Use of Biased Estimators for Improved Channel Estimation in OFDM Systems

This paper addresses the problem of improved channel estimation in OFDM systems. It is a well known fact that the knowledge of the exact tap delay locations enables the construction of the modified least square [m-LS] estimator, which happens to be the best linear unbiased estimator [BLUE].Inherent frailties such as high condition number of the linear transformation matrix and the consequent high noise sensitivity ensures that the m-LS estimator often performs poorly at low to moderate SNRs. This motivates the usage of certain biased estimators which attack these short comings of the least square estimator. We propose two biased channel estimators, derived from the rank shaped least squares [RSLS] estimator and the generalised Liu [GL] estimator to attain superior performance. The simulation results involving downlink channel estimation in WiMAX systems vindicate the usage of the above mentioned estimators.