Performance evaluation of blind channel identification methods based on oversampling
Ernesto Leite Pinto, C.J.A. Silva · 2002
This paper addresses the performance evaluation of two methods that marked the beginning of a brand new direction in blind channel identification using only second-order statistics: TXK and subspace methods. Computer simulations are conducted to investigate the following features of the algorithms: sensitivity to SNR, rate of convergence, computational complexity, sensitivity to noise correlation, robustness to error in model order estimates.