Attacker-Resilient Adaptive Path Following of a Quadrotor with Dynamic Path-Dependent Constraints

Abstract

For most works on constrained motion control in the literature, only constant or time-varying constraints are discussed, which are often conservative and cannot adapt to the dynamically changing operation environment. In this work, in the context of quadrotor operations, we propose a new adaptive path following architecture with dynamic path-dependent constraints, in which the desired path coordinate, desired path speed, and constraint requirements not only depend on a path parameter associated with the desired path, but also can adapt to the presence of an "attacker" nearby. A new concept of "composite barrier function" has been proposed to address both safety and performance constraints in a unified structure. Adaptive laws are introduced to estimate the upper bounds of system uncertainties and unknown "attacker" velocity. Exponential convergence into a small neighborhood around the equilibrium for position tracking error can be guaranteed. In the end, a simulation example further demonstrates the effectiveness of the proposed architecture.