AN EFFICIENT ENERGY-AWARE ROUTING MANAGEMENT TO ENHANCE NODE LIFETIME USING DEEP Q-NETWORK IN WIRELESS SENSOR NETWORKS

Povzetek

Wireless Sensor Networks (WSNs) are used for reliable communication and monitoring, allowing data to be transferred to long-range places. However, because WSNs run on batteries, energy-efficient multi-hop routing is necessary for data transmission. Higher communication delays, increased latency, and a shorter network lifetime are inabilities of traditional WSN routing systems that cannot capture all network information throughout the routing process. In order to overcome these problems, we suggest a routing method that makes use of deep reinforcement learning to send data via multi-hop communication from the origin to the destination. The Deep Q-Network uses Markov Decision Process (MDP) to assist in choosing the best multi-hop path for data transmission. In order to choose the next hop neighbour and provide an ideal path for data packet transmission and network lifetime extension, the DQN takes into account the node's current residual energy, queue capacity, and hop count. Simulations are conducted to analyze the convergence rate, network latency, routing overhead, and network lifetime, the proposed MHRP-DQN protocol performs better random deployment and traditional Q-Learning network. Additionally, it guarantees higher throughput and packet delivery ratio during data transmission in WSNs demonstrating promising performance.