Feasibility study of a novel method for real-time aerodynamic coefficient estimation

Publication Year:
2010
Usage 144
Downloads 130
Abstract Views 14
Repository URL:
https://scholarworks.rit.edu/theses/5799
Author(s):
Gurbacki, Phillip
Tags:
Lyapunov estimation; Nonlinear system ID; Parameter estimation; Real-time estimation; Sliding mode control; Sliding observer
thesis / dissertation description
In this work, a feasibility study of a novel technique for the real-time identification of uncertain nonlinear aircraft aerodynamic coefficients has been conducted. The major objective of this paper is to investigate the feasability of a system for parameter identification in a real-time flight environment. This system should be able to calculate aerodynamic coefficients and derivative information using typical pilot inputs while ensuring robust, stable, and rapid convergence. The parameter estimator investigated is based upon the nonlinear sliding mode control schema; one of the main advantages of the sliding mode estimator is the ability to guarantee a stable and robust convergence. Stable convergence is ensured by choosing a sliding surface and function that satisfies the Lyapunov stability criteria. After a proper sliding surface has been chosen, the nonlinear equations of motion for an F-16 aircraft are substituted into the sliding surface yielding an estimator capable of identifying a single aircraft parameter. Multiple sliding surfaces are then developed for each of the different flight parameters that will be identified. Sliding surfaces and parameter estimators have been developed and simulated for the pitching moment, lift force, and drag force coefficients of the F-16 aircraft. Comparing the estimated coefficients with the reference coefficients shows rapid and stable convergence for a variety of pilot inputs. Starting with simple doublet and sin wave commands, and followed by more complicated continuous pilot inputs, estimated aerodynamic coefficients have been shown to match the actual coefficients with a high degree of accuracy. This estimator is also shown to be superior to model reference or adaptive estimators, it is able to handle positive and negative estimated parameters and control inputs along with guaranteeing Lyapunov stability during convergence. Accurately estimating these aerodynamic parameters in real-time during a flight is essential for advanced control systems that require the estimated parameters as inputs.