Model Identification and Parameter Estimation

The control engineering literature is replete with countless methods of system identification, particularly for linear systems. This chapter discusses the few basic techniques that have been most commonly applied in physiology system analysis. The fact that most physiological systems are closed‐loop systems can introduce some complications into the process of system identification or parameter estimation. The problem of structural identifiability is intimately coupled to the problem of model‐building. The chapter also introduces a number of techniques for system identification of parametric and nonparametric models. It reviews some of the basic computational techniques commonly employed in the identification of single‐input, single‐output, open‐loop systems. Since most physiological control processes operate under closed‐loop conditions, researchers have applied a variety of techniques to "open the loop" by isolating the components of interest from other components that comprise the entire system.