Energy-efficient Q-learning-based routing in wireless sensor networks

Wireless sensor networks (WSNs) are extensively used to collect and transmit data from the environment through sensor nodes for various applications such as agriculture, medicine, and military. The sensor nodes are located in remote locations and harsh environmental conditions. Sensors operate on a limited power source. The lifetime of the entire WSN depends on the power of its nodes. Also, successful data transmission via routing from the source node to the destination depends on the battery life of the sensor node. It therefore becomes necessary to have an energy-efficient routing path to ensure successful packet delivery from source to sink nodes. The work proposes an energy-aware Q-learning-based routing algorithm in WSN. To establish an energy-efficient route, the work exploits the rewards in Q-learning based on the node energy. Experiments conducted using the proposed work demonstrate enhanced packet delivery, throughput, and networks lifetime. A comparison of the work with a basic Q-learning-based routing algorithm exhibits improved performance.