Nonlinear missile planar autopilot design based on PD-spectrum assignment

Presents the nonlinear design of an angle of attack tracking autopilot for a tail fin controlled pitch missile. The autopilot design relies on two separate components. The first is a neural network which implements a pseudo-inversion of the given plant so as to place the states on the nominal trajectory. The second design component is a PD-spectrum assignment controller which exponentially stabilizes the linearized error dynamics about the desired trajectory. A similar missile problem has been treated using this PD-spectrum assignment control but it involved a unique transformation to a tractable form. The design in this paper relies on a Lyapunov transformation of the linearized dynamics which is applicable to any nth-order, uniformly completely controllable system. Moreover, the design method is applied to a more realistic missile model which is not susceptible to the previously published unique transformation. After a brief background on PD-spectrum is given, the particular nonlinear plant is presented along with some simulation results