Periodicity-Enabled Size Reduction of Symbol Based Predistortion for High-Order QAM

• DOI: 10.1109/jlt.2022.3191464
• Publication Date: 2022-07-15T19:36:43Z
• Authors: Zonglong He; Jinxiang Song; Kovendhan Vijayan; Christian Hager; Alexandre Graell ...; Henk Wymeersch; Peter A. Andrekson; Magnus Karlsson; Jochen Schroder
• Keywords: Lookup table; QAM; Modulation (music)

We experimentally demonstrate a novel size reduction approach for symbol-based look-up table (LUT) digital predistortion (DPD) of the transmitter impairments taking advantage periodicity in pattern-dependent distortions. Compared to other reduced-size LUT schemes, proposed method can significantly lessen storage memory requirements with negligible performance penalty high-order modulation formats. To further alleviate restriction, twice scheme is provide reduction. Importantly, given targeted length, we verify importance averaging over sufficient occurrences patterns obtain well-performing LUT. Moreover, it necessary evaluate LUT-based DPD using random data. Finally, neural network (NN) based nonlinear technique, which achieves nearly identical full-size all employed constellations and robust against change format. The techniques are verified back-to-back transmission experiment 20 Gbaud 64-QAM, 256-QAM, 1024-QAM signals considering 3 5 symbol memory. validated noise loading experiment.