(A) study on adaptive control methods for flight vehicle

Abstract

In this dissertation, adaptive control methods for flight vehicles are studied. Conventional flight controllers and existing adaptive flight controllers are reviewed and discussed. Most of the existing adaptive control algorithms which are designed in the point of stability require rather restrictive plant assumptions. Thus, those algorithms are not easily applicable to flight vehicles with multivariable nonlinear fast time-varying dynamics. To overcoem this difficulty, noble stable discrete-time multivariable MRAC algorithms with weak assumptions on the plant are developed in this disertation. In the continuous-time domain, new multiariable MRAC algorithms with prescribed degree of stability are presented to increase the tracking speed of plant states to reference model states. Even though some algorithms with weekend assumptions and faster convergence rate are developed in this dissertation, they still have difficulties to be applied to flight vehicles with nonlinear fast time-varying dynamics. As a possible solution for this difficulty, by augmenting an well-proven conventional flight control algorithm, a Two-Level Adaptive Flight Control (TLAFC) algorithm is proposed. At the first level the gainscheduling control is used to give a desired close loop performance characteristics for an assumingly known nonlinear plant dynamics, and at the second level the model reference adaptive control is used to regulated performance deviations owing to modelling errors and parameter variations ignored int the first level control. With this two-level structure, nonlinear timevarying dynamics as well as modelling errors and parameter variations can be handled simultaneously. TLAFC system is designed in both the discrete-time domain and continuous-time domain with the adaptive control algorithms developed in this disertation. The TLAFC concept is applied to longitudinal and lateral motion of F-8 aircraft for simulation to demonstrate the effectiveness of TLAFC.