This paper extends our previous analysis of an open-loop aircraft flutter model to a more general closed-loop case, as the closed-loop scenario is more practical than the open-loop case. Due to the presence of two measured noises in this general closed-loop aircraft flutter model analysis, the optimal input signal is determined from the perspective of unbiased estimation, i.e., a single unbiased nonparametric flutter model derived solely from the collected closed-loop input-output data. To provide a detailed description of the dependence of the optimal input signal on the nonparametric estimate, we derive an explicit improved form and perform its statistical analysis. Moreover, to guarantee the equivalence between the unbiased estimate and convergence properties, a composite Lyapunov analysis is formulated to account for two different types of noise, similar to convergence-in-noise, i.e., robustness. This composite approach allows for a comprehensive assessment of the system’s behavior under the combined influence of these noises, ensuring that the desired estimation properties are maintained despite uncertainties in the system model, disturbances, and variations in operating conditions.
Finally, a platform is established, and simulations are performed to validate our proposed theoretical results. The main contribution of this paper is to present new ideas on how to simultaneously improve the accuracy of identification within a closed-loop system and address input-output noise. This is demonstrated through experimental studies of the application on an experimental platform for aircraft flutter.