Application of Zernike Moments as Features in KNN and SVM as Semi-Blind Detectors for STBC MIMO-OFDM Systems in Impulsive Noise Environments

In space time block coded (STBC) MIMO-OFDM, most receivers assume the channel noise to be Gaussian, this paper considers the semi-blind detection of symbols in non-Gaussian channels. Our aim is to introduce a new receiver for STBC MIMO-OFDM transmission over Rayleigh fading channels using Zernike moments which apply either k-nearest neighbor (KNN) or support vector machine (SVM) for classification. The detrimental effects of multipath on the transmitted symbols are alleviated using invariance to rotation, and scaling properties of the Zernike moments. The performance of the new method is shown in symmetric α-stable (SaS) and Middleton noise environments.