Non Data Aided parameter estimation for Multi-User ARGOS receivers

In this paper, parameter estimators are analyzed in the context of Successive Interference Cancelation (SIC) receivers for the ARGOS system. A Non Data Aided (NDA) feed forward estimator is proposed for the amplitude and the carrier phase parameters. Time delays are assumed to be known. A Window Accumulator (WA) is used to reduce the influence of the additive noise. In the presence of frequency offset, the window length L cannot be chosen arbitrarily large but an optimal length L_opt can be determined. However, because the estimator induces a different optimal length for each parameter, a trade-off must be made. We show that a window length of around 35 samples induces mean square errors (MSEs) lower than 0.012 for both parameters. The MSE of the proposed estimator is also compared to the Modified Cramér Rao Bound (MCRB).