Date of Award

Summer 1996

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Electrical & Computer Engineering

Program/Concentration

Electrical Engineering

Committee Director

Oscar R. Gonzalez

Committee Member

Stephen Zahorian

Committee Member

Thomas E. Alberts

Call Number for Print

Special Collections LD4331.E55 S646

Abstract

In this thesis a computationally efficient Generalized Predictive Control (GPC) algorithm is presented and implemented. The algorithm is more efficient than others because the number of iterations needed for convergence is significantly lower with Newton-Raphson. The main additional cost with Newton-Raphson algorithm is the calculation of the Hessian. This overhead is not a problem because of the reduced number of iterations, making the algorithm suitable for real-time control. For nonlinear control applications, a neural network is used as a dynamical system predictor leading to a Neural Generalized Predictive Control (NGPC) algorithm which is presented in detail in this thesis. An advantage of GPC is that physical constraints can be easily incorporated. This thesis includes hard actuator constraints using a differentiable function. A real-time procedure to control unstable plants with an untrained neural network is also presented. In this case the neural network is initialized with an embedded linear model of the process.

Rights

In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).

DOI

10.25777/7vak-qn83

Download

Share

COinS