EXPONENTIALLY FADING MLS ESTIMATION OF 2-DIMENSIONAL NONCAUSAL AND NONSTATIONARY SAR MODEL PARAMETERS

This paper studies the problem of recursive estimation of two-dimensio nal (2-D) noncausal simultaneous autoregressive (SAR) model parameters , and develops an exponentially fading spatially recursive algorithm b ased on the recently proposed modified least squares (MLS) estimation method. The number of multiplication operations in the proposed algori thm is only about 55% of the number in its existing sliding counterpar t, while its tracking performance is similar if the fading factor is c arefully selected. A numerical example is included to demonstrate the performance of the proposed algorithm.