Optimal filtering and LQR theory with constraints on system and control matrices
Abstract
Some recent results are presented from the formulation of the classical linear quadratic regulator problem. The approach is based on direct optimization in the space of operators, thereby allowing constraints on the operator types and norms. It is shown that the constrained optimal regulators and optimum filters converge to the classical ones as the constraints are removed. In addition, some generalities and possible generalizations for filtering problem solutions are presented. As a bonus, the developed approach provides a simple and direct derivation of the Kalman-Bucy filter equations. For illustration, some numerical results are shown comparing constrained and unconstrained linear filters.
- Publication:
-
Electrical and Computer Engineering, Volumes 1 and 2 4 p (SEE N93-30215 11-31)
- Pub Date:
- 1990
- Bibcode:
- 1990ecev.confQ....A
- Keywords:
-
- Control Systems Design;
- Control Theory;
- Feedback;
- Kalman Filters;
- Linear Quadratic Regulator;
- Algorithms;
- Matrices (Mathematics);
- Optimization;
- Electronics and Electrical Engineering