A comparison on efficient MA parameter estimation using second and fourth order statistics for communication channels

In this paper we consider efficient and simple, yet reliable parameter estimation techniques for time-varying communications channels. It is noted that in usual communication channel signal-to-noise ratios (SNR) is reasonably high (>30dB) and thus second order autocorrelation based methods provide efficient moving average (MA) parameter estimation using a very small number of samples, which is useful in time-varying channels. However, most communication channel scenarios consist of nonminimum phase (NMP) multipath characteristic. Therefore, it is necessary to use 4th order cumulant in conjunction with the second order statistics for the channel estimation. Efficient channel estimation by the combination of 2nd and 4th order statistics is discussed in this paper.