In general, existing methods to develop an effective input for multiple-input/multiple-output (MIMO) control do not offer flexibility to account for limitations in experimental test setups or tailor the control to specific test objectives. The work presented in this paper introduces a method to leverage global optimization approaches to define a MIMO control input to match a data set representing field data. This contrasts with traditional MIMO input estimation methods which rely on direct inverse methods. Efficacy of the iterative optimization method depends on the objective function and optimization method used as well as the definition of the format of the input cross-power spectral density (CPSD) matrix for the optimization routine. Various objective functions are explored in this work through sampling as well as implementation within the iterative optimization process and their impact on the resulting output CPSD. Performance of iterative optimization is assessed against the traditional, direct pseudoinverse method of obtaining the input CPSD as well as the buzz method and weighted least squares (LS). Constraints can be used within the optimization process to control the magnitude and other aspects of the input CPSD, which allows for shaker limitations to be accounted for, among other considerations. Iterative optimization can provide the best input CPSD possible for a test setup while accounting for any shortcomings the setup may have, including force and voltage constraints, which is not possible with traditional methods.
Pyomo and Dakota are openly available software packages developed by Sandia National Labs. In this tutorial, methods for automating the optimization of controller parameters for a nonlinear cart-pole system are presented. Two approaches are described and demonstrated on the cart-pole example problem for tuning a linear quadratic regulator and also a partial feedback linearization controller. First the problem is formulated as a pseudospectral optimization problem under an open box methodology utilizing Pyomo, where the plant model is fully known to the optimizer. In the next approach, a black-box approach utilizing Dakota in concert with a MATLAB or Simulink plant model is discussed, where the plant model is unknown to the optimizer. A comparison of the two approaches provides the end user the advantages and shortcomings of each method in order to pick the right tool for their problem. We find that complex system models and objectives are easily incorporated in the Dakota-based approach with minimal setup time, while the Pyomo-based approach provides rapid solutions once the system model has been developed.