An intelligent data routing strategy based on deep reinforcement learning for IoT enabled WSNs

Abstract

Routing data across numerous access points becomes complicated due to the frequent relocation of sensor nodes. This can cause packet loss, node selection errors, an inability to extend the lifespan of individual nodes, latency in response times, and an increase in computing complexity. An intelligent data routing strategy based on deep reinforcement learning (DRL) is proposed in this study. It considers factors like message overhead, time complexity, maximum data sum rate, and other factors like reduced communication delay and improved energy efficiency to find an optimistic path for better performance of IoT-enabled WSNs. The fundamental instant data load is divided into different cluster pairs, each containing one strong and one weak sensor node, using the double cluster pairing approach. It can be implemented on any network platform, including mobile and non-mobile nodes, by reducing regulated message overhead and increasing data throughput. The technique proposed here is similar to other existing routing protocols in that it helps nodes last longer while using less power.

Keywords

• Double cluster pairing method
• Intelligent data routing
• Multi-criterion correlation agent model
• Time response delay
• Maximum data rate
• Energy efficiency