Resonance-Forming: Achieving High Gain Metamaterial-Enhanced Magnetic Induction Communication Based on Reinforcement Learning

Limited communication range is the major drawback of Magnetic Induction(MI) communication. Due to the subwavelength aperture of small antennas used in MI communication, radiated energy gathers around the antenna in the form of evanescent waves and cannot propagate away. State-of-theart research surrounds antennas with metamaterials to radiate the evanescent waves away, resulting in increased transmission distance. However, existing Metamaterial-Enhanced Magnetic Induction <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\left(M^{2} I\right)$</tex> communication research ideally assume isotropy and periodicity of metamaterials, which leads to significant reduction of channel gain. To achieve higher channel gain and further transmission distance, in this paper we propose Resonance-Forming, a scheme based on reinforcement learning for adaptively adjusting metamaterial, to increase communication range in <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$M^{2} I$</tex> communication system. numerical results validate that in low-loss and high-loss medium the Resonance-Formed <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$M^{2} I$</tex> communication ranges are respectively increased by 5.11 and 4.96 times than existing uniformly-tuned <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$M^{2} I$</tex> while maintaining a channel capacity of no less than 10 Kbps.