Optimizing RF Energy Harvesting Systems for IoT Applications using Reinforcement Learning

In this work, we discuss Radio-Frequency Energy Harvesting (RFEH) systems, focusing on their application in loT networks. The efficiency of RFEH is critical for low-power devices and sensors in applications like loT networks. By optimizing rectification using fine-tuned Schottky diode parameters, loT application performance is enhanced. This optimization leads to efficient conversion of RF energy into stable DC power. RFEH is especially valuable in remote areas without conventional power sources and for biosensors. The use of AI -driven optimization methods, specifically reinforcement learning, enhances diode performance, leading to improved energy efficiency and autonomy in loT devices. The study compares the AI-optimized diode with commercial Schottky diodes under various conditions. The AI-optimized diode outperforms commercial counterparts in terms of rectification efficiency, consistently maintains an efficiency above 10 percent in scenarios involving lower input powers and higher temperatures, with the potential to achieve efficiency rates of up to 40 percent. These findings present a compelling solution for applications that demand high rectification efficiency and enhanced performance in the face of temperature and input powers variations.