Nth Order Self Adapting Control Systems

Abstract

<p> The very sophisticated control systems of today are built around computers. It is felt that an improved form of cost function in vector or matrix form is needed to fully and most easily utilize the computer's advantages. After defining a vector cost function G , the problem of adapting and learning simplifies to the solution of a partial difference equation. Total system properties are easily defined as matrix arrays which are "learned" in an adapting and "learning" control loop.</p> <p> The relative merits of open and closed loop adaptive systems were investigated. The Nth order adaptive control system was finally chosen to be closed loop after developing two criterion equations in two unknowns which, if satisfied guaranteed improved system sensitivity with the closed loop configuration.</p> <p> Finally, several simple examples are given in experiment form to demonstrate the applicability of the proposed control system techniques.</p>