Higher Convergence Adaptive Equalization Method with Noise Removal Function Using Total Least Squares Method

Abstract

In the communications system, when received signals do not include noises, we can accurately perform blind equalizations. However, when received signals include noises, equalization performance generally deteriorates. To solve this problem, an equalization method using Total Least Squares (TLS) with a noise removal unit was proposed. This method had a problem that it was slower convergence rate because this method was used the gradient method based on TLS for channel estimation and LMS method as the gradient method based on Mean Square Error (MSE) for the calculation of equalizer parameters. Therefore, in this paper, we aim to propose a higher convergence blind equalization method with noise removal unit. In the proposed method, first, for higher convergence rate of channel estimation, we propose a recursive method with an update rule that is like Recursive Least Squares (RLS) method based on TLS, noting that RLS method based on Least Squares is higher convergence rate. Second, the result is used for removing noises. Third, we calculate equalization parameters using estimated channel characteristics. In this calculation, RLS method instead of LMS method as gradient method is used for higher convergence rate. Last, we can achieve higher convergence rate can be obtained with maintaining higher equalization performance by giving these parameters to the equalizer every sample time. The proposed method is evaluated by computer simulation.