AbstractParameter bounding offers a useful alternative to point parameter estimation methods when either statistical hypotheses on the errors are not met or uncertainties are better characterized in a deterministic way (e.g. systematic, round‐off, truncation errors). So far, many efforts have been devoted to the problem of parameter bounding in linear systems, where exact parameter uncertainty intervals can be computed. In contrast, only partial results have been found for general non‐linear parametrization, namely either upper or lower bounds on parameter uncertainties can be evaluated. In this paper we derive approximate parameter uncertainty intervals for a class of discrete bilinear systems with bounded output errors. This work is based on a linear input‐output parametrization and previous results on bounded errors‐in‐variables models. For an extensively simulated example, central estimates by means of the bounded errors‐in‐variables approach and least squares estimates are computed
展开▼
