Leveraging AI in Mixed Hierarchical Topologies to Improve WSN: A Survey

Abstract

The Internet of Things (IoT) as Wireless Sensor Networks (WSNs) holds a significant role in various areas: Military surveillance, Industrial automation, Smart housing, Security Systems, Intelligent Vehicular Traffic control, as well as Healthcare monitoring. So, the sensor nodes performance and life-time with limited power, memory and processing capabilities are significant issues to face in the network. To overcome these limitations, a wide range of energy-aware packet forwarding mechanisms and routing protocols have been proposed to optimize network throughput and lifetime. A WSN's performance is strongly dictated by its node deployment approaches, energy consumption profiles, communication latency, and data aggregation techniques. Furthermore, as the aspects of the network topology (chain-based or cluster-based hierarchical) and the criteria for choosing aggregator nodes to further transmit the sensed data to the sink node strongly affect delivery time and energy balance. Existing hierarchical routing protocols are well studied by several surveys, for example LEACH, PEGASIS, PDCH, CHIRON, CCBRP, CCM, TSCP, DLRP and DCBRP. To do so, it underlines why the addition of Artificial Intelligence (AI) to mixed hierarchical topologies is improving decision-making processes, and adaptive clustering while making WSNs more efficient, scalable and resilient in their performance. We further provide a comparative analysis, primarily emphasizing the advantages and limitations of different chain-based and cluster-based AI-assisted routing solutions.