Symbol detection under channel uncertainties for MIMO systems

Abstract

In this thesis we analysis channel errors of mobile 1VIIMO channels. We classify and analysis channel uncertainties and propose a specified model for channel errors. We propose a modified detection techniques to find a sub-optimal solution. Additionally we propose a model-robust symbol detection technique solving total least squares problem which is not affected by statistical changes of noises or channels. And also we derive an optimal regularization factor for channel inversion precoding matrix . We evaluate these techniques by simulating under proper environments. Simulation results show these techniques mitigate unexpected performance degradations especially at high SNR.