Improved zero-forcing equalizer for ISI reduction in DMT systems

In this work, an improved zero-forcing equalizer method for ISI reduction in DMT systems is presented. It is implemented using the LMSm (least mean squared modified) algorithm in substitution of the recursive algorithm used in the original zero-forcing method. As a result a faster algorithm for update the weights of the digital filter are obtained. Thus, the improved zero-forcing structure, being the most remarkable properties: an increase in the speed of identification and a smaller residual error in the channel to be equalized, properties which makes it adequate for real time applications, and provide a better performance in ISI and ICI cancellation, and BER reduction. The properties of this structure has been analytically derived and experimentally extensively verified by computer simulation. The obtained experimental results (frequency response, eye diagram and BER test) show the important improvements above mentioned.