Enhancing Energy Efficiency and Data Reliability in Wireless Sensor Networks Through Adaptive Multi-Hop Routing with Integrated Machine Learning

Wireless Sensor Networks (WSN) plays an important role in monitoring and data acquisition process in various fields of application such as environmental, health care and smart city. However, WSNs present some acute issues including energy constraints, data credibility, and ability to function in a dynamically changing environment. This paper therefore presents an adaptive multi-hop routing protocol based on machine learning and proposes a novel architecture that focuses on solving these challenges. The adaptive protocol switches to the best paths without prior notice depending on the available node energy, link quality, and data priority the machine leaning estimates the most likely node to fail and makes best routing decisions depending on feature such as residual energy and link quality. To provide balanced load in terms of energy consumption, the proposed framework includes an element of load balancing of traffic periodically. Experiments on NS-3 show that the application of our suggested framework decreases energy consumption on nodes up to 25%, enhances the packet delivery ratio 18%, and network lifetime is 35% higher in contrast with conventional approaches, LEACH and Directed Diffusion. These results suggest that the proposed framework can be readily employed in the context of next generation WSNs to improve performance and longevity.