Neural-Sliding Mode Augmented Robust Controller for Autolanding of Fixed Wing Aircraft

Neural and sliding mode controllers are generally based on the principle of nonlinear dynamic inversion. This leads to control signals containing high frequency components. This can result in actuator rate limiting due to loss of phase at higher frequencies. Large control inputs, for example due to the saturation component of the sliding mode controller can also result in position saturation of the actuator. In this paper we show that by the introduction of suitable phase compensators and an anti-windup scheme the neural-aided sliding mode controller performance can be improved. A novel scheme is proposed for the cascaded feedback controller which addresses practical requirements of both state limiting and control surface saturation respectivel