Adaptive Blind MPSK Constellation Recovery and Equalization for Cognitive Radio Applications

• Published in: IEEE transactions on vehicular technology 2022-11, Vol.71 (11), p.11988-12000
• Main Authors: Marrero, Liset Martinez, Gomez, Jorge Torres, Dressler, Falko, Garcia, M. Julia Fernandez-Getino
• Format: Article
• Language: English
• Subjects: Adaptive filter; Adaptive filters; Adaptive systems; Algorithms; Automatic modulation recognition; Band-pass filters; Binary phase shift keying; blind symbol phase estimation; Cognitive radio; Convergence; Delays; Equalization; Filtering algorithms; Interference; LMS; Phase locked loops; Phase shift keying; PSK; Quadrature amplitude modulation; Recovery; Signal classification; Waveforms
• ISSN: 0018-9545; 1939-9359
• DOI: 10.1109/TVT.2022.3194965

Cognitive radio is considered a relevant communication paradigm to deal with the increasing demands in modern communications systems. Adaptive schemes are required to recognize channel conditions and to properly adjust main transmission parameters to improve the quality of communications. In this direction, blind algorithms to recover constellation, from phase-modulated signals, represent a means to implement cognitive capabilities to allow automatic modulation recognition (AMR) on receivers. Commonly, the most popular approaches for blind constellation recovery are based on a two-step scheme. The first step uses to equalize channel effects and reduce inter-symbol interference (ISI). The second step carries out constellation recovery utilizing phase locked loop (PLL) systems like the Costas Loop, then to classify the incoming signal. This work proposes a novel single-step blind adaptive filter solution, inspired by an adaptive interference canceler, for joint equalization and constellation symbol recovery from received phase shift keying (PSK) waveforms. Furthermore, we propose new coefficients update mechanisms based on the constant amplitude of PSK signals. The proposed solution exhibits reduced computational complexity compared to the state of the art and a reduced time of convergence. Additionally, the proposed scheme does not require a training sequence to operate properly. The obtained results clearly show that the proposed scheme significantly improves accuracy regarding phase symbol estimation and ISI reduction.