Exploring iterative and non-iterative Fourier series-based methods of control optimization in application to a discontinuous capsule drive model

Abstract

The paper explains iterative and non-iterative approaches to control optimization with use of the Fourier series-based method. Both variants of the presented algorithm are used to numerically approximate optimal control of a discontinuous pendulum capsule drive. Firstly, the general algorithm and its two realizations (iterative and non-iterative) are presented. It is shown that the iterative variant assures non-decreasing quality of solutions in subsequent repetitions of the procedure and the background of such guarantees is explained. A numerical example follows: control of a self-propelled capsule drive is optimized using both approaches. Results are compared and discussed. It is expected that the presented methods can be useful in optimal control estimation for complex systems, particularly discontinuous ones.