Capacity-Optimized Design of Reconfigurable Antenna Patterns for MU-MIMO Systems

Reconfigurable antennas have emerged as a promising solution in wireless communication systems to adapt to dynamic environments and enhance system performance. Pattern reconfigurability plays a crucial role in improving antenna adaptability to spatially diverse multi-cluster channels, thus maximizing channel capacity. Existing works primarily address either the mode selection problem from a pre-designed finite pattern set which adjusts only beamwidth and scanning direction and thus can not offer optimal channel adaptation, or the mode design problem of computing continuously adjustable patterns under the assumation of an infinite pattern number but without considering the limit on the pattern number in real-world antennas. To address this gap, this paper proposes a practical design methodology that jointly accounts for the finite number of reconfigurable patterns and their precise optimization over diverse channel conditions. Specifically, channels are grouped based on user location distribution, and a joint optimization approach combining alternating optimization, auxiliary variables, and bisection is employed to design the optimal pattern set under the constraint of limited reconfigurable states. Simulation results demonstrate the effectiveness of the optimized patterns and show that the achieved capacity exhibits an approximately logarithmic relationship with the number of reconfigurable patterns, thereby providing a practical validation of existing theoretical analyses. This work bridges the gap between theoretical mode design and practical implementation by offering a concrete and implementable solution for designing a finite set of reconfigurable antenna patterns tailored to realistic system constraints.