Suboptimal Design of Linear Regulator Systems Subject to Computer Storage Limitations
Abstract¶
The feasibility of taking practical engineering constraints into consideration when designing optimal linear regulator systems is investigated. The study is conducted by prespecifying the structural form of time-varying feedback gains, while leaving various free parameters to be chosen optimally. In this manner, a suboptimal linear regulator problem is precisely formulated. Necessary conditions for its solution are obtained by introducing the concepts of a cost matrix and of a gradient matrix of a trace function. An algorithm is developed for computing the suboptimal control when the feedback gains are constrained to be piecewise constant. A numerical example illustrates the usefulness of the method.
Summary¶
Develops methods for designing suboptimal linear regulators when computer storage constraints limit the complexity of the feedback law. Studies how to approximate optimal LQR controllers under computational restrictions, with relevance to real-time embedded control systems of the era.
Links¶
Tags¶
Similar Papers¶
Linear Controller Design: Limits of Performance via Convex Optimization
Convex Controller Design Limits
The Internal Model Principle for Linear Multivariable Regulators
Internal Model Principle
On the Robustness of LQ Regulators for Discrete-time Systems
Robust Discrete-Time LQ Regulation