Energy Efficient Non-Orthogonal Signalling with Probabilistic Shaping for Wireless Transmissions

This work explores the realistic operation of a recently reported concept by the authors
^1
1
The probabilistically shaped-SEFDM concept was first reported by the authors in the IEEE VTC-Spring 2021. This paper extends earlier transmission systems design by using channel estimation and equalisation system components. The results of this work are published here openly for the first time. They form part of Xinyue Liu's PhD 2023 thesis [1]., which considers the use of probabilistic shaping for the non-orthogonal multicarrier spectrally efficient frequency division multiplexing (SEFDM) system. In this paper, we extend the operation of earlier signal and system designs, which considered only a simple AWGN channel, to realistic wireless environments with the focus of the studies on two frequency selective channels, namely a four-tap channel and Rummler's channel. The special system design for probabilistic shaping adopts a reverse concatenation architecture that cascades the constant composition distribution matching (CCDM) algorithm together with soft-decision forward error correction (SD-FEC) - LDPC code. The paper presents detailed system modelling and simulation results, which clearly establish that, with pilot tone based channel estimation and equalisation, SEFDM with probabilistic shaping shows significant advantages, in both energy efficiency and robust-ness against wireless channel impairments, when compared to OFDM of the same spectral efficiencies.