Minimizing Collision of Fading Channel Using Machine Learning

Energy consumption is considered the main challenge of MAC protocol design. Especially when MAC protocol is employed in an environment of limited energy resources as a wireless sensor network. Parameters optimization of the shared channel in sensor communications is the aim of any MAC protocol designer. In this paper, we suggest a machine learning-based approach for the improvement of the performance parameters using channel prediction learning. Channel predication learning ensures that all the learning process is done by the node. The proposed machine learning algorithm takes into consideration the fading channel parameters and suggests a solution that is best suited to optimize the performance parameters. The proposed machine learning approach incorporates the use of Sensor-MAC (SMAC) protocol and suggests the best tuned MAC protocol based on the supervised learning GRNN algorithm. We investigate the system performance using simulation scenarios under various configurations. The overall performance improvement is more than 80% based on all the output performance parameters.