A STUDY ON FAST CONVERGENCE BLIND ADAPTIVE EQUALIZER WITH FRACTIONAL TAP SPACE

在现代通信系统中，信号均衡是实现快速比特率的必要技术。为了实现这一目标，不需要任何训练信号的盲均衡技术正受到越来越多的关注和需求。本文提出了一种新的对加性噪声具有鲁棒性的盲系统辨识方案。该算法在噪声信号不相关的情况下，即使没有统计信息，也可以进行系统的盲识别。此外，本文还推导了传统方案所没有的梯度型自适应算法，并进行了计算机仿真，验证了该方法的有效性。

In recent telecommunication systems, signal equalization is essential and necessary technique to achieve fast bit rate. In order to make it possible, blind equalization techniques, which do not need any training signals, are gathering a lot of attention and demand. In this paper, we propose a novel blind system identification scheme which is robust for additive noise. The proposed algorithm can carry out blind system identification even if there is no statistical information in terms of noise signals as far as it is uncorrelated. In addition, gradient type adaptive algorithm, which is not available in conventional schemes, is also derived and then some computer simulations are shown in order to show the effectiveness of the proposed method.