A Dynamic System Matching Technique: A State Space Formulation

Abstract

Analog systems are typically modelled by differential equations with several coefficients as design parameters. Manufacturing processes produce systems whose coefficients vary from the desired coefficients of the system designs. In [1], the authors presented a method for combining the measurements from several analog systems, that are nominally the same, using a technique based on the concept of dynamic element matching discussed in [2]. A heuristic analysis in [1] and a simulation in [3], indicated that this method can effectively control the output errors when the outputs of the several systems are corrupted due to manufacturing errors. In a recent paper [4], a rigorous analysis of a dynamic system matching technique (DSMT) applicable to a small class of scalar single input-single output (SISO) analog systems showed that for that class of systems the DSMT reduces the power in the output noise due to manufacturing errors. In this paper, those results are generalized to a class of linear SISO systems defined by general linear state space equations.