Physically-consistent Multi-band Massive MIMO Systems: A Radio Resource Management Model

TL;DR

This paper introduces a joint optimization framework for multi-band massive MIMO systems that considers physical antenna characteristics, aiming to maximize sum-rate through antenna spacing, precoding, sub-carrier, and power optimization.

Contribution

It presents a novel circuit-theoretic model for physically-consistent multi-band MIMO systems and develops a joint optimization approach including a new sub-carrier selection and power allocation algorithm.

Findings

Tightly-coupled antenna arrays outperform loosely-coupled ones in bandwidth gain.

The proposed optimization framework achieves higher sum-rate performance.

Dynamic antenna element selection improves system adaptability.

Abstract

Massive multiple-input multiple-output (mMIMO) antenna systems and inter-band carrier aggregation (CA)-enabled multi-band communication are two key technologies to achieve very high data rates in beyond fifth generation (B5G) wireless systems. We propose a joint optimization framework for such systems where the mMIMO antenna spacing selection, precoder optimization, optimum sub-carrier selection and optimum power allocation are carried out simultaneously. We harness the bandwidth gain existing in a tightly coupled base station mMIMO antenna system to avoid sophisticated, non-practical antenna systems for multi-band operation. In particular, we analyze a multi-band communication system using a circuit-theoretic model to consider physical characteristics of a tightly coupled antenna array, and formulate a joint optimization problem to maximize the sum-rate. As part of the optimization, we also propose a novel block iterative water-filling-based sub-carrier selection and power allocation optimization algorithm for the multi-band mMIMO system. A novel sub-carrier windowing-based sub-carrier selection scheme is also proposed which considers the physical constraints (hardware limitation) at the mobile user devices. We carryout the optimizations in two ways: (i) to optimize the antenna spacing selection in an offline manner, and (ii) to select antenna elements from a dense array dynamically. Via computer simulations, we illustrate superior bandwidth gains present in the tightly-coupled colinear and rectangular planar antenna arrays, compared to the loosely-coupled or tightly-coupled parallel arrays. We further show the optimum sum-rate performance of the proposed optimization-based framework under various power allocation schemes and various user capability scenarios.