2-D non-causal systems: State space modeling for half-plane support

Abstract

In this paper Mendel and Hsueh's [10] state-variable modeling technique for transforming a 1-D non-causal system into a causal system is extended to 2-dimensional systems which have half-plane (HP) supports. The 2-D impulse responses treated here are restricted to a class in which their associated 2-D z-transforms have separable denominators. The final state space model is a special type of so-called Roesser's model. It is a very low-order model, which is important from a computational point of view, when, for example, 2-D recursive estimation algorithms are applied to 2-D systems with HP supports.