Variational Analytic Programming for Synthesis of Optimal Control for Flying Robot

Abstract

The paper presents a novel numerical approach for automatic feedback control system synthesis. The controller is supposed to be described by the control function that depends on objects state. The developed variational analytic programming method allows to construct automatically the mathematical expression of the control function and also to adjust the parameters. The method organizes search of the optimal control function over the set of the small variations of the given basic solution. Search efficiency depends on the quality of the chosen basic solution. An example of automatic control system synthesis for flying robot with state constraints is shown.