Intelligent system for channel prediction in the MIMO-OFDM wireless communications using a multidimensional recurrent LS-SVM

Abstract

In order to resolve channel prediction in the multiple-input multiple-output orthogonal frequency division multiple (MIMO-OFDM) system used in wireless communication, a novel intelligent system based on least squares support vector machines (LS-SVMs) is proposed in this paper. To manipulate the iterative problem, the recurrent multidimensional version LS-SVM has been used. The proposed algorithm used in this system allows us to implement nonlinear decision regions in the channel prediction in MIMO-OFDM systems, and adaptively convergent to minimum mean squared error solutions. It is shown by simulation that the proposed method is able to provide accurate results in channel prediction in these systems. Moreover, this method can be also used in many signal degradations caused by multipath propagation, shadowing from obstacles, etc. © 2013 Springer-Verlag.