Основные понятия
Efficient implementation of massive MIMO transceivers using finite-precision arithmetic to offset performance gaps.
Аннотация
The article discusses the proposal of a new massive MIMO architecture utilizing finite-precision arithmetic to reduce computational complexity. It covers rounding error analysis, achievable rate bounds for SIMO and MISO systems, multi-user scenarios, and proposes a mixed-precision architecture. The content also explores hardware and computational complexities in transceiver design, GPU acceleration, analog computing, and channel property exploitation. Various architectures like hybrid beamforming and dynamic hybrid architecture are discussed.
Статистика
"Only a few works have investigated reducing the computational complexity of communication systems using finite-precision arithmetic."
"For ZF detection or precoding in such a system, computational complexity is O(MK^2 + K^3)."
"Single-precision data requires half as much storage as double-precision data."
Цитаты
"The challenges are twofold. First, the impact of finite-precision arithmetic on communication performance remains unknown."
"Can we replace full-precision arithmetic with finite-precision or low-precision arithmetic in communication systems?"