Recursive least-squares identification of glucose dynamics

A recursive least-squares (RLS) parameter estimation procedure for identifying metabolic features of the glucose-insulin system in man is developed. The proposed approach is based on the method of computing a discrete-time mathematical model of blood glucose and insulin concentrations in response to externally excited glucose loads. The discrete-time model is derived from Bergman´s minimal model, a continuous-time model for identifying glucose and insulin dynamics. The estimated parameters consist of two factors indicating a metabolic disorder in the ability to dispose of glucose from blood at normal rates. These two factors are glucose effectiveness and insulin sensitivity index. The estimates from the RLS estimators are compared with results from Bergman´s minimal model method of parameter estimation. The results of the RLS identification agree well with the minimal model prediction