Abstract

Wireless Sensor Networks (WSNs) offer versatile deployment options, particularly in battery-powered scenarios, addressing energy consumption concerns among sensor nodes. However, the data-intensive nature of WSNs poses challenges in routing, particularly in maintaining balanced paths while accommodating rapid data acquisition. This paper presents an innovative approach called Similarity Estimation-Based Stochastic Gradient Descent (SESGD) routing for WSNs, designed to establish stable routing paths that align with the speed of data acquisition. Sensor nodes play a crucial role in data collection and acquisition, while WSNs facilitate data routing through multiple hops from source to sink nodes. SESGD effectively manages data routing, synchronizing it with data acquisition rates, thereby ensuring network stability. Simulation results assess key performance metrics, including average delay, throughput, and network energy efficiency. The findings demonstrate that the proposed machine learning method outperforms existing algorithms, achieving superior network throughput.