Optimal input design for flight vehicle system identification via dynamic programming and the direct method for optimal control

An experiment on a dynamic system to generate data for parameter estimation includes exciting the system with a particular input signal. The outputs of the system are measured and recorded during the experiment. The input signals are to be designed in a way that the information content in the recorded data is maximized. This leads to a more precise estimation of the model parameters based on the measurement data for a given test runtime. The task of designing inputs for parameter estimation can be formulated as an optimal control problem. Developing an appropriate method to solve this non-convex optimal control problem, with the focus on flight vehicle system identification, is the main purpose of this paper. In this study we apply a combination of dynamic programming and optimal control using the direct method in order to solve the problem for linear and nonlinear aircraft dynamic systems. The developed solution methodology has been tested and verified based on numerical experiments.     «

An experiment on a dynamic system to generate data for parameter estimation includes exciting the system with a particular input signal. The outputs of the system are measured and recorded during the experiment. The input signals are to be designed in a way that the information content in the recorded data is maximized. This leads to a more precise estimation of the model parameters based on the measurement data for a given test runtime. The task of designing inputs for parameter estimation can...     »