Ph.D Thesis

Ph.D StudentLevy Maital
SubjectFull-State Autopilot and Guidance for Multi-Input
Multi-Output Missile Engagements
DepartmentDepartment of Aerospace Engineering
Supervisors PROFESSOR EMERITUS Shaul Gutman
PROF. Tal Shima
Full Thesis textFull thesis text - English Version


Nowadays, interceptor missiles have to overcome a variety of ever-evolving threats. This need imposes strict performance requirements on the missiles, in particular an improved accuracy and extended kill envelope are usually mandatory. The traditional approach to designing autopilot-guidance systems, where each subsystem is designed separately (e.g. commonly the guidance law is designed using a low order approximation of the closed loop missile dynamics), is not expected to improve the overall system performance, since it does not exploit the synergy between these subsystems. On the other hand, the full-state approach to design autopilot-guidance systems has the potential to improve the missile's performance and meet these advanced design requirements, since it provides tighter integration between the guidance and autopilot loops.

This research is concerned with the full-state approach to designing autopilot and guidance systems. Two full-state G&C architectures are considered in this research: full-state single-loop and full-state two-loop. In both cases, there is a full-state feedback into the guidance law. In the full-state single-loop case, the guidance law commands the actuators directly, thus it is being related directly to the airframe dynamics, and is expected to achieve the optimal performance subject to a given dynamical system. In the full-state two-loop case, the inner autopilot loop is designed separately from the outer guidance one.

The autopilot is a crucial component in practical autopilot-guidance systems since it ensures both the inner stability of the airframe and robustness to model uncertainties. As a consequence, the main goal of this research is to identify the merits of the full-state two-loop architecture and determine if under certain conditions it can achieve the superior performance of the full-state single-loop design as well as stabilize the airframe if the guidance law is inactive. Such conditions imply that the two full-state designs are equivalent, i.e. the two minimize the same cost function, under the same differential equations and constraints.

It is proven that under linear quadratic optimal control and differential games formulations the two full-state architectures are equivalent for multi-input systems if and only if the number of guidance commands is identical to the number of available controllers. Furthermore, equivalence conditions between the two full-state architectures are also established for nonlinear systems with bounded controls, regardless of the chosen cost function. In this case, the two full-state guidance laws are equivalent if there is a unique mapping from the servo command and state sets into the guidance command set. The theoretical results and guidance laws are illustrated via simulations to allow further investigation of their performance and characteristics.