An interval-valued recursive estimation framework for linearly parameterized systems

This paper proposes a recursive interval-valued estimation framework for identifying the parameters of linearly parameterized systems which may be slowly time-varying. It is assumed that the model error (which may consist in measurement noise or model mismatch or both) is unknown but lies at each time instant in a known interval. In this context, the proposed method relies on bounding the error generated by a given reference point-valued recursive estimator, for example, the well-known recursive least squares algorithm. We discuss the trade-off between computational complexity and tightness of the estimated parametric interval.