Peak to Average Power Ratio Reduction for Space-Time Codes That Achieve Diversity-Multiplexing Gain Tradeoff

TL;DR

This paper analyzes the impact of peak-to-average power ratio constraints on the diversity-multiplexing gain tradeoff in multi-antenna systems and proposes a constellation shaping method for PAPR reduction that preserves optimality.

Contribution

It demonstrates that PAPR constraints above one do not affect the D-MG tradeoff and introduces a PAPR reduction technique compatible with existing optimal space-time codes.

Findings

D-MG tradeoff remains unchanged for PAPR > 1

Proposed PAPR reduction method does not affect code optimality

Sphere decoding confirms performance preservation

Abstract

Zheng and Tse have shown that over a quasi-static channel, there exists a fundamental tradeoff, known as the diversity-multiplexing gain (D-MG) tradeoff. In a realistic system, to avoid inefficiently operating the power amplifier, one should consider the situation where constraints are imposed on the peak to average power ratio (PAPR) of the transmitted signal. In this paper, the D-MG tradeoff of multi-antenna systems with PAPR constraints is analyzed. For Rayleigh fading channels, we show that the D-MG tradeoff remains unchanged with any PAPR constraints larger than one. This result implies that, instead of designing codes on a case-by-case basis, as done by most existing works, there possibly exist general methodologies for designing space-time codes with low PAPR that achieve the optimal D-MG tradeoff. As an example of such methodologies, we propose a PAPR reduction method based on constellation shaping that can be applied to existing optimal space-time codes without affecting their optimality in the D-MG tradeoff. Unlike most PAPR reduction methods, the proposed method does not introduce redundancy or require side information being transmitted to the decoder. Two realizations of the proposed method are considered. The first is similar to the method proposed by Kwok except that we employ the Hermite Normal Form (HNF) decomposition instead of the Smith Normal Form (SNF) to reduce complexity. The second takes the idea of integer reversible mapping which avoids the difficulty in matrix decomposition when the number of antennas becomes large. Sphere decoding is performed to verify that the proposed PAPR reduction method does not affect the performance of optimal space-time codes.