Parameter estimation of noisy autoregressive signals

The problem of estimating the parameters of a noisy autoregressive (AR) signal is considered. We propose a new least-squares (LS) method for estimating AR parameters that uses both low-order and high-order Yule-Walker equations in a new way. This estimate is biased. We derive a new method for noise variance estimation to yield unbiased LS estimate of the AR parameters. To evaluate the performance of the proposed method, computer simulations are performed. Simulation results illustrate that the performance of the proposed method is much better than the other estimation methods.