Robust estimation for correlated observations: two local sensitivity-based downweighting strategies

An iterative robust estimation procedure for correlated observations is proposed, in which the a-prior correlation coefficient matrix is not updated to alleviate the computational burden. Selection of the downweighting strategy plays a key role in the proposed method. Two local sensitivity-based strategies, one is based on the uniformly most powerful test statistics, the other is based on the standardized least squares residuals, are developed and analyzed. Monte Carlo simulations in the GPS network adjustment scenario demonstrate that, the two strategies can provide a certain resistance against the deteriorating effect of outlying observations on the parameter estimates; the former downweighting strategy is superior to the latter one, both in terms of robustness and computational efficiency.