Beamforming with Free Energy Principle under Hierarchical Codebook
Tatsuya Otoshi, Masayuki Murata · 2024
Beamforming plays a crucial role in enhancing the performance of wireless communication systems. However, achieving optimal beamforming entails a trade-off between exploration and exploitation, where the system needs to balance the exploration of different beam directions with the exploitation of the best available beam. Motivated by the exploration-exploitation trade-off, we propose the FEP method, which leverages hierarchical modeling and adaptive beam switching to optimize this trade-off. Through simulations in a dynamic environment, we evaluate the performance of the FEP method in terms of expected free energy minimization and signal-to-interference-plus-noise ratio (SINR) maximization. The results demonstrate that the FEP method effectively maintains high SINR levels through adaptive beam direction selection, reflecting efficient exploitation. Comparative analysis with other beamforming methods further highlights the superior performance of the FEP method in terms of average signal quality and stability.