Image and Speech Processing; Sleep; WSN-IoT; Sleep scheduling; Data redundancy; Emergency event; Grid; Data gathering
• ISSN: 0929-6212; 1572-834X
• DOI: 10.1007/s11277-021-08904-3

Heterogeneous sensors are equipped with a limited battery source that is concerned with network lifetime problems. However, this problem can be tackled with the effective design of WSN-IoT by clustering and sleep scheduling mechanisms. This paper addresses this issue by presenting novel ideas involved in the WSN operations such as grid construction, cluster head selection, sleep scheduling, and data gathering by intelligent Agents (iAgents). An energy-efficient dual iAgents based Heterogeneous WSN (E2IA-HWSN) is proposed. iAgents are used in this paper to automatically collect the sensed data from IoT sensors. In this E2IA-HWSN, a 3 × 3 grid is built and each cell is sub-divided into four in which cluster heads (CH) are selected in each sub-division, followed by ring partitioning for selecting a CH present at the center. Multi-Objective Harris Hawks optimization (MO-HHO) algorithm is used to select CH and supernode, here to minimize the energy consumption of CH, the supernode takes responsibility to assign sleep schedules to devices. The scheduling slots are assigned only after a sensor reaches below the energy threshold. For scheduling, the Bayes rule-based Markov model (BR-MM) is applied with the determination of residual energy and sensed packet counts. Generator de Bits Pseudo Aleatorios (GBPA) eliminates redundant data in CH and then inter-cluster routing is performed in case of emergency events. If not, then the CH waits for the arrival of iagents, the trajectory of iAgents is dynamically predicted with Deep Policy Gradient (DDPG). The implementation is carried out in NS3.26 and the results show betterment to the well-known methods.