A Deep Policy Dynamic Programming Based Intelligent Data Routing Scheme for IoT-Enabled Wireless Sensor Networks

Nowadays, the Internet of Things (IoT) plays a significant role in the development of various real-life applications such as smart cities, healthcare, precision agriculture, and industrial automation. Wireless Sensor Networks (WSNs) are a major ingredient of these IoT-based applications. In WSNs, sensor nodes that are close to the Base Station (BS) relay more data packets compared to other nodes, which creates high energy consumption at nodes close to the BS. As a result, an energy imbalance is created among the sensor nodes. Therefore, sensor nodes close to BS die early as compared to the faraway sensor nodes. These early dead nodes drastically increase data collection delay within the network. Furthermore, the early death of the sensor nodes partitions the network into different isolated sub-networks/segments. The formation of isolated segments causes premature death of the network. This paper proposes a Deep Policy Dynamic Programming (DPDP) based intelligent data routing scheme for IoT-enabled WSNs. The proposed scheme identifies an optimal number of Cluster Heads (CHs) and forms clusters to reduce the energy consumption of the deployed sensor nodes and prevent the early death of sensor nodes. Furthermore, the proposed scheme identifies an optimal number of Rendezvous Points (RPs) and designs an optimal path for Mobile Sink (MS) based data collection. Optimal RP selection and path design algorithms prevent the premature death of the network and significantly improve the overall performance of the network. Extensive simulations and test-bed experiments are conducted to test the performance of the proposed scheme. The simulation and test-bed results show that the proposed scheme outperforms as compared to the existing state-of-the-art approaches in terms of network lifetime, network stability, data loss due to buffer overflow, residual energy, and delay.