Efficient Minimum Finders for Wi-Fi 7 Dimension Reduction Soft Demodulator

In this paper, we propose efficient minimum finders for dimension reduction soft demodulators (DRSD) of Wi-Fi 7 modems with 2x2 MIMO. Minimum finder hardware architectures of DRSD can consists of multiple tree based minimum finders with multiplexers those select valid Euclidean distances (ED) of symbols for each bit value. Firstly, the proposed minimum finder separates quadrature amplitude modulation (QAM) into two pulse amplitude modulations (PAMs) by finding intermediate real and imaginary minimum EDs among EDs of symbols with the same real parts and those with the same imaginary parts respectively. Then, the minimum EDs for each bit value are found among the intermediate minimum EDs in the first step. The proposed minimum finder is efficient for symbols of all constellation points and reduced candidate symbols with determined shape. For symbols of all constellation points of 2m-QAM, the complexity of the number of minimum values is O(2m) compared with O(m·2m) of the conventional method. For reduced candidate symbols of 4096-QAM, we propose the initial candidate reduction (ICR) method using 5x5 square symbols and 6 mirror symbols for each spatial stream. The areas of the proposed minimum finder and previous minimum finders without using shapes are compared when synthesized with 4 nm CMOS technology for 480 MHz operation. The synthesized area of the proposed minimum finder is 35 % smaller than that of a previous minimum finder with shared minimum values.

Efficient Minimum Finders for Wi-Fi 7 Dimension Reduction Soft Demodulator

Soonwoo Choi, Minki Ahn, Taeshin Park, Junyoung Jeong, Samsung Electronics