An autonomous systems control theory: an overview

Abstract

A formal theory for characterizing the control of autonomous, dynamical systems is outlined. The main goal is to establish a coherent blend of conventional feedback control and logic-based model theories that provides a unified framework for design and analysis of autonomous systems. A descriptive style is used to highlight the elements of the theory and some basic results. The theory is based on the representation of dynamical systems by evolution models. Although the aim is to develop the theory for general dynamic models, the focus is on a class of systems whose dynamics and requirements are properly represented by an amalgamation of an evolution object and a logic object. The evolution object is characterized by coupled ordinary differential, integral (with time as the independent variable), and algebraic equations and inequalities over a suitable vector space. The logic object is a set of generic and customizable (problem dependent) axioms and theorems that prescribe the requirements imposed on the system, its interaction with the environment over time, and the design principles that permit inference-based online design of control laws. >