Improving Duty Cycle-based MAC Protocol in Wireless Networks using AI and Machine Learning

Abstract

Duty cycle of a Medium Access Control (MAC) protocol is made up of sleep phase, wake-up phase and listen phase. MAC protocols usually proposes to optimize the duration of the wake-up and listen phases, in order to increase the duration of the sleep phase, thereby reducing the unwanted energy consumption of the wireless node. In this paper, we propose an Artificial Intelligence (AI) and machine learning (ML) based approach, which uses a hybrid combination of Time Division Multiple Access (TDMA), Bitmap Assisted MAC (BMA) and Sensor MAC (SMAC). The machine learning layer utilizes the duty cycle in the MAC layer, and generates multiple solutions for a given wireless communication. The AI layer then selects the best solution from the generated solutions by incorporating a duty cycle factor in the selection function, thereby optimizing the duty cycle of the protocol. The proposed system shows a 15% improvement in communication speed, and a 10% reduction in energy consumption across multiple communications. We plan to further extend this work for rural India, and apply it to real time agricultural applications.