ICA Based MIMO Transceiver For Time Varying Wireless Channels Utilizing Smaller Data Blocks Lengths

Independent component analysis (ICA) is a signal processing technique used for blind source separation of the mixed received signals. In wireless communication, it is widely used in multiple input multiple output (MIMO) systems. Generally, the ICA algorithms assume mixing in static or quasi static wireless channels. Performance of these algorithms degrade for time varying scenario, where the quasi stationarity does not hold. Furthermore, if data blocks lengths are reduced for achieving quasi stationarity, the performance of these algorithms further degrade. In this paper, we propose an ICA based transceiver called the MIMO UD-transceiver system that performs well for time varying mixing scenario in the wireless channels. The proposed system also performs well for smaller blocks lengths. We use the well known ICA algorithm, the FastICA algorithm as a bench mark to evaluate the performance of the proposed system. Simulation is performed over 16 QAM signals. We compare results of the MIMO UD-ransceiver with the conventional MIMO transceiver, when both employ FastICA algorithm for separation. We also compare performance of the proposed system with the only available ICA algorithm for time varying wireless channels, the OBAICA algorithm. Results show that the proposed MIMO UD-transceiver outperforms the conventional MIMO transceiver and OBAICA algorithm for time varying mixing scenario in terms of signal to interference ratio and bit error rate.