Optimum controller design of an overhead crane: Monte Carlo versus pre-filter-based designs

Abstract

In this paper, we consider an overhead crane model with some of its parameters known within given upper and lower bounds. It is shown that a controller with single valued gains based on the nominal values of the parameters could lead to unsatisfactory performance. Two design methodologies are applied to guarantee the desired performance of the considered overhead crane model and their results are compared. The first method utilizes the Monte Carlo approach to find optimum values of the controller gains based on minimizing a least-square error function representing the deviation of the swing angle from a desired trajectory. The second method uses a pre-filter to ensure that the closed-loop performance lies within two given desired responses. Simulation results are provided to compare the performance of the two methods where both show reasonable responses. The Monte Carlo-based approach, however, does not require additional hardware components, compared with the pre-filter technique.