Aircraft system modeling error and control error

A method for modeling error-driven adaptive control of an aircraft, the method comprising:

• providing a selected aircraft variable, y(k+1), at a time index having a value k+1, as a matrix sum of Wf βf(k) and B u(k), where βf(k) includes the at least one aircraft variable y(k) in a linear or nonlinear format, Wf is a matrix of selected aircraft variable weighting coefficients, u(k) is a control variable vector for the at least one aircraft variable, and B is a matrix of control variable weighting coefficients that is not yet known;
• modeling an aircraft plant operation using a first neural network modeling mechanism, where the first neural network mechanism incorporates an assumption that the aircraft plant is operating within a normal range, without perturbations and without a tracking error vector e(k) that would cause the aircraft plant to experience an excursion outside a normal range of operation;
• providing a finite bound for the tracking error vector e(k) for operation of the aircraft within the normal range;
• when (1) at least one component of the tracking error vector e(k) is experiencing an excursion, determining if (2) return of the at least one component of the tracking error vector e(k) toward a selected reference vector does not lie on or adjacent to a selected controller error characteristic;
• when the conditions (1) and (2) are satisfied for at least one value of the time index k, introducing at least one change in at least one parameter of the neural network modeling mechanism and modeling the aircraft plant operation according to a modified neural network mechanism with the at least one changed modeling parameter;
• and
• when the conditions (1) is satisfied and condition (2) is not satisfied, continuing to model the aircraft plant operation using the first neural network mechanism, with little or no change in any modeling parameter of the first neural network mechanism.