Stored Response Modeling

Abstract

A direct “brute force” method of system identification is presented. The method is based on the definition of a deterministic system and applicable to nonlinear nonstationary systems with measurement noise. The approach is to discretize the state of the system (or equivalent measurable state), the input vector and time (in the case of a nonstationary system). For these discretized sets of values, the response i.e. the state at t + Δt is determined and stored, thus giving a “stored response” model SRM. The response for arbitrary input vector (within the class for which the model was made) is then obtained by interpolating stored responses for the current state vector, input vector and time thus yielding the state at the next Δt. Repeating this procedure produces the model’s dynamic response. The method of building the SRM table and using it is discussed and several examples are given. An optimal control problem is solved using the SRM model.