Perpustakaan Digital ITB

Dynamic characteristics of an aircraft is significantly determined by its configuration, whose design is dictated by the purpose and mission of the aircraft. as example, a fighter aircraft usually has a fast and often unstable dynamics. Therefore, automatic flight control system is a mandatory for this kind of aircraft since controlling its dynamics is very demanding for a human operator. The control system for a fighter aircraft must be able to anticipate its fast dynamics and possible significant change of its parameters due to its wide operating regime and extreme maneuvers. In this research, adaptive model predictive control method is implemented as a multivariable flight control system for fighter aircraft. The method has feature to optimize and generate the required control action based on prediction of the future input and plant response. The predicted variables are computed based on predictor model which are determined via identification process. Hence, this method can be viewed as an optimized control scheme that considers the change of the plant dynamics and possible response of the plant in the future. These features are then explored to accommodate the need for controlling a fighter aircraft by considering some aspects that may arise during the implementation, i.e., change of plant parameter, time-delay dynamics, nonlinearity, and complexity of the maneuver. Adaptive model predictive control scheme is developed and implemented in multivariable linear and nonlinear models of two fighter aircrafts. The first is a linear model with redundant feature and ability to provide asymmetric input that induces its dynamics in six degrees-of-freedom. Control method performance to handle time-delay is also able to be observed using the linear model. The second model, nonlinear one, is exploited to investigate the performance of the control method to anticipate nonlinearity. The results show that the developed control method has good capability to anticipate the change of plant parameter and nonlinearity. While to some extent, the method can overcome delay in the input and achieve good multivariable tracking performance for a complex maneuver.