David Di Ruscio

On State Space Model Based Predictive Control

Abstract An input and output model is used for the development of a model based predictive control framework for linear model structures. Different MPC algorithms which are based on linear state space models or linear polynomial models fit into this framework. A new identification horizon is introduced in order to represent the past.

Model predictive control and identification: a linear state space model approach

In this paper a linear state space model predictive control algorithm is applied to a thermo-mechanical pulping refiner. The paper shows that input and output constraints can be incorporated into the state space model based control algorithm. The properties of the model predictive control (MPC) algorithm as well as comparisons with other MPC algorithms have been presented earlier by the author. A short review of the predictive control algorithm which is extended to handle constraints, is presented in this paper.In this paper a linear state space model predictive control algorithm is applied to a thermo-mechanical pulping refiner. The paper shows that input and output constraints can be incorporated into the state space model based control algorithm. The properties of the model predictive control (MPC) algorithm as well as comparisons with other MPC algorithms have been presented earlier by the author. A short review of the predictive control algorithm which is extended to handle constraints, is presented in this paper.

Methods for The Identification of State Space Models from Input and Output Measurements

Abstract In this paper we present a simple and general algorithm for the combined deterministic stochastic realization problem directly from known input and output time series. The solution to the pure deterministic as well as the pure stochastic realization problem pulls out as special cases of the method presented.

A method for the stabilization of linear feedback systems

A method for the design of controllers with a specific arbitrary structure for linear multivariable time-invariant systems is presented. This method gives a solution to the problem of designing decentralized controllers and includes feedback from a reduced state vector and the output feedback gain matrix. The method does not need to be initialized by a stabilizing controller and is therefore a solution to the stabilization problem. A proof that the algorithm will converge to a stabilizing controller under nonrestrictive assumptions is given. The solution corresponds, at least, to a local minimum for a design objective.<<ETX>>A method for the design of controllers with a specific arbitrary structure for linear multivariable time-invariant systems is presented. This method gives a solution to the problem of designing decentralized controllers and includes feedback from a reduced state vector and the output feedback gain matrix. The method does not need to be initialized by a stabilizing controller and is therefore a solution to the stabilization problem. A proof that the algorithm will converge to a stabilizing controller under nonrestrictive assumptions is given. The solution corresponds, at least, to a local minimum for a design objective. >