An Artificial Intelligence Framework for Holographic Beamforming: Coexistence of Holographic MIMO and Intelligent Omni-Surface
Apurba Adhikary, Md. Shirajum Munir, Avi Deb Raha, Yu Qiao, Sang Hoon Hong, Eui‐Nam Huh, Choong Seon Hong
Abstract
The forth-coming 6G wireless communication systems are required to meet the increasing demand for network connectivity that requires power savings for generating effective beamforming. Therefore, joint sensing and communication framework is proposed with the coexistence between holographic MIMO (HMIMO) and Intelligent Omni-Surface (IOS) which ensures the extension of the coverage area resulting in lower power consumption for beamforming. An optimization problem is formulated maximizing the utility function considering the channel capacity, beampattern gains with zenith and azimuth angles, distances, and losses during the sensing-communication process. An artificial intelligence framework is designed for solving the formulated NP-hard problem. Firstly, long short-term memory (LSTM) based scheme is utilized to determine the zenith and azimuth angles with a view to obtaining the current location of all the users that are served directly by the HMIMO and users that are served with reflective channel and refractive channel from the IOS. Afterward, a reinforcement learning (RL) based method is developed for allocating the communication resources to the intended users for generating the required beamforming based on the results obtained from the LSTM-based scheme. Finally, simulation results illustrate that the proposed artificial intelligence framework achieves a power gain of 1.5% compare to the baseline method to perform holographic beamforming for serving the users.