An Improved SIC-Based Detection Scheme for Non-Uniform Constellations in ATSC 3.0 MIMO

• Published in: IEEE transactions on broadcasting Vol. 68; no. 2; pp. 286 - 294
• Main Authors: Kim, Hyeongseok, Kim, Jeongchang, Park, Sung-Ik, Hur, Namho
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.06.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Advanced television systems committee 3.0; Cancellation; Codes; Complexity; Complexity theory; Interference; Interference cancellation; Maximum likelihood decoding; MIMO communication; multiple-input multiple-output; Quadrature amplitude modulation; real-valued representation; Representations; Robustness; Silicon carbide; suboptimal decoding; successive interference cancellation; Symbols; Television systems
• ISSN: 0018-9316; 1557-9611
• DOI: 10.1109/TBC.2022.3140675

This paper presents an alternative approach to real-valued signal representation and proposes an improved successive interference cancellation (SIC)-based detection scheme using the alternative signal representation for non-uniform constellations (NUCs) in Advanced Television Systems Committee (ATSC) 3.0. Since I/Q polarization interleaving in ATSC 3.0 multiple-input multiple-output (MIMO) precoding performs a non-linear operation, it makes it difficult to use conventional suboptimal detection schemes based on interference cancellation. Therefore, the alternative signal representation is designed to easily apply various suboptimal detection schemes. Further, to obtain more robustness and lower complexity, a suboptimal detection scheme based on the alternative signal representation and interference cancellation with QR decomposition is proposed. The proposed detection scheme exploits the structural properties of NUCs and jointly detects the real and imaginary parts as a block. In addition, by considering one or more candidate symbols in the interference cancellation for the proposed detection scheme, the performance and the complexity are trade-offs according to the number of the considering candidate symbols. Simulation results show that the proposed detection scheme with a slight increase of the complexity significantly outperforms a linear detection under fixed and mobile channels.