Determination and stable tracking of feasible aircraft trajectories using adaptive inverse dynamics

This dissertation investigates a novel approach for generating and tracking of feasible trajectories for autonomous systems like futuristic Uninhabited Aerial Vehicles. An aircraft, for most of its flight envelope, is inherently an under-actuated system, which poses a difficult non-flat problem for applying “Inverse Dynamics” to generate a feasible trajectory, due to rank deficiency it, the system in our approach we introduce an innovative technique to overcome the problem of rank deficiency by using pseudo forces. An interesting feature of a proposed trajectory representation approach is the use of basis functions dig allows for systematic perturbation of trajectories and hence determination of a feasible solution. In this dissertation we also present an n-dimensional “Finite Element Method” to represent a large aircraft model accurately, by using piecewise local approximations, each approximation having its node at a centroid of validity associated in a hypercube with dud local approximation.; Another interesting feature, used throughout this dissertation, is to recognize the “truth” that the kinematics of most mechanical systems are exact differential equations, which presents an important structure in the mathematical model of the system dynamics, a structure that is vigorously incorporated into an adaptive control formulation. The dissertation also presents Structured Adaptive Model Inversion (SAMI), which is used for designing adaptive flight control laws to track a target reference trajectory. It is realized that for a rigid aircraft problem, the mathematical model is uncertain only in the aerodynamics, and propulsive influences. The SAMI controller continuously estimates the uncertain parameters using an adaptation law. Further the nonlinear control law presented here realms prescribed linear tracking error dynamics in the kinematic states' vector. The overall closed loop dynamics is non-linear in two respects, the non-linear plant itself, and the non-linear adaptation process. The controller designed using SAMI enforces the desired kinematic states' error dynamics and guarantee tracking stability in the presence of model errors and unknown external disturbances. We show that the resulting closed loop system is globally stable, although no claim can be made with regards to convergence of the adaptation parameters, as usual for controllers of this class.