Semi-Automatic Generation of Control Law Parameters for Generic Fighter Aircraft

Abstract: Control law design can be an iterative and time-consuming process. The design procedure can often include manual tuning, not uncommonly in the form of trial and error. Modern software tools may alleviate this process but are generally not developed for use within any specific industry. There is therefore an apparent need to develop field-specific tools to facilitate control law design.The main contribution of this thesis is the investigation of a systematic and simplified approach to semi-automatic generation of control law parameters for generic fighter aircraft. The investigated method aims to reduce human workload and time spent on complex decision making in the early stages of aircraft development. The method presented is based on gain scheduled LQI-control with piece-wise linear interpolation. A solution to the automated tuning problem of the associated weighting matrices Q and R is investigated. The method is based on an LQ-optimal eigenstructure assignment. However, the derived method suffers from problem regarding practical implementation, such as the seemingly narrow LQ-optimal root-loci of the linearized aircraft model.Furthermore, the inherent problem of hidden coupling is discussed in relation to gain scheduled controllers based on conventional series expansion linearization. An alternative linearization method is used in order to circumvent this problem. Moreover, the possible benefits and disadvantages of control allocation is addressed in the context of actuator redundancy. It is concluded that one may achieve a somewhat simpler handling of constraints at the expense of some model accuracy due to the inevitable exclusion of servo dynamics.