In this paper, a new multistage complex fractionally spaced bilinear perceptron (MCFSBLP) model with faster convergence rate is introduced. It divides the task of equalizing multilevel quadrature amplitude modulation (M-QAM) system into simple, fast and reliable sub-tasks. The computational complexity does not increase as the level of QAM increases in this model. Applications for frequency selective Rayleigh fading (FSRF) channel equalization and co-channel interference (CCI) suppression in 16-level QAM receiver system are considered. Computer simulation results are presented, which suggest that MCFSBLP is quite effective in higher order QAM channel
Published in:
Neural Networks Proceedings, 1998. IEEE World Congress on Computational Intelligence. The 1998 IEEE International Joint Conference on
(Volume:2
)
Date of Conference: 4-9 May 1998