A Pilot Based MIMO MMSE Channel Estimation with Channel Covariance Matrix

Abstract

Recently in the past from one decade onwards the improving of channel capacity value and high data rate for transforming information is dominantly attracts the researchers to work in this field so we use MIMO wireless broadband communication system which is one of the method to increase these features as they are proposing many algorithms for achieving better performance. In MIMO system max users share the same bandwidth so quality fades away when the user moves away from the bandwidth. However, for achieving maximum channel capacity we should know the complete knowledge of the channel. So here my aim is to estimate the channel in an MIMO wireless environment. And we are estimating different channel parameters such as amplitude and phase and the results are compared with the ideal parameters. These parameters are estimated by Maximum Likelihood Technique (MLE). So in this paper we study the estimated minimum mean square error (LMMSE) channel with respect to estimated channel correlation matrix where the information of channel is imperfect. We investigate how the statistical information of the channel correlation matrix, affects the system performance of MMSE estimator for multipath channels. We can analyze the MMSE by providing the upper and lower bounds. The accuracy of estimated channel correlation matrix shows how it impacts on system performance, average signal to noise ratio and multipath channel conditions. The minimum mean square error (MMSE) estimator is to exploit the spatially correlated channels but it needs prior knowledge of estimated channel correlation matrix. In practical, the perfect knowledge of estimated channel correlation matrix is not available in wireless communication system, due to this the estimated channel state information is not accurate and errors occur in the system. It results to degrade the performance of system. Channel estimators are analyzed based on imperfect channel state information. Channel estimators may experience irreducible error if the channel is not sampled at regular interval spaced. This thesis will analyze and presents the low-rank approximation of the linear minimum mean square error (LMMSE) and singular value decomposition of the channel estimator. The proposed channel estimators are designed for fixed values and signal to noise ratio. The resulting channel estimators choose with fixed designed parameters, low complexity and better performance