A novel channel estimation strategy with chaotic coded signals

Abstract

A novel dynamic based deconvolution (DBD) algorithm is presented that can be exploited in blind channel estimation as well as in non-blind applications like echo cancellation in telephony circuits. For generating the chaotic signal the unimodal logistic map is employed while the channel is represented by an autoregressive (AR) model. The applicability is demonstrated in a speech transmission scenario using chaotic coded speech showing a modelling misadjustment improvement (MMI) of 24 dB/100 iterations which is 6 fold that obtained by the LMS for a 128 tap digital adaptive filter (DAF) using a white noise (WN) process.